1 UNIVERSIDAD NACIONAL AUTÓNOMA DE MÉXICO DOCTORADO EN CIENCIAS BIOMÉDICAS INSTITUTO DE FISIOLOGÍA CELULAR CONTRIBUCION DEL TÁLAMO A LA ACTIVACIÓN DE LOS GANGLIOS BASALES DURANTE LA INICIACIÓN Y REALIZACIÓN DE SECUENCIAS DE ACCIONES TESIS QUE PARA OPTAR POR EL GRADO DE DOCTOR EN CIENCIAS PRESENTA EDGAR ARTURO DÍAZ HERNÁNDEZ DIRECTOR DE TESIS DR. FATUEL TECUAPETLA AGUILAR INSTITUTO DE FISIOLOGIA CELULAR COMITÉ TUTOR DRA. DIANA MARIA ESCALANTE ALCALDE INSTITUTO DE FISIOLOGIA CELULAR DR. VÍCTOR DE LAFUENTE FLORES INSTITUTO DE NEUROBIOLOGIA UNAM – Dirección General de Bibliotecas Tesis Digitales Restricciones de uso DERECHOS RESERVADOS © PROHIBIDA SU REPRODUCCIÓN TOTAL O PARCIAL Todo el material contenido en esta tesis esta protegido por la Ley Federal del Derecho de Autor (LFDA) de los Estados Unidos Mexicanos (México). El uso de imágenes, fragmentos de videos, y demás material que sea objeto de protección de los derechos de autor, será exclusivamente para fines educativos e informativos y deberá citar la fuente donde la obtuvo mencionando el autor o autores. Cualquier uso distinto como el lucro, reproducción, edición o modificación, será perseguido y sancionado por el respectivo titular de los Derechos de Autor. 2 Agradecimientos académicos • Al posgrado en Ciencias Biomédicas de la UNAM por la oportunidad de formar parte del programa, que fomentaron mi crecimiento como persona y científico. • El desarrollo de este proyecto fue gracias al apoyo económico por parte del Consejo Nacional de Ciencia y Tecnología de Ciencia Básica 220412, fronteras de la ciencia 2022 y Beca Nacional CONACyT para estudios de posgrado número 574086. Así como los donativos por parte del programa de apoyo a proyectos de investigación e innovación tecnológica (PAPIIT-DGAPA) IA200815 y IN226517. 3 Agradecimientos Personales 4 Índice Resumen---------------------------------------------------------------------------------------------------------7 Abstract----------------------------------------------------------------------------------------------------------8 Introducción----------------------------------------------------------------------------------------------------9 Antecedentes-------------------------------------------------------------------------------------------------10 Secuencias de acciones-------------------------------------------------------------------------10 Modelo serial o jerárquico para la formación de secuencias de acciones-----11 Actividad neuronal dentro del estriado que apoya al modelo serial o jerárquico en la generación de secuencias ----------------------------------------------11 Circuito corteza-ganglios-basales-tálamo-corteza en la generación de secuencias de acciones-------------------------------------------------------------------------13 Proyección tálamica hacia el estriado------------------------------------------------------16 Participación de la proyección tálamo-estriatal a la generación de acciones----------18 La proyección tálamo-estriatal en patologías donde se afectan las secuencias de movimientos--------------------------------------------------------------------------------------------------18 Planteamiento del problema-----------------------------------------------------------------------------19 Pregunta de investigación-------------------------------------------------------------------------------19 Hipótesis-------------------------------------------------------------------------------------------------------20 Objetivos-------------------------------------------------------------------------------------------------------20 Métodos--------------------------------------------------------------------------------------------------------21 Animales empleados-------------------------------------------------------------------------------21 Transfección viral-----------------------------------------------------------------------------------21 Implantación de electrodos---------------------- -----------------------------------------------22 Entrenamiento de los animales a realizar secuencias de acciones----------------22 Registros electrofisiológicos en animales en libre movimiento y foto- identificación-----------------------------------------------------------------------------------------23 Inhibiciones optogenéticas----------------------------------------------------------------------23 UNAM – Dirección General de Bibliotecas Tesis Digitales Restricciones de uso DERECHOS RESERVADOS © PROHIBIDA SU REPRODUCCIÓN TOTAL O PARCIAL Todo el material contenido en esta tesis esta protegido por la Ley Federal del Derecho de Autor (LFDA) de los Estados Unidos Mexicanos (México). El uso de imágenes, fragmentos de videos, y demás material que sea objeto de protección de los derechos de autor, será exclusivamente para fines educativos e informativos y deberá citar la fuente donde la obtuvo mencionando el autor o autores. Cualquier uso distinto como el lucro, reproducción, edición o modificación, será perseguido y sancionado por el respectivo titular de los Derechos de Autor. 5 Electrofisiología ex vivo--------------------------------------------------------------------------23 Evaluación de efectos de las manipulaciones optogenéticas sobre la motricidad gruesa en el campo abierto ------------------------------------------------------24 Análisis de datos------------------------------------------------------------------------------------24 Inmunohistoquímica-------------------------------------------------------------------------------25 Determinación de las proyecciones tálamo-estriatales --------------------------------25 Estimación de neuronas marcadas retrógradamente ----------------------------------26 Análisis Estadístico--------------------------------------------------------------------------------26 Resultados-----------------------------------------------------------------------------------------------------27 Proyección tálamo-estriatal--------------------------------------------------------------------27 Entrenando ratones a iniciar / realizar secuencias de acciones-------------------33 Registro de neuronas talámicas mientras los animales inician o desarrollan una secuencia de acciones---------------------------------------------------------------------34 Registro de las neuronas tálamo-estriatales mientras inician / desarrollan una secuencia de acciones---------------------------------------------------------------------------37 Inhibición de las neuronas talámicas durante el inicio / realización de una secuencia de acciones---------------------------------------------------------------------------41 Inhibición de las terminales sinápticas tálamo-estriatales--------------------------43 Inhibición de las terminales tálamo-estriatales durante el inicio de una secuencia de acciones---------------------------------------------------------------------------45 Inhibición de las terminales tálamo-estriatales durante la realización de una secuencia de acciones---------------------------------------------------------------------------46 La inhibición de la vía PFs-DLS no evoca cambios durante el inicio / realización de secuencias de acciones---------------------------------------------------------------------47 La inhibición de la vía VPs-S1 no afecta el inicio / realización de secuencias de acciones----------------------------------------------------------------------------------------------48 6 La inhibición de las sinapsis VPs→DLS en la primera sesión de entrenamiento no afecta el inicio / realización de secuencias de acciones--------------------------49 La inhibición de las sinapsis PFs→DLS o VPs→DLS no modifica el desplazamiento horizontal de los animales en la prueba de campo abierto---50 La eficacia sináptica en las sinapsis Pf→DMS se encuentran disminuida en un modelo de compulsiones SAPAP3 (-/-) ----------------------------------------------------51 Discusión------------------------------------------------------------------------------------------------------54 Conclusiones ------------------------------------------------------------------------------------------------59 Referencias----------------------------------------------------------------------------------------------------60 Información suplementaria------------------------------------------------------------------------------66 7 Resumen Las secuencias de movimientos son esenciales para la vida de los animales, esto hace necesario conocer la actividad neuronal responsable de la generación de estas. La actividad neuronal del estriado (principal entrada de los ganglios basales) es necesaria para el inicio y realización de secuencias de acciones. El estriado tiene dos principales aferencias excitatorias: la corteza cerebral y el tálamo. La corteza motora parece no participar en la ejecución de secuencias de acciones mientras que se sabe que el tálamo procesa información sensorial, que podría contribuir a la elección de acciones, pero no sabemos si esta actividad contribuye al inicio/ejecución de secuencias. El presente trabajo investigó la contribución tálamo-estriatal durante el inicio y realización de secuencias de acciones. Primero identificamos las proyecciones tálamo-estriatales hacia los compartimentos del estriado dorsal. La región tálamica del núcleo parafascicular (PF) presentó proyecciones preferentes al estriado dorso-medial (DMS), y la región ventral posterior (VPs) hacia el estriado dorsolateral (DLS). Una vez que identificamos estas dos proyecciones preferentes (PF→DMS y VPs→DLS), registramos la actividad de neuronas en el PF y los VPs e identificamos que la actividad neuronal de esta estructura correlaciona con el inicio y realización de secuencias de acciones. Posteriormente en un experimento de inhibición optogenética de las terminales tálamo-estriatales mostramos que su inhibición al estriado, retrasan el inicio de la secuencia de acciones, y solo la inhibición de los VPs→ DLS alteró el patrón de la realización de la secuencia. Por último, en experimentos ex vivo mostramos que la conectividad funcional de las sinapsis PF→DMS se encuentra disminuida en un modelo genético de compulsiones. Como conclusión de nuestros experimentos identificamos que las sinapsis tálamo→estriatales contribuyen al inicio y ejecución de secuencias de acciones y sugerimos que la desregulación de estas sinapsis podría estar implicada en la generación de conductas compulsivas. UNAM – Dirección General de Bibliotecas Tesis Digitales Restricciones de uso DERECHOS RESERVADOS © PROHIBIDA SU REPRODUCCIÓN TOTAL O PARCIAL Todo el material contenido en esta tesis esta protegido por la Ley Federal del Derecho de Autor (LFDA) de los Estados Unidos Mexicanos (México). El uso de imágenes, fragmentos de videos, y demás material que sea objeto de protección de los derechos de autor, será exclusivamente para fines educativos e informativos y deberá citar la fuente donde la obtuvo mencionando el autor o autores. Cualquier uso distinto como el lucro, reproducción, edición o modificación, será perseguido y sancionado por el respectivo titular de los Derechos de Autor. 8 Abstract The generation of sequences of movements is essential for the life of animals, this makes it necessary to understand the neuronal activity responsible for their generation. The neuronal activity of the striatum (the main entrance of the basal ganglia) is necessary for the initiation and the execution of sequences of actions. The striatum has two mains excitatory afferences, one from the cerebral cortex and the other from the thalamus. Recently the cortex has been challenged as the main driver in the initiation and execution of action. On the other hand, the thalamus is known as a stage to process sensory information, which can contribute to the choice of actions, but it is not known whether its activity contributes to the start / execution of sequences. The present work investigated the thalamus-striatal contribution to the initiation and execution of action sequences. First, we identified the thalamo-striatal projections innervating the dorsal striatum. We observed that the parafascicular thalamic region (PF) preferentially innervate the dorsomedial striatum (DMS), while ventral posterior thalamic region (VPs) preferentially innervate the dorsolateral striatum (DLS). Once we identified these two preferred projections to the dorsal striatum (PF→DMS and VPs→DLS). We recorded the neuronal activity of the thalamo→striatal neurons (in the PF and the VPs) and identified that it correlated with the initiation and execution of an action sequence. Afterward in an experiment of optogenetic inhibition of the thalamo-striatal terminals we observed that their inhibition to striatum, delayed the initiation of action sequences, and only the inhibition of VPs → DLS altered the execution of the action sequence. Finally, in ex vivo experiments we show that the functional connectivity of the PF→DMS synapses is diminished in a genetic model of compulsions. As conclusion of our experiments, we identify that the thalamo-striatal synapses contribute to the initiation and execution of sequences of actions and we suggest that the downregulation of thalamo→striatal synapses could be involved in the generation of compulsive behaviors. 9 Introducción El cambio y la apropiada selección entre acciones son necesarias para tomar decisiones de manera flexible y eficiente (Gillan et al., 2011). El conocimiento de núcleos blanco sobre la actividad neuronal que controla la selección, inicio y ejecución de acciones, nos permitirá hacer predicciones acerca de en dónde existen alteraciones, por ejemplo, en la enfermedad de Parkinson, y/o cuando se presentan síndromes de compulsiones. Los circuitos responsables de la selección de acciones se ubican en conexiones de retroalimentación entre la corteza cerebral con los ganglios basales, el tálamo con la corteza y el tálamo con los ganglios basales (Redgrave et al., 2010). El estriado tiene dos principales entradas excitatorias, la corteza y el tálamo. Estudios recientes sugieren que la corteza podría no participar en la ejecución de secuencias aprendidas (Kawai et al., 2015; Kupferschmidt et al., 2017). El tálamo procesa información sensorial que contribuye a la elección de acciones, participando en la demanda atencional (Bradfield and Balleine, 2017; Minamimoto and Kimura, 2002), y está documentado que la inhibición tálamo-estriatal revierte los déficits motores provocados por la ausencia de dopamina (modelo de la enfermedad de Parkinson) (Parker et al., 2016). Estos antecedentes sugieren que las sinapsis tálamo→estriatales son importantes para el control del movimiento, pero no dejan claro si estas proyecciones participan durante el inicio y/o ejecución de secuencias de acciones en condiciones normales. En este estudio establecemos que las sinapsis tálamo-estriatales son requeridas para el apropiado inicio y ejecución de una secuencia de acciones, teniendo particularidades dependiendo del núcleo talámico de origen y la región estriatal en donde se conectan. Y además sugerimos que la desregulación de estas sinapsis podría estar detrás de la generación de compulsiones. UNAM – Dirección General de Bibliotecas Tesis Digitales Restricciones de uso DERECHOS RESERVADOS © PROHIBIDA SU REPRODUCCIÓN TOTAL O PARCIAL Todo el material contenido en esta tesis esta protegido por la Ley Federal del Derecho de Autor (LFDA) de los Estados Unidos Mexicanos (México). El uso de imágenes, fragmentos de videos, y demás material que sea objeto de protección de los derechos de autor, será exclusivamente para fines educativos e informativos y deberá citar la fuente donde la obtuvo mencionando el autor o autores. Cualquier uso distinto como el lucro, reproducción, edición o modificación, será perseguido y sancionado por el respectivo titular de los Derechos de Autor. 10 ANTECEDENTES Selección de secuencias de acciones El aprendizaje de secuencias de movimientos es esencial para la vida de los animales; ej. en nuestra vida cotidiana las realizamos a cada momento: manejar un auto, una bicicleta, tocar un instrumento o escribir en la computadora. Dichas secuencias las aprendimos y ahora las realizamos casi de manera automática. Una secuencia de acciones puede denominarse como una serie de movimientos que llevan a una consecuencia (Jin and Costa, 2015). Las secuencias pueden descomponerse en un inicio (momento en el cual se selecciona la secuencia), realización y fin de la misma (Calabresi and Filippo, 2010). El inicio de una secuencia de acciones puede emerger como una consecuencia de una señal interna, o de actividad preparatoria que sucede antes de realizarla (Schultz and Romo, 1992). Las primeras evidencias experimentales, mostraron que existe una señal interna que se genera en el cerebro. Es decir, en experimentos donde se lesionó la medula espinal de ratas, los sujetos son capaces de iniciar secuencias, que sugiere que la señal desencadenante de una secuencia se encuentra en el cerebro (Lashley, 1952). Una secuencia de acciones, parte de comportamientos innatos o a partir de comportamientos aprendidos, por ejemplo aprender a escribir en un teclado [en este caso primero se aprende la ubicación de las letras, posteriormente, los sitios donde se encuentran los conjuntos de letras para conformar sílabas, palabras cortas, etc. (Miller, 1956)]. En el ejemplo anterior, las sílabas pueden llamarse pedazos (chunks), particularmente existe evidencia de registros de actividad neuronal en el cerebro que apoya la idea de la construcción de chunks (Fujii and Graybiel, 2003; Jin and Costa, 2010). La generación de chunks está basada en la selección de secuencias; actualmente existen dos teorías de como se da esta selección, una basada en la recompensa y otra en el procesamiento del entorno. La primera teoría, basada en la recompensa, sugiere que la formación de estos chunks sucede a partir de la exploración, en donde la variabilidad sobre la conducta es muy alta. Una conducta se selecciona en base a su asociación con un mayor reforzador; las secuencias se van formando de fragmentos que dan como consecuencia un reforzador o la obtención de éste en mayor frecuencia (Dhawale et al., 2017). La segunda teoría se basa en el entorno, en esta, la parte sensorial juega un papel muy importante; en la primera etapa cada acción tiene una retroalimentación importante, posteriormente pasa a una etapa donde la información UNAM – Dirección General de Bibliotecas Tesis Digitales Restricciones de uso DERECHOS RESERVADOS © PROHIBIDA SU REPRODUCCIÓN TOTAL O PARCIAL Todo el material contenido en esta tesis esta protegido por la Ley Federal del Derecho de Autor (LFDA) de los Estados Unidos Mexicanos (México). El uso de imágenes, fragmentos de videos, y demás material que sea objeto de protección de los derechos de autor, será exclusivamente para fines educativos e informativos y deberá citar la fuente donde la obtuvo mencionando el autor o autores. Cualquier uso distinto como el lucro, reproducción, edición o modificación, será perseguido y sancionado por el respectivo titular de los Derechos de Autor. 11 espacial se vuelve relevante para que finalmente la respuesta motora sea consolidada para que se formen estas secuencias. El modelo que explica el ejemplo anterior se llama de acumulación de evidencia o información (Hikosaka et al., 1999). Modelos serial o jerárquico para la formación secuencias de acciones Existen dos modelos para explicar la ejecución de secuencias de acciones. El modelo serial propone que un movimiento lleva a otro dentro de una secuencia. El modelo jerárquico propone que una secuencia está formada por módulos, es decir, no se necesita la primera evidencia sensorial para resolver una parte de la secuencia (Hikosaka et al., 1999; Lashley, 1952; Miller, 1956). Actividad neuronal dentro del estriado que apoya al modelo serial o al modelo jerárquico de la generación de secuencias El modelo serial, propone que en una secuencia las acciones son realizadas en cadena y si la serie es interrumpida, no es posible completarla. La evidencia experimental mostró que la actividad del estriado se genera en serie cuando un animal realiza una secuencia natural una conducta innata, como el acicalamiento en ratas (Figura 1A-B) (Aldridge and Berridge, 1998); en la Figura 1C se muestra la actividad del estriado alineada al inicio de una secuencia de acicalamiento. Se contrastó que la actividad tiene un orden especifico ya que se observó durante la secuencia de acicalamiento (Figura 1C) y no así en los movimientos que componen la secuencia fuera de manera aislada (Figura 1D) esta idea sugiere que la actividad del estriado se genera en serie a la ejecución de secuencias de movimientos naturales (Aldridge and Berridge, 1998). El modelo jerárquico, propone que las secuencias van resolviéndose por módulos. Es decir, la secuencia tiene diferente orden que un chunk y a su vez una acción (Figura1E). La evidencia experimental que mostró que la generación de secuencias puede ser por módulos, se observó por primera vez sobre manipulaciones en el estriado sobre una secuencia de palanqueos (Tecuapetla et al., 2016). Otra evidencia más muestra que al entrenar ratones a presionar dos palancas con un orden especifico. La manipulación de la actividad de células específicas del estriado, puede substituir los elementos de la secuencia sin cambiar el tamaño de esta, y cuando se manipuló un mando mayor de la secuencia jerárquica, por ejemplo antes de iniciar, se modificó toda la secuencia (Geddes et al., 2018; 12 Tecuapetla et al., 2016) y cuando se manipuló un chunk se modifica solo el chunk, y cuando se manipuló un elemento de la secuencia este no se ve alterado (Geddes et al., 2018) (Figura 1 E-H). Ambos modelos pueden emerger en el estriado, pero, es importante resaltar que el modelo jerárquico se mostró en una innata mientras que el modelo jerárquico se reportó en secuencias aprendidas. Figura 1. Modelo serial versus modelo jerárquico para explicar la ejecución de secuencias de acciones. A) Esquema del modelo serial, modificado de (Jin and Costa, 2015). B) Esquema de las acciones individuales que componen a una secuencia de acicalamiento. C y D) Actividad neuronal registrada en el estriado durante la ejecución de una secuencia de acicalamiento y durante la ejecución de los movimientos individuales que componen la secuencia fuera de la secuencia. E) Esquema del modelo jerárquico, (Jin and Costa, 2015). F) Esquema de una tarea operante en donde se entrenó ratones a presionar 2 palancas en secuencia primero 2 veces izquierda y luego 2 veces derecha (esquema ordenado en forma jerárquica), en la parte inferior se ilustra a un ratón mostrando que antes de presionar la palanca izquierda recibió activación optogenética de las neuronas de proyección estriado-palidales. G) Distribución de palanqueos en ensayos sin y con estimulación optogenética (barra gris), nótese que la estimulación suprimió el primer par de palanqueos, pero no el segundo. H) En este caso la manipulación optogenética fue en la segunda presión de primer par de palanqueos, note que la manipulación no altero la secuencia. I) En este caso la manipulación fue en la primera presión del segundo par de palanqueos, note que en este caso se suprimió la terminación de la secuencia. A B e Serial model EllipticaI Strok./ UniLateral StroM Bilateral Stroke Flank l ic:k ~~\~t¡)~Q Pnase 1 Phase 2 p t\ase 3 Pha5e " "j:,: ,,',,' , ' " "" [:,:"" , ' : " ::' :'.~ : ;: .:~" :':; :' ':, "' , : , ~: :'" " "N" II""IO " ~ '" 1 ' ,,: ,' , ~l ,,' ;:,',' ,', ', ',' , "',' , , ' , ,':, ' ,,' ' . , n I i • '" ' ,,", : , ".',',,:,', ,", :,' ."" ,: " .' , r .: .~ ',' • " I ~ f~: L j~ l ~ L- W~'. : ~ : ~~ o ~ . , Q 1 2 -.2 _, O , 1 -.2 .1 O , 2 -.2 ., O , Time (lee) Activity of each action :.'~t:: ·, ~ , ' : ".:'~ ' : : L :"~:~~;::·;~· : ,f : ~',:: ~ Time (lec) 1l1I 1 ... t 11111' l' 11. l ' • . .. , I I I 11111111' I!! I IRI!':U 1 11 11 ' E Hierarchical model Q:) I s= sequence C= chunk a=action C§J - ~ -------+End 01 seque / I \ I \ F ~ G H Q) 1§Ñ' RR H .. -'--.. ,.-- ,""' ... -0,-----,,'- I 1(8) LL RR -7 LL R 13 Circuito corteza - ganglios basales – tálamo - corteza en la generación de secuencias de acciones Los circuitos responsables de la selección de acciones se ubican dentro de la interconexión entre la corteza cerebral, los ganglios basales, y el tálamo (Redgrave et al., 2010). El estriado es la entrada principal a los ganglios basales, su funcionamiento es fundamental para un apropiado aprendizaje, inicio y ejecución de secuencias de acciones (Jin and Costa, 2010; Jin et al., 2014; Tecuapetla et al., 2016). La actividad del estriado se modula de manera específica durante la ejecución de secuencias de acciones (palanqueos) (Figura 2F) (Jin et al., 2014). Y aunque se sabe que de las dos principales entradas excitatorias al estriado, la corteza presenta modulaciones de actividad especificas durante la generación de secuencias de acciones (Figura 2B-D) (Vandermaelen & Kitai, 1980) (Fujii and Graybiel, 2003) (Tanji, 2001), en el caso del tálamo no es claro como esta entrada modula la actividad estriatal durante la ejecución de estas secuencias de acciones, la evidencia experimental sugiere que el tálamo presenta actividad antes de un movimiento en tareas estimulo-respuesta (Figura 2H-J) (Z. V. Guo et al., 2017), pero no es claro si solo es un relevo sensorial hacia la corteza o contiene información sobre el inicio y ejecución de secuencias de acciones en neuronas talámicas que proyecten al estriado. 14 Figura 2. Actividad del circuito Cortico-ganglios basal-talámico durante secuencias motoras. A) Esquema del loop conectando la corteza ganglios basales y tálamo. B) Esquema de paradigma de secuencias de movimientos sacádicos C y D) Registros en la corteza prefrontal en una tarea de movimientos sacádicos modificadas con (C) y sin recompensa (D) (Fujii & Graybiel, 2003). E) Esquema de un ratón entrenado a presionar una palanca en secuencia. F y G) Raster-plot e histograma alineado al inicio de la secuencia de presiones de una neurona de inicio e Histograma de una neurona de ejecución (Jin et al., 2014). H) Esquema del paradigma, un ratón entrenado a responder a un polo dependiendo de la vibrisa estimulada y en la parte inferior se muestran las etapas de la tarea. I) Raster-plot de una neurona tálamica que incrementa su tasa de disparo cuando se emite la secuencia de lengüeteos. J) Similar a I, pero mostrando una neurona tálamica que presenta actividad preparatoria a la secuencia de lengüeteos (Z. V. Guo et al., 2017). A ( '""5 B E H _ .... , .............. _ $Kutndo.dep'e""'" _ ... ~ I ! 1 - ......... - I M :roo ,. .., fBI f ~~~~~ .\ t I·I~ ' ·'0 \.'0 '·'0 I -- eo...._ .. _ '-- -- • • • I c:tF at -- 15 Los experimentos mostrados en la Figura 2 muestran patrones de la modulación de la corteza y el estriado durante la ejecución de secuencias de acciones, pero no ponen a prueba si estos circuitos contribuyen a la generación de secuencias de acciones. En 2016, Tecuapetla y colaboradores realizando manipulaciones optogenéticas en roedores antes y durante la ejecución de secuencias de palanqueo, concluyeron que inhibir o activar la actividad estriatal hace que los animales tarden más tiempo en iniciar una secuencia de acciones, también observaron que la inhibición del estriado disminuyó el número de acciones dentro de la secuencia, argumentando que el estriado contiene información necesaria para iniciar y ejecutar una secuencia de acciones (Figura 3A-B). Figura 3. El estriado es necesario para iniciar y ejecutar una secuencia de acciones. A) Esquema de inhibición optogenética antes de que el ratón inicie (izquierda), o al ejecutar una secuencia de acciones (palanqueos) (derecha). B) Izquierda, cuantificación del tiempo para iniciar una secuencia durante la inhibición de las neuronas del estriado, derecha, número de presiones que realizó el ratón durante periodos de inhibición. Modificada (Tecuapetla et al., 2016). A B infrared beam triggers lighl Light-on 11 * Light-off Light-on Ma~ne MagaziM ........ ~ ... ;ul .......... · ... ? .~. I B ~' EJ '~' g:¡ 10 1st lever press triggers lighl al ....J 8ft O n ~ Prcss 1 Ociar (targc t .. 700 ms) IPI dist ríbution 700 1 00 Trialnumber(dOOO) 2; O Motor corte)! lesioned 700 .-... .. : ....... _ ... ,.,. ( ~ "._" " ' ~ ... I e; o Trial number (xl000) 30 O * Off On Prcss 2 Rcward M ean IP I 700 --- ----- "--- ." 400>- ---1~0--~2" 0~-;30 Trial Number (xl000) 16 Figura 4. La corteza motora es requerida para el aprendizaje de secuencias motoras. A) Esquema de para desarrollar secuencias de palanqueo con un intervalo especifico. Modificado de (Kawai et al., 2015). B) Izquierda, cuantificación de los intervalos entre presiones antes de la lesión de la corteza motora. En medio, ejemplo de la lesión en la corteza. Derecha, Intervalos de presiones después de la lesión. C) Izquierda, intervalos de ratas lesionadas antes de aprender la tarea. Derecha se muestra la media de los intervalos entre presiones de animales lesionados (rojo) y de animales intactos (negro) antes de aprender. Proyección tálamica hacia el estriado Actualmente muy pocos estudios han cuestionado la causalidad de la proyección tálamo-estriatal sobre la ejecución de secuencias de acciones. Dentro del estriado existen dos tipos de poblaciones de neuronas de proyección, la vía directa (neuronas estriato-nigrales) y la vía indirecta (neuronas estriato-palidales), además de varios tipos de interneuronas (Redgrave et al., 2010; Tepper et al., 2018). En estudios preguntando la conectividad tálamo→estriatal, utilizando trazadores retrógrados no se encontró una preferencia de las proyecciones talámicas por uno de los tipos de neuronas de proyección del estriado (Guo et al., 2015; Wall et al., 2013). Aunque las proyecciones talámicas si muestran una inervación preferente sobre algunas interneuronas (Assous et al., 2017; Choi et al., 2018; Guo et al., 2015). En específico, por medio del marcaje retrogrado tipo celular especifico a través de virus de la rabia modificados, se reportó un mayor número de células marcadas retrógradamente en el núcleo PF con proyección hacia interneuronas positivas a parvalbumina, que neuronas de proyección de la vía directa y mayor que interneuronas colinérgicas en el DMS (Choi et al., 2018; Guo et al., 2015). La proyección tálamo-estriatal se puede clasificar en dos grandes grupos de acuerdo a su blanco de proyección, 1) la proyección preferente al estriado y 2) la proyección de neuronas que dejan colaterales en el estriado pero que preferentemente inervan a la corteza. Estos dos grupos de proyecciones se originaba en los núcleos intralaminares y no intralaminares del tálamo respectivamente (Smith et al., 2004). Actualmente existe una amplia discusión sobre si los diferentes núcleos del tálamo inervan regiones específicas del estriado o presentan proyecciones con diferentes propiedades sinápticas (Alloway et al., 2014; Erro et al., 1999; Guo et al., 2015; Lacey et al., 2007; Pan et al., 2010; Smith et al., 2012; Wall et al., 2013). En particular, se sugiere que la sinapsis tálamo-estriatal, proveniente del núcleo central lateral (CL), el cual es más eficiente en modular la despolarización de las neuronas. Evocando potenciales postsinápticos de gran 17 amplitud (mediados preferentemente por receptores AMPA), mientras que las sinapsis provenientes del PF exhiben potenciales postsinápticos de menor amplitud y depresión a corto plazo, expresando predominantemente receptores de tipo NMDA (Ellender et al., 2013). Además, se sugiere que la proyección tálamo-estriatal es heterogénea hacia los compartimentos del estriado, (Figura 5D) (Hunnicutt et al., 2016), que se refleja provocando diferentes patrones de disparo en las neuronas del estriado (Nanda et al., 2009) (Figura 5 E). Figura 5. Proyección tálamo→estriatal. A) Diagrama representando la proporción de aferencias al estriado versus la corteza (Smith et al., 2014). B) Esquema representando las posibilidades de inervación del PF hacia las neuronas de proyección (SPN) o los diferentes tipos de interneuronas en el estriado (INT) modificada (Assous et al., 2017). C) Corrientes postsinápticas excitatorias (EPSC) evocadas de fibras provenientes del PF o el CL, mostrando que existen diferencias entre las corrientes evocadas dependiendo del núcleo talámico dentro de los núcleos intralaminares modificada (Smith et al., 2014). A e E Types of thalamostriatal projections B -'0 EPSP AMPLlTUDE I Ch"2 stlmulatlon eL madia lvd r8sponse - P1 me- ro * * Q)~ CI) ¡§ 60 .!!1Q) ~~40 "C 2:1 -o g¡ :: 20 Cl)Q) <0""0 .0* o i.L ~ DMS DLS DMS DLS PFs VPS * 32 Figura 8 (página anterior). Proyecciones por trazado anterógrado de los PFs y VPs al estriado dorsal A) 1: Esquema de la inyección a nivel de los PFs, 2: muestran los ejemplos de la inyección del vector viral (en el inserto se muestran la suma de varias fotos en el sitio de inyección), 3: las fibras en el estriado de la inyección en 2 (en el inserto se muestran la suma de varias fotos en el estriado), 4: un zoom de las fibras en el DMS/DLS. B) similar al panel A para la inyección de los VPs. C) Porcentaje de células transfectadas en los sitios de inyección y 300 micras alrededor. D) % de fibras en el DMS vs. DLS por las inyecciones de los PFs o VPs. Figura 9. Registro de la vía tálamo-estriatal es específico por compartimento del estriado dorsal. A) Esquema donde se muestra el experimento. B) Ejemplos de corrientes postsinápticas excitatorias registradas en cada una de las condiciones y en la parte inferior se muestra el gráfico del tamaño de las corrientes postsinápticas por célula y rebanada, los puntos y el error muestran la media de todas las células. C) Probabilidad de encontrar conexión en los diferentes tipos de experimentos. D) inhibición de las corrientes con CNQX (antagonista AMPA) y en la parte derecha se muestra un ejemplo representativo. A B e PFs or VPs : AAV-DIO-ChR2-eYFP (VGluT2Cre) PFs VPs ~~, s ~ ~s ?f!Y) +::jJ ~ "'::jJ PFs + • DMS DLS y----- DMS DLS VPs + • DMS DLS ~ -v ----l 100 pA "'m, DMS DLS Connectividad (%) + DMS PFs VPs • + DLS DMS . (9 t DLS 100 % 70 % 62.5 % 100 % PFs~ DMS D VPs ~ DLS Control Y;X+4.AP y- CNQX ~ :::~ \"\', (f) \ D- o " . W Control CNQX SOms l 100 pA E SPN In! ~ ~ 2omv f- (... 200 pA L 2 00 pA L 100 ms 100 ms F PFs .. ;¡, DMS DLS VPs .SPN ;¡, • No SPN DLS 33 E) Ejemplo de registro de una neurona clasificada como espinosa mediana. F) Distribución de células identificadas por cada circuito. Una vez identificadas las proyecciones del tálamo, identificando que los PFs y los VPs proyectan preferentemente al DMS y DLS respectivamente. Decidimos estudiar las implicaciones de estas proyecciones en el inicio y la ejecución de secuencias de acciones. Entrenando ratones a iniciar / realizar secuencias de acciones Para evaluar la contribución de las sinapsis tálamo→estriatales durante el inicio/realización de una secuencia de acciones entrenamos animales a autoiniciar y realizar secuencias presiones de una palanca (Jin and Costa, 2010; Jin et al., 2014; Tecuapetla et al., 2016). Para esto primero cada animal recibió un reforzador de manera aleatoria cada 30 segundos por media hora (0.02 ml de una solución de sacarosa al 10%) en una caja de comportamiento operante (equipada con una palanca y un comedero). En una segunda etapa de entrenamiento cada animal fue expuesto con una palanca por tres sesiones en las que cada vez que el animal presionó la palanca recibió un reforzador (tres sesiones de reforzamiento continuo; CRF5, 15 y 30). En la última etapa del entrenamiento la contingencia con que los animales recibieron recompensa fue modificada a razón fija 8 (cada que el animal acumulo 8 presiones recibió un reforzador); se mantuvo en esta etapa a los animales por 12 sesiones previo a los protocolos de inhibición optogenética o registros electrofisiológicos (Figura 10). A lo largo del entrenamiento los ratones fueron incrementando el porcentaje de ensayos donde realizaron presiones en secuencia y disminuyeron el porcentaje de ensayos en que realizaron presiones individuales (Figura 10D). Al pasar los días de entrenamiento, los ratones estereotiparon el movimiento de ir del comedero a palanca, por lo que utilizando un sensor infrarrojo entre el comedero y la palanca se midió la latencia para iniciar una secuencia (latencia para iniciar; Figura 10E) (Tecuapetla et al., 2016) (latencias C57BL6/6J = 1± 0.1 s; PF→DMSArch3.0 = 1.2 ± 0.2; PF→DMSeYFP = 1.1 ± 0.1; VPs→DLSArch3.0 = 1.2 ± 0.1; VPs→DLSeYFP = 1.3 ± 0.1; n=6 animales por grupo; Figuras 10E y 10F; p > 0.05, Kruskal-Wallis test). En promedio los ratones realizaron 4.2 ± 0.4 presiones (para los animales Cre: PF→DMSArch3.0= 3.7 ±0.6; PFs→DMSeYFP = 3.3 ± 0.2; VPs→DLSArch3.0= 3.7 ± 0.3; VPs→DLSeYFP= 3.5 ± 0.3; 6 animales por grupo; p< 0.05 Kruskal-Wallis test; Figura 10F, izquierda); con estos experimentos estandarizamos un modelo que nos permitió estudiar el inicio y la realización de una secuencia de acciones. 34 Figura 10. Animales desarrollan secuencias de palanqueo. A) Esquema temporal del paradigma. B) Ejemplo de la medición de la latencia para iniciar. C) Raster-plot de la conducta de un ratón alineada al inicio del palanqueo. D) Gráfico de la proporción de ensayos con presiones en secuencia y presiones individuales. E) Gráfico de las latencias a lo largo del aprendizaje. F) Gráfico del número de presiones y latencia para iniciar en el día previo a las manipulaciones optogenéticas. Registro de neuronas talámicas mientras se inicia / realiza una secuencia de acciones Para determinar la actividad tálamica durante el inicio y ejecución de una secuencia de acciones se decidió implantar ratones con arreglos de 16 electrodos en la región PFs o la región VPs; se registraron 7 animales (3 para el PFs y 4 para los VPs, en un total de 6 y 8 posiciones, respectivamente) (Figura 11). A D Secuencias >= 2 presiones ... C57BL/6J ... PFs-Arch3.0 .... PFs-eYFP ..... VPs-Arch3.0 ... VPs-eYFP Dia test CRF FR8 .J,.. --+ -4 (Dias) B e Inicio =it 30 "' ro ·0 20 e Q) " ~ 10 (f) -1 1 11-12 (Sesiones en FR8) Comedero a IR sensor Ejecución ;I .r ~ ;t .. ,;· .. y'h' 1:. '. 'YLatencia I \': ~ .... + I Presiones li ,,1 T • I .\'(::' ": . .... Recompensa I J. .... : . • Comedero ¡',:";{ '. o " .., .... . - ',; <# :: .- . 3 Tiempo (s) U) QJ U) (/) -O- C57BL/6J ~ 50 ... PFs-Arch3.0 ~ l ~~~~~t~ !!~~~~ l -O- PFs-eYFP ID -o- VPs-Arch3.0 ~ ... VPs-eYFP --' o E 3 5 3 5 F Presiones/secuencia o 6 ~ Q) E 4 ," e .~ "O 2 Q) :2' O ns. "' o "O e " '" Q) (f) 7 9 11 Sesiones (FR8) 7 9 11 Sesiones (FR8) Latencia para iniciar 2 ns. o 35 Se registraron 86 unidades en los PFs y 122 unidades en los VPs. Encontramos neuronas que incrementaron su taza de disparo antes de la primera presión (Figura 11C), y unidades que incrementaron su actividad durante la realización de las presiones (Figura 11D). La modulación de las neuronas registradas se determinó por curvas ROC con 1000 permutaciones, preguntando si la actividad era diferente a una basal tomada 2.5 segundos antes del palanqueo. Se calculó el porcentaje de unidades reclutadas por unidad de tiempo (Figura 11G). El PFs mostró un mayor porcentaje de unidades moduladas en el inicio que los VPs (PFs inicio = 65% versus VPs inicio= 41%; p < 0.05; ×2 test; Figura 11G). Mientras que en los VPs se mostró un mayor porcentaje de unidades moduladas positivamente durante la ejecución de la secuencia (VPs ejecución (+) = 31% versus PFs ejecución (+) = 23%; pastel2 versus 4, p < 0.05, ×2 test). Posteriormente, se determinó sí las unidades eran específicas del inicio de una secuencia o se modulaban en cada presión, para esto alineamos la actividad a la primera presión, segunda penúltima y última; encontrando que el PFs contiene un mayor número de unidades moduladas en el inicio, mientras que la mayor cantidad de unidades moduladas en la ejecución fue en los VPs (Figura 12) (PFs = 42% versus VPs = 27%; p < 0.05, ×2 test; Figura 12C) (PFs = 22% versus VPs = 55%; p < 0.05, ×2 test; Figura 12C). 36 Figura 11. La actividad del PF y los VPs incrementan su tasa de disparo al inicio y durante la realización de secuencias aprendidas. A) Fotografía de un ratón implantado con un electrodo. B) Ejemplo de la lesión representativa por la cánula del electrodo y los esquemas muestran los sitios de los demás animales. C) Ejemplo de 2 unidades registradas en los PFs alineadas al inicio de la secuencia de presiones. D) similar a C, registros en los VPs. E y F) Población de unidades alineadas al inicio de la secuencia de palanqueo registradas en los PFs o en los VPs. G) Gráfico de reclutamiento de unidades a lo largo del inicio y realización de la secuencia. Los pasteles muestran el promedio de unidades 2 segundos antes o 2 segundos después de la primera presión. A B e (PFs) Inicio Electrodo Primera presión ~ Ejecución Palanca ~ ~~;~~~r~~;:~,&tI i' ~ ~ ~ -1 O 1 2 Tiempo (s) Tiempo (s) E (PFs) G . ro -2 -1 o Tiempo (s) 2 ~ Inicio , Ejecución !. ~ (J) 40 ' . ' u ~ 20 , e ' o O ' 2 * cJ. -2 -1 O 1 2 F (VPs) 4 "' 3 " -o ro 2 -o "c: ::> '" O -1 -2 -1 O 2 Tiempo (s) * • 3 4 ~ Inicio : Ejecución "S ' ~ 60 ~ i + moduladas (J) 40 : - moduladas • u ' ~ 20 ' e , ~ ~' 2 ~"-- - - ~ ' -- O;;- ---;-'--"'-' ~ 2 Tiempo (5) N 8 ro 37 Figura 12. Actividad alineada a diferentes momentos de la secuencia de presiones. A) Unidades alineadas a los diferentes eventos de la secuencia, a la primera, segunda, penúltima y última presión registradas en los PFs. B) Similar a A, registros en los VPs. C) Proporción de las unidades por categoría, de acuerdo a su máxima activación. Registro de las neuronas tálamo→estriatales mientras se inicia/ realiza una secuencia de acciones Con el fin de identificar la actividad de neuronas del tálamo que proyectan al estriado mientras los animales iniciaban/realizaban la secuencia de palanqueos, transfectamos ChR2 en neuronas del tálamo y colocamos una fibra óptica en el estriado (DMS o DLS según fue el caso). La idea fue colectar la actividad durante el palanqueo y al final de la sesión preguntar si la neurona registrada respondía a la activación antidrómica desde el estriado (Figura 13) identificándolas como neuronas tálamo-estriatales (Lima et al., 2009). La Figura 13 se muestra un ejemplo de una unidad que presentó estimulación antidrómica; es posible observar la espiga extracelular, su respuesta a los pulsos estimulados y el clúster de distribución de los componentes principales, para determinar que se trataba de la misma unidad; aceptamos un coeficiente de correlación entre la forma de la espiga extracelular registrada en la conducta y la forma registrada durante la foto- estimulación de 10ms, y una latencia de respuesta menor a 8 ms (Vandermaelen & Kitai, 1980) (latencia media PFs→DMS = 5.4 ± 0.3 ms; VPs→DLS = 5.4 ± 0.5 ms; Figure 12B). A Presión 1 -2 -1 o B Presión 1 -2 -1 U Parafascicular región Presión 2 Penultima presión 2 -2 -1 o 2 -2 -1 o Ventroposterior región Ultima presión 2 -2 -1 O 4 3 2N 00' 18: O (ji! -1 :\ 2 liempo(s) Presión 2 Penultima presión -1 Ultima presión 2 -2 -1 O 2 -2 -1 U 2 -2 -1 O 2 Tiempo (5) e PFs VPs % Inicio % Ejecución %Final % Otras categorias 38 Figura 13. Registro de neuronas tálamo-estriatales durante el inicio y realización de secuencias. A) Esquema del experimento para identificar neuronas tálamo→estriatales. B) Ejemplo de una unidad foto- estimulada anti-dromicamente, se muestra la forma de la espiga extracelular registrada. En la parte inferior se muestran las latencias registradas por las neuronas tálamo-estriatales. C y D) Ejemplo de una unidad foto-identificada alineada al inicio de la secuencia registrada en el PF y una en los VPs respectivamente. E y F) Población de unidades foto-identificadas alineadas al inicio del palanqueo para PFs→DMS y VPs→DLS respectivamente. G) Gráfico de reclutamiento del porcentaje de unidades por unidad de tiempo. Los pasteles de los insertos muestran el promedio de las unidades 2s antes vs. 2 s después de la primera presión, en color gris se muestran las unidades que no se modularon. A AAV + Electrodo arreglo (ChR2) e \ Fibra ~ '"" t-_ loL:: 0Ptica VGluT2-Cre (PFs .. DMS) - .:,." i:b1:~ ~, ~ -1 o 1 2 Tiempo (5) E (PFs .. OMS) U> al -o ro 10 -o "E ::l '" 20 G -1 -1 o 1 2 Tiempo (5) * 2 o 2 Tiempo (5) B unidad tálamo .. estr Luz Wf-secuencias 1I I rI' Wf-estlmulacl6n i",' {~': : : " 5O ~V U ~ " . , : ,. Q.3ms ~c" "" " :n,' '... ~ , ". ' ~:': t: e,.:;' PC, pe, ~ ~20 1 IIII w 0\JLa. PFs->OMS VPs->DLS ~1 0 lL.. "05 {lo 'c 048048 · 1 O 1 2 Tiempo (5) =:J Latencia (ms) D (VPs~DL~) .' " "'" Tiempo (5) F (VPs .. OLS) -2 -1 * 3 012 Tiempo (5) 4 I :~ ~ 40 I 20 I o -2 -1 o 1 2 Tiempo (5) 39 Siguiendo esta metodología se logró foto-identificar al 34% de las unidades registradas en los PFs mientras que en los VPs→DLS se logró foto-identificar el 12% (Figura 13E-F). El reclutamiento de neuronas en el tiempo mostró una mayor proporción de unidades moduladas al inicio que durante la realización para el circuito de VPs→DLS (PFs→DMS inicio= 51% de las unidades, en comparación a VPs→DLS inicio= 88% p < 0.05; ×2 test; PFs→DMS ejecución= 24% en comparación con VPs→DLS ejecución= 58%, p<0.05; ×2 test; Figuras 13E–G). Al cuestionar si la actividad observada era específica del inicio de una secuencia o estaba alineada solo a cualquier presión, se evidenció que el PFs→DMS contiene mayor porcentaje de unidades específicas del inicio de la secuencia, mientras que la región VPs mantenía un mayor porcentaje de neuronas asociadas específicamente a la realización, (PFs→DMS inicio = 43% versus VPs→DLS inicio = 33% de las unidades; PFs→DMS realización = 13% versus VPs→DLS realización = 20%; p<0.05, ×2 test; Figura S1G). En conjunto estos resultados mostraron que las unidades tálamo-estriatales que conectan con el estriado dorsal incrementan su taza de disparo alrededor del inicio y ejecución de secuencias de movimientos. Posteriormente decidimos preguntar si la actividad de estas neuronas tálamo- estriatales contenían información del desarrollo o inicio de la secuencia. Para esto realizamos ajustes lineales entre la actividad neuronal y parámetros de la conducta como lo es la latencia en iniciar una secuencia, o el número de presiones. Para esto se calculó la tasa de disparo por ensayo y después agrupamos los ensayos por el tiempo de latencia (Figura 14A) o por el número de presiones (Figura 14E), posteriormente a través de un análisis de regresión calculamos donde la beta ajustada fuera significativamente diferente de cero. Con los datos anteriores, se calculó el porcentaje de codificación positiva y negativa de las unidades tálamo-estriatales. Identificamos un porcentaje alto de unidades que codifican las presiones para los VPs→DLS. También se encontró un mayor porcentaje de unidades que realizan una regresión negativa con la latencia para el PFs→DMS. Al preguntar si existían neuronas específicas que sólo codificaban un parámetro, encontramos que la mayoría de las neuronas tálamo-estriatales codifican ambos parámetros, pero el PFs→DMS contiene unidades que codifican sólo latencia y por su parte el VPs→DLS contiene unidades que sólo codifican presiones y no latencia para iniciar (PF→DMS latencia unidades = 22% versus VPs→DLS unidades sólo latencia = 0%; PFs→DMS unidades sólo relacionas con las presiones = 7% VPs→DLS relacionadas sólo con las presiones = 33%; p<0.05 ×2; Figura 14I). Estos datos en conjunto sugieren que las neuronas tálamo- estriatales contienen información de cómo se codifica una secuencia de presiones. 40 Figura 14. Las neuronas tálamo- estriatales codifican el inicio y número de presiones en una secuencia. A) Ráster plot de una neurona que muestra regresión entre las espigas y el tiempo de latencia para iniciar. B) Regresión de los datos en A. C) Porcentaje de unidades que tuvieron una regresión significativa con la latencia, beta positiva (azul) o negativa(rosa). D) Pasteles que muestran el promedio de unidades en 2s antes o después de la primera presión, que hicieron regresión con la latencia. E) Una neurona que hace regresión con el número de presiones. F) Regresión de los datos mostrados en E. G y H) similar a C y D, en este caso haciendo regresión con el número de presiones. I) Diagrama de Ven que resume las poblaciones registradas y su distribución con la regresión lineal a los 2 parámetros evaluados. A e '" Q) 40 "O ro o "O 'c 60 I " ;f'. o E -2 o 2 Tiempo (s) ~ ~ -2 o 2 Tiempo (s) • Lalency B espigas-Latencia R=O.70 !!! 12 ~ ~ 8 • Ol "g. 4 . - Q) D 123456 Latencia (s) + beta (PFs_DMS) + beta(VPs_DLS) * 1 4 b (PF! ... DMS) _ - beta (VPs- DLS) (PFs .. OMS) (VPs .. OLS) Inicio Ejecución F • Lever press • Splke # se\o ~ , If:>\ L ~'''' , ,1.10 .. • .. " •• ' 4 :1:!;. 40 ' : i',''' ~ ; ~ ' ' ~ : i ' ; ' ~ . ,0, ~ 20 ; ,. ~ ~. @ • ' 0: " ~ a. O ' - -- ~. , .... espigas-Presiones '" 51 R=-O.82 } 31"-.- lfi1 ~ esp~gas: ~ -2 O 2 G H ~ 4~ LI----' ..... +.-.... ~ 6°1 + beta (PFs_DMS) ......... 1 + beta (VPs- DLS) " O L_ --"F=::hS"' .... ~ ,- -2 2 cft.. TIme (sec) - b ta (PF -DMS) .... - beta (VPs - DLS) Inicio 2 4 6 8 Presiones(#) (PFs .. OMS) (VPs .. OLS) Ejecución Latencia Latencia & Presiones Presiones (PFs +-;:..OM~S,--) ___ ---'(_V_PS.:..-,+OLS) 7.5 % 33 % 41 Inhibición de las neuronas talámicas durante el inicio/realización de una secuencia de acciones De acuerdo con los registros in vivo, las neuronas tálamo→estriatales mostraron modularse en el inicio y realización de la secuencia de movimientos. Para comprobar si las neuronas talámicas contribuyen a iniciar/realizar la secuencia, se realizaron inhibiciones optogenéticas, en algunas regiones talámicas, tales como, la región anterior talámica (ANT), la región central medial (CM), PFs y los VPs por medio de Arch.3.0. Las inhibiciones optogenéticas se realizaron en 2 momentos: el primero, cuando el animal cruzó el sensor de latencia. La Figura 15 muestra que la inhibición de los cuerpos celulares de neuronas del tálamo provoco un retraso en el inicio de la secuencia de movimientos cuando se inhibió los somas del PFs y VPs (Figura 15D). Interesantemente la inhibición de los VPs mostro un incremento en el número de presiones en la secuencia, mientras que la inhibición en los PFs mostró un incremento en la duración de la secuencia cuando es estimulado antes que inicie la secuencia (Figura 15F). Por otro lado, cuando se realizó la inhibición cuando el animal se encontraba realizando la primera presión de la secuencia, solo encontramos efecto en la duración de la secuencia que incrementó cuando manipulamos la región del PFs y ANT (Figura 15K). Aunque estos resultados apoyan la idea de que la actividad del tálamo es necesaria para la apropiada ejecución de la secuencia, no informan específicamente acerca de las proyecciones tálamo→estriatales, ya que las neuronas manipuladas en el tálamo pueden proyectar a diferentes blancos, como la corteza, estriado u otras regiones. Por lo tanto, los siguientes experimentos se centraron en conocer la interacción tálamo→estriatal a través de la inhibición de los axones que llegan del tálamo al estriado dorsal. 4 2 F ig u ra 15. L a in h ib ició n d el tálam o en el in icio y ejecu ció n d e secu en cias d e accio n es afecta la ap ro p iad a realizació n d e secu en cias. A ) E squem a del sitio de m anipulación. B y G ) E squem a de m anipulación antes de iniciar y durante la realización de la secuencia. C y H ) R áster-plot de la conducta en condición sin luz y con luz alineado a la prim era presión. D ) G ráficos de latencia para iniciar, en el gráfico pareado se m uestran 3 y 4 anim ales realizados en los V P s y P F s respectivam ente; en los boxs plot se m uestra a cada grupo probado, com o cociente entre estim ulación y no estim ulación on /off durante el día de estim ulación y un día previo. E ) S im ilar a C , cuantificando las presiones. F ) sim ilar a D , cuantificando la duración de la secuencia. I-K ) S im ilar a D -F en este caso para la inhibición del protocolo en G . "T1 Segundos ? N ~ O> o o ~ 'ft " o o On/Off o N ~ O> ,; ,YFPpe" ~ eYFP test PFs prev I • ... •• -:--Pt- ~ PFs test ' VPs prev r«I'" VPs test ' CM prev f ·· CM test DML prev DMLtest ..- ANT prev m .... ANTtest Segundos ~ : ~ On/Off ~O>ooC; ¡¡¡ Q, o' o o. ro ¡¡¡- ~ g ro o n ¡;' " o o ¡¡¡ Q, o' o o. ro ¡¡¡- ~ n eYFP prev eYFP test PFs prev PFs test VPs prev VPs test CM prev CM test ~ '==-:J' ~ o n ¡;' DML prev DML test ANT prev ANTtest ~ g # presiones ~ ~ On/Off eYFPprev ~ eYFPtest --- PFs prev PFs test VPs prev VPs test CM prev CM test DML prev DML test ANT prev ANTtest # presiones ~ m Segundos z o 3 ro a : ~ o. ro 'O ~ ro o o' eYFP prev ::J eYFP lest ~ PFs prev c.... PFs test VPs prev VPs test CM prev CM test DML prev DML test ANT prev ANTlast 9 On/Off~ o Segundos o N ~ T ~ \ ~ :\1 ro g M"" '.'. ~ eYFP prev eYFP test PFs prev PFs test VPs prev VPs test CM prev CM test DML prev DML test ANT prev ANTtest On/Off ro 9 On/Off ~ ~ ~ o o N ~ ~ !!?, eYFP prev :::::lZ: g eYFPtest ro PFs prev (f) PFs test VPs prev VPs test CM prev CM test DML prev DML lest ANT prev ANTtest e r • ro o n ¡;' 11 ¡¡¡ S' ñ' ~' .- ro o n ¡;' 'O • ¡¡¡ S' ñ' ~' ,. n " w 6 ~ ~ ." ." "U "U Secuencias(#) n N ~o ~ o 00 , o N o ...... ~ ......... ~ ... ~ : : ~~~~~~ ,· ~ · ~i ~ .. : ........ ~ ~ I -:;;.~ " ~ ro' N "-> -.:'.- 3 ",' "8 ~ - -- -:. ' ~ - - .. - " " O' ñ' 6: o o 2- Secuencias(#) :::I: ~ N ~0r-:t0 00 ~ g .. '. -o _ . ,; ' - '-'- ' ~, ' - '- - "' . '-"= ' ,: ' _ ":.. _ '" (3 N ~~ , ~ ' ~-r , ~. ~ ' ~ ~ ~ . > - ~ " ", ~ :I ~ : ~ ~ \ l ~( '8 ' O' Q: o' o o 2- ~ ... .. - 8 E' ¡ji""O 3 N ro q¡ ro o· ít 5:2, o' a W * G5 '" ~ h o" 9;' ,,~ ¡; ro 0 0 w o, '? g' ro -- r ' ~" CJ o " O' N_ ·o ~N Q* ~. ." ~¡¡¡ .~ Q. -o ~¡¡¡ .0 tt:f "' ~. ¡;; ,' _. ro\ ', :!. 5. 'o' R Xi o "O • o o ~. ~ , g- .- ce N . ~ ro 3 • o ro - . ~ • o ro o - -g o o~ ~~ ~ '§l, ¡¡¡ ~ " \ H ¡V { roo V ~ ~.~ .. N"O "O ro ro ~ g. g ~ a • G') o " O' Q: Q' o Q, ro g n o: o o o 43 Inhibición de las sinapsis tálamo→estriatales durante el inicio/realización de una secuencia de acciones La inhibición de las terminales axónicas por medio de optogenética es una técnica poco utilizada (Kim et al., 2017). Por lo tanto para evaluar esta técnica llevamos a cabo registros de célula completa y estimulación eléctrica de los axones tálamo→estriatales, en una preparación donde se conservan las fibras tálamo-estriatales(Arias-García et al., 2018; Smeal et al., 2007), esto en animales previamente tratados para expresar Arch 3.0 en el tálamo. La idea fue que, al aplicar luz verde sobre corrientes postsinápticas registradas en el estriado, esta produjera un decremento en la probabilidad de generar despolarizaciones postsinápticas en comparación a los ensayos sin luz verde (Figura 16L-M). La segunda manera de verificar la inhibición de axones usando Arch3.0 fue registrando in vivo en el estriado mientras se iluminaba a las fibras tálamo→estriatales, esto con el propósito de observar los cambios generados por la luz. Al realizar este experimento encontramos que la inhibición de las terminales tálamo→estriatales redujo la actividad en al menos el 37% de las unidades registradas. Adicionalmente se realizaron experimentos control, en los cuales solo se expresó eYFP, en los cuales no se observó una modulación prácticamente nula con la luz (Figura 16F). Como última prueba del funcionamiento de la inhibición de terminales axónicas, también analizamos los cambios registrados en el LFP, encontrando una modulación clara del LFP al inhibir las sinapsis del PF→DMS. Por otro lado, la inhibición de los VPs→DLS es más compleja ya que se observó un efecto menor que pudo ser debido a una circuitería más compleja (Figura 16I). Una posibilidad colateral del querer inhibir a los axones tálamo→estriatales es que al manipular los axones en el estriado afectáramos a los axones de paso hacia la corteza. Por este motivo se implantó a un animal transfectado con Arch3.0 en los VPs, una fibra óptica en el estriado y un electrodo de registro en la corteza somatosensorial; los registros de este último experimento no mostraron cambios ni en la actividad unitaria (Figura 16G) ni en el LFP (Figura 16I). 44 Figura 16. Verificación de la inhibición tálamo-estriatal. A) Esquema del experimento para registrar en el estriado mientras se inhiben las fibras tálamo→estriatales. B) Ejemplo de una unidad en el estriado que decremento su tasa de disparo cuando se le colocó pulsos de luz verde. La unidad mostrada tiene una frecuencia más alta que el promedio. C) Registro del total de unidades registradas en el estriado mientras se inhibió a las terminales PFs→DMS. D) Promedio de la actividad de las unidades presentes en C. E) Similar a C, pero registrando un animal que tiene la expresión de eYFP. F) Promedio de los registros en E. G) Similar a D, pero ahora en un animal infectado en los VPs con la fibra en el estriado y los electrodos en la corteza somatosensorial. H) Promedio de G. I) LFP registrado para todas las inhibiciones probadas en este estudio. J) Registro en rebanada para despolarizar al estriado por estim. eléctrica de aferentes tálamos→estriatales. K) Fotomicrografía de una célula registrada bajo el protocolo en J. L) Despolarizaciones evocadas por la estimulación eléctrica (ver J) en presencia o ausencia de luz verde. M) Cuantificación de los trazos presentados en L. Registro In vivo extracelu lar A AAvmo- A rchJ.O-eYFP B TVGluT2-Cre .;- , Estriado 50~N~ Estimulación O.3ms Arch3_0-eYFP eYFP e D LFP DLS(inhib ~ LFP DMS (inhibit PFs->DMS) I 0.5 mV .... .... .... ,. ........ .."'1'- LFP 51 (inhibit. VPs->DLS) ~~·NN~. __ ~'W. __ ~~.~~Y~~'-..~I~ ' ~ ., AAV 010- Arch3.0-eYFP VGluT2-Cre + Estimu lación Estriado E AAV DIO-eYFP VGluT2-Cre + Estriado Estimulación Inhibidos Act ivadas Neo Todas 4 J Tiempo (5) F Tiempo (5) 4% ,ti'! \96'91 -1 Tiempo(s) -'-'" ~ ;;----O - -';- ~ 6" ~ Tiempo(s) Ex vivo reqistros célula completa stim lación Registro Estimulación I z ldlJ el.e ctrica e, ~ SIr - '" ~ L + Luz estim . ~ ~e : .. ~s.ti~. • • ••••••• •• G -, H AAV 010 - Arch3.0-eYFP VGluT2-Cre + Estimulación 51 (Corteza) 4 6 Tiempo (s) Tiempo(s) ·1 -, K # Estimulación electrica 45 Inhibición optogenética de las terminales tálamo-estriatales durante el inicio de una secuencia de acciones Para evaluar la contribución de las sinapsis tálamo→estriatales primero se realizó una inhibición de estas justo antes de que los animales iniciaran la secuencia acciones. Los resultados de este experimento mostraron un incremento en la latencia para iniciar la secuencia, tanto al inhibir los circuitos PFs→DMS como VPs→DLS (Figura 17) (PFs→DMS latenciaon = 2.9 ± 0.7s versus PFs→DMS latenciaoff = 0.8 ± 0.1; VPs→DLS latenciaon = 2.7 ± 0.5 versus VPs→DLS latenciaoff = 0.9 ± 0.9, p<0.05, Wilcoxon test). Interesantemente la inhibición de los VPs→DLS incrementó significativamente el número de presiones (VPs→DLS presioneson = 5.4 ± 0.5 presiones versus VPs→DLS presionesoff = 3.8 ± 0.5, p<0.05, Wilcoxon test) sin presentar cambios en la duración de la secuencia. Figura 17. La inhibición de la sinapsis tálamo-estriatales retrasa el inicio de la secuencia de acciones A) Esquema de la inhibición. B) Cuantificación de la latencia cuando se inhibió las sinapsis PFs→DMS. A la derecha se muestra el cociente comparando el efecto de la luz a un día previo a la estimulación. C) Similar a B, pero para el circuito de VPs→DLS. D y E) Número de presiones. F y G) Duración de la secuencia. A Inhibición - inicio VGluT2-Cre + inyección talámica AVV DIO-Arch3.0IeYFP Fibras ~ ópticas Comedero f' ", / O . i Sensor IR LaÍ: : n~ Palanca La luz es activada cuando cruza el sensor B PFs-+ DMS Arch3.0 eYFP e VPs-+DLS Arch3.0 eYFP Latencia para iniciar eYFP ~: ~ J: ~ e o~ - 5, 4 --~ 4 w c~ (f) 2 ~ o~ 2 O off on off on 20 D (/)10 ~ 8- .~ 6 ~ 4 o. .. 2 O Número de presiones F Duración de la secuencia 20 (/) 015 "O 310 Ol ¡jj 5 O E (/)10 ~ 8 .~ 6 ~ 4 o. .. 2 O Número de presiones > ~ ~ * __________ 15. 2 - .-. b.~ -'2 ~ c-1 ~ O* off on off on :6 o G Duración de la secuencia ~:: / :;2 él :1 T ~5 ~ ~ 6~2 ~ ~ o o ~ off on i5 o .:::::;- 46 Inhibición de las terminales tálamo-estriatales durante la realización de una secuencia de acciones La inhibición previa de las sinapsis VPs→DLS produjo un aumento en el número de presiones, por lo que preguntamos si el inhibir estas sinapsis cuando el animal presiona por primera vez, tiene el mismo efecto que cuando inhibimos antes de que el animal presione. La inhibición durante la ejecución produjo un incremento en el número de presiones, sin alterar la duración de la secuencia (Figura 18E) (VPs→DLS presioneson = 5.1 ± 0.4 versus VPs→DLS presionesoff = 3.2 ± 0.3, p<0.05, Wilcoxon test), concluyendo que la participación de la sinapsis VPs→DLS produce un incremento en el número de acciones. Figura18. La inhibición de las sinapsis VPs→DLS genera un mayor número de acciones en la secuencia. A) Esquema de la inhibición. B) Cuantificación de la latencia para cuando se inhibió a las sinapsis PFs→DMS. A la derecha se muestra un cociente que compara el efecto de la luz contra un día previo a la estimulación. C) Similar a B, pero para la inhibición de las sinapsis VPs→DLS. D y E) Número de presiones. F y G) Cuantificación de la duración de la secuencia. A B PFs+DMS Arch3.0 eYFP e VPs+DLS Arch3.0 eYFP eYFP Inhibición - ejecución Latencia para iniciar eYFP i~o "ª ~ j ~ol off on off on ~ VGluT2-Cre + inyección talamica AW DIO-Arch3.0/eYFP D F ; bras ~ ópti cas () Comedero / Palanca (.í. ' La luz es activada cuando presiona por primera vez '" 1 o il1 8 o "00 6 i'! 4 c. .. 2 O F ~ 20 -g 15 ::l O> 1 O Ql rJ) 5 O Número de presiones ~ ::::=- .-------- '" ~ 8 - >10_ ~ ~ 0'6 6- ~ --z:: c_ 4 o g¡ - ~ 2- off on off on :o 0- Duración de la secuencia ,; 10 ~ ___ ~ 8 "'ro 6 2~ 4 f4 ó~ 2 :o O off on off on Latencia para iniciar ~ ( ~ ~ j j L off on off on E '" 10 Ql § 8 "00 6 i'! c. 4 .. 2 O G ~20 -g 15 ::l 0>10 Ql rJ) 5 O Número de presiones ~ /;;;; ~ ~ ~r: ~ ;,....-'A ___ c::::: 4 --______ o g¡ """"- - 2 off on off on ~ O Duración de la secuencia ~f: l- c:::: 4 ----- °2 2 ~ ~ o1:o off on off en 47 La inhibición de la vía PFs-DLS no evoca cambios durante el inicio / realización de secuencias de acciones Dado que existe la posibilidad de que cualquier manipulación de las fibras excitatorias sobre el estriado dorsolateral afecten la ejecución de una secuencia de acciones, preguntamos si la inhibición de las sinapsis PFs→DLS afectaba la ejecución de la secuencia de acciones. Los resultados de este experimento no mostraron efecto alguno (Figura19), por lo que pensamos que la modulación de las sinapsis VPs→DLS en la modulación del número de presiones es específica. Figura 19. La inhibición de la sinapsis PF→DLS no altera el inicio ni la ejecución de secuencias A) Esquema de la inhibición. B y C) Protocolo de inhibición antes e iniciada la secuencia de presiones. D) Fotos representativas de la infección y un track de una fibra óptica. E y F) Latencia para iniciar cuando se inhibió las sinapsis PFs→DLS en los protocolos B-C. G) Puntas de las fibras de los animales de este experimento. H) Número de presiones cuando se inhibió las sinapsis PFs→DLS en los protocolos B-C. A PFs+DLS VGluT2-Cre + inyección tálamica AAV DIO-Arch3.0-eYFP D PFs+DLS ~ _ . -\\ G:;:,\ .! ·' ! .... W 'I l' _1) -~;J _2 .• Bfe,;¡ma-- 05 f"(1"" Sitio de inyección Fibra óptica B E 8 S1 3 • InhibiciOn in icio • InhibiciOn ejecution R"=O.03 • p=O.570 1, _ f ! . _ O l-~~~~ O 2 4 6 8 # Presiones ( off) L Inh ibición VPs->DLS 3 "'1 ~ : o ~~ ; 06 ;~ .2 o • ! § : ' :0 O 2 4 6 8 # Presiones ( 011) 49 La inhibición de las sinapsis VPs→DLS en la primera sesión de entrenamiento no afecta el inicio / realización de secuencias de acciones Para evaluar la posibilidad de que la modulación ejercida por la inhibición de las sinapsis VP→DLS sobre la ejecución de la secuencia suceda solo sobre la realización de secuencias de presiones aprendidas o simplemente es un efecto motor y está sucediendo en cualquier momento independientemente de una secuencia aprendida decidimos correr un grupo de animales a los que les inhibimos las sinapsis VPs→DLS en el primer día de entrenamiento a realizar secuencias de acciones (Figura 21). En este experimento no se observó modulación sobre la realización de las secuencias sugiriendo que se necesita haber aprendido las secuencias para que el circuito VPs→DLS modifique el número de acciones del animal. Figura 21. La inhibición de los axones VPs→DLS temprano en el aprendizaje de secuencias de acciones, no modifica el número de acciones de una secuencia de acciones. A) Esquema del diseño de este experimento. B) Esquema de la inhibición. C) Protocolo de inhibición en la ejecución. D) Ubicación de las fibras ópticas de los animales de este experimento. E) Número de presiones que realiza el animal por secuencia en periodos de luz y no luz, en el box- plot se muestra el cociente del día de inhibición normalizado a un día sin manipulación. A VPs-+-DLS Inhibición temprana I nyecci6n Arch Prueba CRF ,j. FR8 [ -14 ---+ 1 11-1 2 (dias) (Sesiones en FR8) B , ~ ) ~ VGluT2-Cre + inyección tálamica AAV DIO-Arch3 .0-eYFP e Inhibición ejecución Fibras ~ comedero ópticas O Palanca {.í" , La luz se activa cuando el ratón presiona por primera vez o Arch3.0-eYFP E Número de presiones en la secuencia 10 8 • N • -"6 e o 6 ",o "' .e~4 ~ ~ §~2 o. " 2 B 2 O off on 50 La inhibición de las sinapsis PFs→DLS o VPs→DLS no modifica el desplazamiento horizontal de los animales en la prueba de campo abierto Debido a las técnicas empleadas y los circuitos manipulados era necesario saber si la inhibición de las vías tálamo-estriatales modificaban el desplazamiento de los animales. Para ello, se llevó a cabo la inhibición de las fibras tálamo-estriatales para ambos circuitos (PFs→DMS y VPs→DLS) en la prueba de campo abierto encontrando que la inhibición de estas vías no altero el desplazamiento horizontal de los animales (Figura 22). No obstante la inhibición de las fibras VPs→S1 si se modificó el desplazamiento horizontal (Mathis et al., 2017). Esto sugiere que las sinapsis tálamo→estriatales estudiadas tienen su modulación sobre el inicio y realización de secuencias de acciones no por efectos motores generales. Figura 22. La inhibición de las proyecciones tálamo-estriatales no afecta el desplazamiento horizontal de los animales en la prueba de campo abierto. A y B) Vista superior y seguimiento de la trayectoria de un animal mientras recibió inhibición optogenética. C y D) Gráfico de la distancia normalizada antes y después de la estimulación con luz para la proyección PFs-DMS y VPs→DLS respectivamente. E) Comparación de la distancia normalizada durante los 5 segundos de inhibición para los diferentes grupos. e ro "O .§ 2 ro E o c:: ro ·0 c:: ro A B on en OO~---- ~ ---- ~ -- ---- -5 O 5 10 Tiempo(s) DI' ¡ c:: 1 ro ·0 on !J _~5------ ~ 0------ 3- ----~ , 5 10 Tiempo(s) E ro "O 2 ro .~ ro E o c:: ro '-' c:: $ ti) O O * "' ~ .... $ $$ ~ - -- ---- . --_ ....... . PFs-DMS eYFP on PFs-DMS Arch3_0 on VPs-DLS eYFP on VPs-DLS Arch3_0 on VPs-S1 Arch 3.0 on 51 La eficacia sináptica en las sinapsis Pf→DMS se encuentran disminuida en un modelo de compulsiones - SAPAP3 (-/-) Los resultados hasta ahora mostrados sugieren fuertemente una contribución específica de las sinapsis del tálamo al estriado en condiciones normales al iniciar o realizar una secuencia de acciones. Dada la Hipótesis de que la generación de compulsiones pudiese deberse a una desregulación en la apropiada iniciación/realización de acciones(Bernardo Barahona-Correa et al., 2015) y que existen reportes sugiriendo que el tálamo podría estar alterado en sujetos con compulsiones (Smith et al., 2009, 2014). Existen modelos animales deficientes de proteínas de andamiaje de sinapsis excitatorias que mostraron conductas repetitivas, estos modelos animales suelen utilizarse para el estudio de compulsiones (Pappas et al., 2014), un ejemplo de este modelo es el animal deficiente de la proteína SAPAP3(-/-), este modelo mostró presentar alteraciones especificas sobre las sinapsis excitatorias que llegan al estriado (Wan et al., 2014), por ello nos dimos a la tarea de saber si las sinapsis tálamo→estriatales están alteradas cuando los sujetos presentan compulsiones. Para contestar esta pregunta tomamos ventaja de la existencia del modelo transgénico de compulsiones SAPAP3 (-/-) y caracterizamos la trasmisión sináptica en las conexiones sinápticas entre el PFs→DMS de estos ratones y un grupo de respectivos controles [como controles utilizamos a los hermanos de la misma camada SAPAP3 (+/+)]. Decidimos explorar la sinapsis PFs→DMS debido a que estas células son las que mostraron mayor modulación durante el inicio y fin de las secuencias de acciones (Figura 12,13 y S1). Para la activación selectiva de las sinapsis PFs→DMS se siguieron dos estrategias: A) En la primera se inyectó en el PF de animales SAPAP3 (-/-) y (+/+) un virus que expresa ChR2 en neuronas glutaminérgicas (AVV1-CamKII-ChR2-eYFP; Figura 23A). Dos semanas después de la expresión de ChR2 en los axones PFs→DMS se procedió a registrar neuronas estriatales en rebanadas coronales, esto con el fin de evocar corrientes postsinápticas provenientes del PFs sobre las neuronas del estriado. Los resultados de este experimento muestran que la proyección PFs→DMS de animales SAPAP3(-/-) se encuentra desreguladas presentando corrientes postsinápticas de menor amplitud que los controles (PF→DMSSapap3(-/-) =40±11pA versus PF→DMSWT =89±17pA, n=10 y 9 respectivamente, Prueba U de Mann Whitney p<0.05). También evaluamos si las latencias para evocar corrientes postsinápticas pudiesen variar entre grupos, pero no fueron diferentes (Figura 23D). Con la intención de valuar alteraciones presinápticas, otro 52 parámetro que se evaluó fue la relación de pulso pareado. La medición de pulso pareado (PPR) a varios intervalos (50, 100, 150 y 200 ms) mostraron que los animales SAPAP(-/-) presentan una mayor depresión de corto plazo (PPR100msSapap3(-/-) =0.34±0.06 versus PPR100msWT =0.62±0.07 y PPR200msSapap3(-/-) =0.60±0.06 versus PPR200msWT =0.92±0.08 p<0.05 n=6 y n=9 múltiple t test con corrección de p corrección usando el método Holm- Sidak) (Figura 23E-F). B) En una segunda estrategia para medir la conectividad entre el PFs→DMS en los animales SAPAP se realizó un experimento más específico para marcar las sinapsis PFs→DLS. En esta ocasión se inyectaron un virus para expresar la Cre recombinasa de manera retrograda desde el DMS virus y se inyectó un segundo virus en el PF que expresa ChR2 en una manera Cre-dependiente, de este modo la segunda inyección solo marcaria células que tomaron el Cre desde el DMS y que por tanto son tálamo-estriatales. Los resultados de este experimento a la fecha tienen una n preliminar, pero ya muestran la misma tendencia de los datos previamente mencionados (comparar Figura 23). Las sinapsis PFs→DMS de los animales SAPAP3 (-/-) muestran una disminución en la eficacia sináptica respecto de sus controles (PF→DMSSapap3(-/-) =54±7.1pA versus PF→DMSWT =95±9pA, n=10 ensayos por célula provenientes de 6 y 5 células respectivamente, Prueba U de Mann Whitney p<0.05). A aav1 ChR2Kmll-mCherry e * •. 0 ... ~ . 7"f, .8 .. •.. .. fK"\ + :. \ I // B 300 ;¡- I::J Sapap3 (-1-) ~ D WT Jo * "" ID . ~ ~ () ",,, E F , Interyalo entre pulsos 200ms ~ 100ms = 50ms :.v::::::;r , '" ~ 50m, ~ ~ ~ D 1.5 1.0 0.5 0.0 '" .s ti ID " 51- ~ ID ~ • g fj ~ SOpA -+- Sapap3 (-1-) -+- WT L 50ms D Sapap 3 (- j - ) D WT ~ J * 50 100 150 200 250 Intervalos entre pulsos de luz (ms) G aav1 ChR2 DIO-eYFP PF, H 1.8 -+- WT ..... SAPAP3 (-1-) .0 ~ '" ~ ~ 0.5 0.0 50 100 150 200 250 Intervalos entre pulsos de luz (ms) 53 Figura 23. Foto-estimulación de fibras tálamo-estriatales en animales Sapap3 Knockout. (Página anterior) A) Esquema del experimento expresando Chr2 en neuronas del PF por medio de CaMKII B) Ejemplo de dos células registradas en el DMS en fijación de voltaje que exhiben corrientes postsinápticas excitatorias (EPSC) en color rojo se muestra los animales que carecen de la proteína Sapap3 y en azul los animales control. C) Gráfico de los EPSC en animales control y animales carentes de la proteína. D) Latencias de respuesta a la foto-estimulación. E) ejemplos de EPSC con estimulación en diferentes intervalos. F) Gráfico de Pulsos pareados normalizados (PPR) en diferentes intervalos. G) Esquema del experimento expresando Chr2 en neuronas del PF por medio de inyección de Cre de manera retrograda desde el estriado y luego ChR2 Cre dependiente en el PF. H-I) Similar a C y F, pero con la estrategia en G Por último, también evaluamos la probabilidad de encontrar conexión ya que sabemos en de manera normal es muy alta (Figura 9). Esta medición muestra que existe una menor probabilidad de encontrar sinapsis funcionales en el circuito PF→DMS en animales compulsivos (SAPAP3 (-/-) =51% versus SAPAP (+/+) =82% n=31 y n=17 células ×2 Figura 24). Figura 24. Probabilidad de encontrar conexión funcional en la conexión PF→DMS en animales SAPAP. A) Proporción de células que mostraron EPSC en la población de los animales SAPAP3 (-/-) y sus respectivos controles SAPAP3 (+/+), en color purpura se muestran neuronas de proyección, en color azul interneuronas y en color blanco neuronas que no se logró clasificar. B) Ejemplo de 2 células registradas en fijación de corriente en color purpura se muestra una neurona de proyección y en color azul se muestra a una interneurona. A WT Sapap3 (-/-) ~ ~------ ~--------~ B SPN ~JJjJJ jj JI ji ~-------~ -J~ ______________ ~ 1 ~ OO ~ p = A ~ &Q ~ m ~ ' __ -r __ __ Interneurona 54 Discusión Los resultados de este trabajo, abordan el estudio de la proyección tálamo-estriatal y su contribución al inicio y ejecución de secuencias de movimientos. Primero, identificamos proyecciones preferentes del tálamo hacia el estriado dorsal (Figuras 5-9). Segundo, identificamos que las neuronas tálamo-estriatales presentan actividad correlacionada con el inicio y ejecución de secuencias de movimientos (Figuras 11-14). Tercero, identificamos que la inhibición transitoria de las proyecciones durante en el inicio y ejecución de secuencias de movimientos afecta la iniciación y en el caso particular de las sinapsis VPs→DLS afecto la ejecución (Figuras 17 y 18). Finalmente, en las sinapsis tálamo→estriatales están afectadas en un modelo animal de compulsiones (Figuras 23 y 24). Proyecciones preferentes del tálamo hacia el estriado dorsal La proyección tálamo-estriatal ha sido objeto de estudios desde roedores hasta primates (Elena Erro et al., 2002; Hunnicutt et al., 2016; Nanda et al., 2009; Pan et al., 2010; Sadikot and Rymar, 2009; Wall et al., 2013). Aunque ya existían reportes sugiriendo que la proyección del tálamo hacia el estriado puede ser específicas hacia sus compartimentos (Berendse and Groenewegen, 1990; Elena Erro et al., 2002; Hunnicutt et al., 2016). En el presente trabajo realizamos una cuantificación de células marcadas retrógradamente y axones tálamo→estriatales marcados anterogradamente corroborando que la región PFs proyecta preferentemente al estriado dorso medial (Bradfield and Balleine, 2017; Choi et al., 2018; Thorn and Graybiel, 2010) y mostramos por primera vez que los núcleos ventrales del tálamo inervan preferentemente al estriado dorsolateral. Dentro de los experimentos de marcaje retrogrado nuestras imágenes preliminares mostraron poblaciones de células marcadas en el PF en función del compartimento postsináptico, DLS vs. DMS (Figura 7), idea que previamente había sido resaltada por Elena Erro et al., (2002). Es ta idea fue claramente corroborada en un estudio reciente en el que se caracterizó a tres poblaciones de neuronas PF→Estriado, neuronas del PF que preferentemente inervan al estriado dorsomedial, dorsolateral y medial (Mandelbaum et al., 2019). UNAM – Dirección General de Bibliotecas Tesis Digitales Restricciones de uso DERECHOS RESERVADOS © PROHIBIDA SU REPRODUCCIÓN TOTAL O PARCIAL Todo el material contenido en esta tesis esta protegido por la Ley Federal del Derecho de Autor (LFDA) de los Estados Unidos Mexicanos (México). El uso de imágenes, fragmentos de videos, y demás material que sea objeto de protección de los derechos de autor, será exclusivamente para fines educativos e informativos y deberá citar la fuente donde la obtuvo mencionando el autor o autores. Cualquier uso distinto como el lucro, reproducción, edición o modificación, será perseguido y sancionado por el respectivo titular de los Derechos de Autor. 55 Como la presencia de marca anatómica de las proyecciones tálamo→estriatales no significaba que esta conexión fuese funcional realizamos una comparación de las corrientes postsinápticas evocadas en células del estriado dorsal por estimulación de los axones provenientes del tálamo (Figura 9) corroborando nuestros propios datos de anatomía, sugiriendo que las regiones PFs y VPs inervan preferentemente a los compartimentos dorsomedial y dorsolateral del estriado. En un esfuerzo por que la conectividad reforzada no fuese producto de interacciones poli-sinápaticas realizamos nuestros registros resultados de corrientes postsinápticas en presencia de TTX y 4AP, una metodología que se utiliza para argumentar la conexión monosináptica (Petreanu et al., 2009), en este caso entre las sinapsis del tálamo y del estriado. Actualmente se sugiere que las proyecciones talámicas al estriado presentan especificidad hacia un tipo celular. En el estriado existen dos tipos de neuronas de proyección y un número grande de interneuronas (Tepper et al., 2010). Hasta el momento los reportes acerca de si la proyección del PFs es específica hacia un tipo neuronal no es clara, aunque diversos estudios muestran que esta inervación sucede sobre neuronas de proyección y sobre interneuronas (Arias-García et al., 2018; Ellender et al., 2013; Wallace et al., 2017). Con respecto a los estudios que reportan que los PFs tienen una proyección preferencial hacia las interneuronas colinérgicas (Bradfield and Balleine, 2017; Ding et al., 2010; Guo et al., 2015; Matsumoto et al., 2001; Yamanaka et al., 2018), de nuestros solo observamos una tendencia de una mayor probabilidad de conexión de las sinapsis del PF sobre interneuronas en el DLS (Figura 9). Actualmente existe una Hipótesis en la cual se plantea que la información llega a varios tipos de interneuronas y juega un rol importante en cómo se procesa dicha la información (Assous et al., 2017). Con respecto a la proyección de los VPs al estriado, ya existían reportes de esta región dejando colaterales dentro del estriado (Elena Erro et al., 2002; Hunnicutt et al., 2016) (Figuras 6-8). Canónicamente se reconoce a la región VPs como una estructura primaria del tálamo, procesa información sensoriomotora y proyecta principalmente hacia la corteza somatosensorial (Hunnicutt et al., 2014). Este trabajo significa la primera demostración de la conexión monosináptica que los VPs dejan en el estriado [esto gracias a que la estimulación de axones conteniendo ChR2 se hizo en presencia de TTX y 4AP; (Petreanu et al., 2009)], y aun que no encontramos diferencias en su inervación a neuronas de proyección versus interneuronas, esta posibilidad debe ser abarcada en estudios posteriores. 56 Las neuronas tálamo-estriatales presentan actividad correlacionada con el inicio y ejecución de secuencias de movimientos Después de encontrar proyecciones preferentes del tálamo al estriado, identificamos la actividad de neuronas talámicas (Figura 11) y de neuronas tálamo→estriatales, con ayuda de la foto-identificación antidrómica (Figura 13)(Lima et al., 2009). Previamente existían reportes mostrando que las neuronas del PF responden a estímulos visuales o táctiles, y que pueden modularse durante la demanda de atención y son dependientes del estado del animal (Kimura et al., 2004; Matsumoto et al., 2001; Minamimoto et al., 2005). Dado que en el paradigma empleado en este estudio no se presentó ningún estímulo explícito para realizar la secuencias, esto enfatiza nuestros hallazgos respecto a que las neuronas del PFs presentan un incremento justo antes del inicio de una secuencia de palanqueos, cuando estas secuencias son autoiniciadas (Figura 11); cabe mencionar que existe la posibilidad de que el incremento observado pueda ser debido al movimiento del ratón (lo cual se discutirá más adelante). Interesantemente, también observamos que algunas de las neuronas tálamo-estriatales mostraron tener una correlación con la latencia para iniciar la secuencia (Figura 14A-D) o la realización de la misma (Figura 14E-H). La región PFs procesa estímulos multisensoriales (Matsumoto et al., 2001) y la región VPs estímulos vibro táctiles (Vazquez et al., 2013). De hecho, recientemente se publicó un artículo que también reportó que la manipulación del VPL al DLS modifica la ejecución de una secuencia de movimientos aprendidos además de mencionar que existe una somatotopia sensorial en el estriado activados por estos núcleos del tálamo (Hidalgo- balbuena et al.2019). Nuestra especulación supone que el estriado se provee de información sensorial proveniente del tálamo que puede permitir el adecuado inicio y ejecución de secuencias de acciones, pero habría que dilucidar porque esta retroalimentación parece ser específica en la realización de secuencias aprendidas (Figura 21). 57 La inhibición transitoria de las proyecciones tálamo→estriatales afecta la iniciación y en el caso particular de las sinapsis VPs→DLS afecta la realización La validación de la técnica de inhibición de axones en un sitio blanco sin afectar axones de paso se abordó por medio de la evaluación de registros en S1 mientras se inhibía en el estriado (Figura 16), y al correr experimentos conductuales de inhibición en S1 (Figura 20). De estos resultados podemos concluir que los efectos de inhibición son específicos del sitio blanco basándonos además de nuestros propios resultados en artículos previos haciendo esta misma pregunta (Otchy et al., 2015), y en el hecho de que específicamente Arch parece tener un mecanismo de acción a través de modular el pH localmente y no necesariamente a través de la hiperpolarización de axones (El-Gaby et al., 2016) . Por otro lado, la recién evidenciada posibilidad de que el calor debido a la intensidad de luz utilizada pueda per se generar inhibición se refuta ya que no pudimos documentar esto en nuestros experimentos, ni al registran en un animal que sólo expresó la proteína reportera eYFP (Figura 16) ni al evaluar un efecto de desplazamiento neto (Figura 22). La inhibición de los axones tálamo-estriatales con la proteína Arch3.0, mostró que ambos circuitos retrasan el inicio de la secuencia (Figura 17). Sugiriendo que en parte la actividad observada en el estriado mientras los animales inician/realizan secuencias de palanqueo (Jin et al., 2014) y el hecho de que la inhibición estriatal afecte la iniciación ejecución de estas mismas secuencias (Tecuapetla et al., 2016) pueda deberse a la excitación que proveen las sinapsis tálamo→estriatales al estriado. El mecanismo por el cual se da el retraso del inicio puede ser debido a una baja en tasa de disparo en las neuronas espinosas medianas (Figura 16). Nuestros resultados mostraron que la inhibición de las fibras tálamo-estriatales no muestran un cambio en el desplazamiento horizontal (Figura 22), sugiriendo que el efecto observado en el retraso durante el inicio no es motor. Especulamos que el “feedback” que el tálamo provee se establece mientras se aprende la realización de secuencias (Figura 21), quizás implicando cambios plásticos en el peso sináptico de las sinapsis tálamo→estriatales. La inhibición del circuito VPs→DLS modificó la ejecución de una secuencia de acciones (Figura 18). Existen reportes que sugieren que es principalmente el compartimento DLS es indispensable la para la ejecución de secuencias seriales (Yin, 2010) (Geddes et al., 2018; Tecuapetla et al., 2016), sin embargo apenas se empieza a 58 dilucidar si son sinapsis específicas las que son requeridas. Una posibilidad era que la desestabilización de cualquier entrada al DLS afectaría la realización de secuencias sin embargo la inhibición del circuito del PF→DLS, no mostró afectar la realización de secuencias (Figura 19), sugiriendo que el sistema necesita una entrada específica (sensorial primaria en este caso), para ejecutar de manera adecuada la secuencia. Dado que en los registros de campo cuando se inhibió a las sinápsis VPs→DLS se observó una menor modulación (Figura 16) y observamos una mayor proporción de interneuronas registradas en el DLS (Figura 9), suponemos un mecanismo a través de interneuronas que podrían estar modulando la actividad estriatal, mecanismo que aun tendría que ser dilucidado. Las sinapsis tálamo→estriatales están afectadas en un modelo animal de compulsiones En este trabajo mostramos que las sinapsis PF→DMS contribuyen al apropiado inicio de secuencias de acciones (Figura 21). Sabiendo que en patologías como la enfermedad de Parkinson el inicio de acciones voluntarias se ve afectado (Redgrave et al., 2010) y que de hecho la inhibición del circuito PF→estriado disminuye los síntomas por la lesión de dopaminérgica (Parker et al., 2016), especulamos que en la generación de compulsiones el circuito tálamo→estriado también pueda estar alterado (Bernardo Barahona-Correa et al., 2015). Para aclarar esta especulación, y dado que los experimentos que han abordado esta pregunta no son contundentes (Wan et al., 2014), preguntamos si las sinapsis PF→DMS estaban afectadas en un modelo genético del trastorno obsesivo compulsivo [TOC; modelo SAPAP3(-/-) (Welch et al., 2007)]. La evaluación de esta sinapsis mostró que, en animales compulsivos, respecto de sus controles, se observó una menor entrada sináptica por parte del PF al DMS (Figura 23). Al comparar nuestros experimento con la literatura previa, se diferencian en que nosotros sí encontramos afectada la sinapsis tálamo→estriatal y otros no (Wan et al., 2014). Esto se podría explicar debido a diferencias metodológicas, nosotros a diferencia de trabajos previos utilizamos el marcaje y activación selectivo de sinapsis tálamo-estriatales mientras que previamente se hizo de manera eléctrica lo cual podría activar poblaciones heterogéneas de fibras (Wan et al., 2014). No obstante, el hecho de activar terminales presinápticas por la expresión de ChR2 podría generar complicaciones, nos permite enfatizar que las sinapsis PF→DMS presenta una disminución en conectividad en animales compulsivos. 59 Conclusiones ➢ La conectividad de los núcleos talámicos de la región PFs y VPs presentan una inervación preferencial al estriado, los primeros al DMS y los últimos al DLS. ➢ La actividad de las neuronas tálamo-estriatales registradas en los PFs o VPs correlaciona con el inicio y ejecución de secuencias de movimientos. ➢ La inhibición optogenética de las vías tálamo-estriatales (PFs→DMS y VPs→DLS) retrasa el inicio de una secuencia. ➢ La inhibición de la vía VPs→DLS afectó la realización de una secuencia de acciones, incrementando el número de acciones. ➢ El circuito PF-DMS se encuentra regulado a la baja en animales de un modelo genético de compulsiones. El presente trabajo muestra es un primer acercamiento al estudio de la proyección tálamo- estriatal y su participación en la realización de secuencias, mostrando que las sinapsis PF→DMS y VPs→DLS son requeridas para la iniciación y de manera específica las sinapsis VPs—DLS son requeridas para la apropiada ejecución de una secuencia. UNAM – Dirección General de Bibliotecas Tesis Digitales Restricciones de uso DERECHOS RESERVADOS © PROHIBIDA SU REPRODUCCIÓN TOTAL O PARCIAL Todo el material contenido en esta tesis esta protegido por la Ley Federal del Derecho de Autor (LFDA) de los Estados Unidos Mexicanos (México). El uso de imágenes, fragmentos de videos, y demás material que sea objeto de protección de los derechos de autor, será exclusivamente para fines educativos e informativos y deberá citar la fuente donde la obtuvo mencionando el autor o autores. Cualquier uso distinto como el lucro, reproducción, edición o modificación, será perseguido y sancionado por el respectivo titular de los Derechos de Autor. 60 Referencias Aldridge, J.W., and Berridge, K.C. (1998). Coding of serial order by neostriatal neurons: a “natural action” approach to movement sequence. J. Neurosci. 18, 2777–2787. Alloway, K.D., Smith, J.B., and Watson, G.D.R. (2014). Thalamostriatal projections from the medial posterior and parafascicular nuclei have distinct topographic and physiologic properties. J. Neurophysiol. 111, 36–50. Arias-García, M.A., Tapia, D., Laville, J.A., Calderón, V.M., Ramiro-Cortés, Y., Bargas, J., and Galarraga, E. (2018). Functional comparison of corticostriatal and thalamostriatal postsynaptic responses in striatal neurons of the mouse. Brain Struct. Funct. 223, 1229–1253. Assous, M., Kaminer, J., Shah, F., Garg, A., Ko?s, T., and Tepper, J.M. (2017). Differential processing of thalamic information via distinct striatal interneuron circuits. Nat. Commun. 8, 15860. Berendse, H.W., and Groenewegen, H.J. (1990). Organization of the thalamostriatal projections in the rat, with special emphasis on the ventral striatum. J. Comp. Neurol. 299, 187–228. Bernardo Barahona-Corr??a, J., Camacho, M., Castro-Rodrigues, P., Costa, R., and Oliveira-Maia, A.J. (2015). From thought to action: How the interplay between neuroscience and phenomenology changed our understanding of obsessive-compulsive disorder. Front. Psychol. 6, 1–12. Bradfield, L.A., and Balleine, B.W. (2017). Thalamic Control of Dorsomedial Striatum Regulates Internal State to Guide Goal-Directed Action Selection. J. Neurosci. 37, 3721–3733. Bradfield, L.A., Bertran-Gonzalez, J., Chieng, B., and Balleine, B.W. (2013). The thalamo-striatal pathway and cholinergic control of goal-directed action: Interlacing new and existing learning in the striatum. Neuron 79, 10.1016/j.neuron.2013.04.039. Bradfield, L.A., Bertran-Gonzalez, J., Chieng, B., and Balleine, B.W. (2017). The Thalamostriatal Pathway and Cholinergic Control of Goal-Directed Action: Interlacing New with Existing Learning in the Striatum. Neuron 79, 153–166. Calabresi, P., and Filippo, M. Di (2010). Brain ’ s traffic lights. 466, 2066. Caparros-Lefebvre, D., Blond, S., Feltin, M.P., Pollak, P., and Benabid, A.L. (1999). Improvement of levodopa induced dyskinesias by thalamic deep brain stimulation is related to slight variation in electrode placement: Possible involvement of the centre median and parafascicularis complex. J. Neurol. Neurosurg. Psychiatry 67, 308–314. Choi, K., Holly, E.N., Davatolhagh, M.F., Beier, K.T., and Fuccillo, M. V. (2018). Integrated anatomical and physiological mapping of striatal afferent projections. Eur. J. Neurosci. 1–14. Dhawale, A.K., Smith, M.A., and Ölveczky, B.P. (2017). The Role of Variability in Motor Learning. Annu Rev Neurosci. 2017 May 10. Doi 10.1146/Annurev-Neuro-072116-031548. Ding, J.B., Guzman, J.N., Peterson, J.D., Goldberg, J.A., and Surmeier, D.J. (2010). Thalamic Gating of Corticostriatal Signaling by Cholinergic Interneurons. Neuron 67, 294–307. UNAM – Dirección General de Bibliotecas Tesis Digitales Restricciones de uso DERECHOS RESERVADOS © PROHIBIDA SU REPRODUCCIÓN TOTAL O PARCIAL Todo el material contenido en esta tesis esta protegido por la Ley Federal del Derecho de Autor (LFDA) de los Estados Unidos Mexicanos (México). El uso de imágenes, fragmentos de videos, y demás material que sea objeto de protección de los derechos de autor, será exclusivamente para fines educativos e informativos y deberá citar la fuente donde la obtuvo mencionando el autor o autores. Cualquier uso distinto como el lucro, reproducción, edición o modificación, será perseguido y sancionado por el respectivo titular de los Derechos de Autor. 61 Doya, K., Ishii, S., Pouget, A., and Rao, R.P.N. (2007). Bayesian Brain: Probabilistic Approaches to Neural Coding. El-Gaby, M., Zhang, Y., Wolf, K., Schwiening, C.J., Paulsen, O., and Shipton, O.A. (2016). Archaerhodopsin Selectively and Reversibly Silences Synaptic Transmission through Altered pH. Cell Rep. 16, 2259–2268. Elena Erro, M., Lanciego, J.L., and Giménez-Amaya, J.M. (2002). Re-examination of the thalamostriatal projections in the rat with retrograde tracers. Neurosci. Res. 42, 45–55. Ellender, T.J., Harwood, J., Kosillo, P., Capogna, M., and Bolam, J.P. (2013a). Heterogeneous properties of central lateral and parafascicular thalamic synapses in the striatum. 1, 257–272. Ellender, T.J., Harwood, J., Kosillo, P., Capogna, M., and Bolam, J.P. (2013b). Heterogeneous properties of central lateral and parafascicular thalamic synapses in the striatum. J. Physiol. 591, 257–272. Erro, E., Lanciego, J.L., and Gim ńez-Amaya, J.M. (1999). Relationships between thalamostriatal neurons and pedunculopontine projections to the thalamus: A neuroanatomical tract-tracing study in the rat. Exp. Brain Res. 127, 162–170. Fremeau, R.T., Troyer, M.D., Pahner, I., Nygaard, G.O., Tran, C.H., Reimer, R.J., Bellocchio, E.E., Fortin, D., Storm-Mathisen, J., and Edwards, R.H. (2001). The expression of vesicular glutamate transporters defines two classes of excitatory synapse. Neuron 31, 247–260. Fujii, N., and Graybiel, A.M. (2003). Representation of Action Sequence Boundaries by Macaque Prefrontal Cortical Neurons. Science (80-. ). 301, 1246 LP – 1249. Geddes, C.E., Li, H., and Jin, X. (2018). Optogenetic Editing Reveals the Hierarchical Organization of Learned Action Sequences. Cell 174, 32-43.e15. Gillan, C.M., Papmeyer, M., Morein-Zamir, S., Sahakian, B.J., Fineberg, N.A., Robbins, T.W., and De Wit, S. (2011). Disruption in the balance between goal-directed behavior and habit learning in obsessive-compulsive disorder. Am. J. Psychiatry 168, 718–726. Grider, M.H., Chen, Q., and David Shine, H. (2006). Semi-automated quantification of axonal densities in labeled CNS tissue. J. Neurosci. Methods 155, 172–179. Guo, Q., Wang, D., He, X., Feng, Q., Lin, R., Xu, F., Fu, L., and Luo, M. (2015). Whole-Brain Mapping of Inputs to Projection Neurons and Cholinergic Interneurons in the Dorsal Striatum. PLoS One 10, e0123381. Guo, Z. V., Inagaki, H.K., Daie, K., Druckmann, S., Gerfen, C.R., and Svoboda, K. (2017). Maintenance of persistent activity in a frontal thalamocortical loop. Nature 545, 181–186. Hidalgo-balbuena, A.E., Luma, A.Y., Pimentel-farfan, A.K., Peña-rangel, T., and Rueda-orozco, P.E. motor habits. Nat. Commun. 1–15. Hikosaka, O., Sakai, K., Lu, X., Nakahara, H., Rand, M.K., Nakamura, K., Miyachi, S., and Doya, K. (1999). Parallel neural networks for learning sequential procedures. Trends Neurosci. 22, 464–471. Hunnicutt, B.J., Long, B.R., Kusefoglu, D., Gertz, K.J., Zhong, H., and Mao, T. (2014). r e so u r c e A 62 comprehensive thalamocortical projection map at the mesoscopic level. Nat. Publ. Gr. 17, 1276–1285. Hunnicutt, B.J., Jongbloets, B.C., Birdsong, W.T., Gertz, K.J., Zhong, H., and Mao, T. (2016). A comprehensive excitatory input map of the striatum reveals novel functional organization. Elife 5, e19103. Jin, X., and Costa, R.M. (2010). Start/stop signals emerge in nigrostriatal circuits during sequence learning. Nature 466, 457–462. Jin, X., and Costa, R.M. (2015). ScienceDirect Shaping action sequences in basal ganglia circuits. Curr. Opin. Neurobiol. 33, 188–196. Jin, X., Tecuapetla, F., and Costa, R.M. (2014a). Basal ganglia subcircuits distinctively encode the parsing and concatenation of action sequences. Nat. Neurosci. 17, 423–430. Jin, X., Tecuapetla, F., and Costa, R.M. (2014b). Basal ganglia subcircuits distinctively encode the parsing and concatenation of action sequences. Nat. Neurosci. 17, 423–430. Jin, X., Tecuapetla, F., and Costa, R.M. (2014c). Basal ganglia subcircuits distinctively encode the parsing and concatenation of action sequences. Nat Neurosci 17, 423–430. Kato, S., Fukabori, R., Nishizawa, K., Okada, K., Yoshioka, N., Sugawara, M., Maejima, Y., Shimomura, K., Okamoto, M., Eifuku, S., et al. (2018). Action Selection and Flexible Switching Controlled by the Intralaminar Thalamic Neurons. Cell Rep. 22, 2455–2468. Kawai, R., Markman, T., Poddar, R., Ko, R., Fantana, A.L., Dhawale, A.K., Kampff, A.R., and Ölveczky, B.P. (2015). Motor Cortex Is Required for Learning but Not for Executing a Motor Skill. Neuron 86, 800–812. Kim, C.K., Adhikari, A., and Deisseroth, K. (2017). Integration of optogenetics with complementary methodologies in systems neuroscience. Nat Rev Neurosci 18, 222–235. Kimura, M., Minamimoto, T., Matsumoto, N., and Hori, Y. (2004). Monitoring and switching of cortico-basal ganglia loop functions by the thalamo-striatal system. Neurosci. Res. 48, 335–360. Kupferschmidt, D.A., Juczewski, K., Cui, G., Johnson, K.A., and Lovinger, D.M. (2017). Parallel, but Dissociable, Processing in Discrete Corticostriatal Inputs Encodes Skill Learning. Neuron 96, 476-489.e5. Lacey, C.J., Bolam, J.P., and Magill, P.J. (2007). Novel and Distinct Operational Principles of Intralaminar Thalamic Neurons and Their Striatal Projections. J. Neurosci. 27, 4374–4384. Lashley, K. (1952). The problem of serial order in behavior. Jeffress LA (Ed), Cereb. Mech. Behav. New York Wiley 112–131. Lima, S.Q., Hromádka, T., Znamenskiy, P., and Zador, A.M. (2009). PINP: A New Method of Tagging Neuronal Populations for Identification during In Vivo Electrophysiological Recording. PLoS One 4, e6099. Mandelbaum, G., Taranda, J., Haynes, T.M., Robertson, K., Osten, P., Sabatini, B.L., Mandelbaum, G., Taranda, J., Haynes, T.M., Hochbaum, D.R., et al. (2019). Distinct Cortical-Thalamic-Striatal Circuits through the Parafascicular Nucleus Article Distinct Cortical-Thalamic-Striatal Circuits through the Parafascicular Nucleus. 1–17. 63 Mathis, M.W., Mathis, A., and Uchida, N. (2017). Somatosensory Cortex Plays an Essential Role in Forelimb Motor Adaptation in Mice. Neuron 93, 1493-1503.e6. Matsumoto, N., Minamimoto, T., Graybiel, a M., and Kimura, M. (2001). Neurons in the thalamic CM-Pf complex supply striatal neurons with information about behaviorally significant sensory events. J. Neurophysiol. 85, 960–976. Melief, E.J., McKinley, J.W., Lam, J.Y., Whiteley, N.M., Gibson, A.W., Neumaier, J.F., Henschen, C.W., Palmiter, R.D., Bamford, N.S., and Darvas, M. (2018). Loss of glutamate signaling from the thalamus to dorsal striatum impairs motor function and slows the execution of learned behaviors. Npj Park. Dis. 4, 23. Miller, G.A. (1956). Miller GA Magical Seven Psych Review 1955.pdf. 101, 343–352. Minamimoto, T., and Kimura, M. (2002). Participation of the thalamic CM-Pf complex in attentional orienting. J. Neurophysiol. 87, 3090–3101. Minamimoto, T., Hori, Y., and Kimura, M. (2005). Complementary Processes to Response Bias in the Centromedian Nucleus of the Thalamus. Science (80-. ). 308, 1798–1801. Nanda, B., Galvan, A., Smith, Y., and Wichmann, T. (2009). Effects of Stimulation of the Centromedian Nucleus of the Thalamus on the Activity of Striatal Cells in Awake Rhesus Monkeys. Eur. J. Neurosci. 29, 588– 598. Otchy, T.M., Wolff, S.B.E., Rhee, J.Y., Pehlevan, C., Kawai, R., Kempf, A., Gobes, S.M.H., and Ölveczky, B.P. (2015). Acute off-target effects of neural circuit manipulations. Nature 528, 358–363. Pan, W.X., Mao, T., and Dudman, J.T. (2010). Inputs to the Dorsal Striatum of the Mouse Reflect the Parallel Circuit Architecture of the Forebrain. Front. Neuroanat. 4, 1–14. Pappas, S.S., Leventhal, D.K., Albin, R.L., and Dauer, W.T. (2014). Mouse Models of Neurodevelopmental Disease of the Basal Ganglia and Associated Circuits (Elsevier Inc.). Parker, P.R.L., Lalive, A.L., and Kreitzer, A.C. (2016a). Pathway-Specific Remodeling of Thalamostriatal Synapses in Parkinsonian Mice. Neuron 1–7. Parker, P.R.L., Lalive, A.L., and Kreitzer, A.C. (2016b). Pathway-Specific Remodeling of Thalamostriatal Synapses in Parkinsonian Mice. Neuron 89, 734–740. Petreanu, L., Mao, T., Sternson, S.M., and Svoboda, K. (2009). The subcellular organization of neocortical excitatory connections. Nature 457, 1142–1145. Redgrave, P., Rodriguez, M., Smith, Y., Rodriguez-Oroz, M.C., Lehericy, S., Bergman, H., Agid, Y., DeLong, M.R., and Obeso, J. a (2010a). Goal-directed and habitual control in the basal ganglia: implications for Parkinson’s disease. Nat. Rev. Neurosci. 11, 760–772. Redgrave, P., Rodriguez, M., Smith, Y., Rodriguez-Oroz, M.C., Lehericy, S., Bergman, H., Agid, Y., DeLong, M.R., and Obeso, J.A. (2010b). Goal-directed and habitual control in the basal ganglia: implications for Parkinson’s disease. Nat. Rev. Neurosci. 11, 760–772. 64 Sadikot, A.F., and Rymar, V. V. (2009). The primate centromedian-parafascicular complex: Anatomical organization with a note on neuromodulation. Brain Res. Bull. 78, 122–130. Schultz, W., and Romo, R. (1992). Role of primate basal ganglia and frontal cortex in the internal generation of movements I . Preparatory activity in the anterior striatum. 363–384. Smeal, R.M., Gaspar, R.C., Keefe, K.A., and Wilcox, K.S. (2007). A rat brain slice preparation for characterizing both thalamostriatal and corticostriatal afferents. J. Neurosci. Methods 159, 224–235. Smith, J.B., Mowery, T.M., and Alloway, K.D. (2012). Thalamic POm projections to the dorsolateral striatum of rats: potential pathway for mediating stimulus-response associations for sensorimotor habits. J. Neurophysiol. 108, 160–174. Smith, Y., Raju, D. V., Pare, J.F., and Sidibe, M. (2004). The thalamostriatal system: A highly specific network of the basal ganglia circuitry. Trends Neurosci. 27, 520–527. Smith, Y., Raju, D., Nanda, B., Pare, J.-F., Galvan, A., and Wichmann, T. (2009). The thalamostriatal systems: Anatomical and functional organization in normal and parkinsonian states. Brain Res. Bull. 78, 60–68. Smith, Y., Galvan, A., Ellender, T.J., Doig, N., Villalba, R.M., Huerta-, I., Wichmann, T., and Bolam, J.P. (2014). The thalamostriatal system in normal and diseased states. 8, 1–18. Tanji, J. (2001). Sequential Organization of Multiple Movements: Involvement of Cortical Motor Areas. Annu. Rev. Neurosci. 24, 631–651. Tecuapetla, F., Matias, S., Dugue, G.P., Mainen, Z.F., and Costa, R.M. (2014). Balanced activity in basal ganglia projection pathways is critical for contraversive movements. Nat. Commun. 5, 1–10. Tecuapetla, F., Jin, X., Lima, S.Q., Costa, R.M., Tecuapetla, F., Jin, X., Lima, S.Q., and Costa, R.M. (2016a). Complementary Contributions of Striatal Projection Pathways to Action Initiation and Execution Article Complementary Contributions of Striatal Projection Pathways to Action Initiation and Execution. Cell 166, 703– 715. Tecuapetla, F., Jin, X., Lima, S.Q., and Costa, R.M. (2016b). Complementary Contributions of Striatal Projection Pathways to Action Initiation and Execution. Cell 166, 703–715. Tepper, J.M., Tecuapetla, F., Koós, T., and Ibáñez-Sandoval, O. (2010). Heterogeneity and diversity of striatal GABAergic interneurons. Front. Neuroanat. 4, 150. Tepper, J.M., Koós, T., Ibanez-Sandoval, O., Tecuapetla, F., Faust, T.W., and Assous, M. (2018). Heterogeneity and Diversity of Striatal GABAergic Interneurons: Update 2018. Front. Neuroanat. 12, 1–14. Thorn, C. a., and Graybiel, A.M. (2010). Pausing to regroup: Thalamic gating of cortico-basal ganglia networks. Neuron 67, 175–178. Vandermaelen, C.P., and Kitai, S.T. (1980a). Intracellular analysis of synaptic potentials in rat neostriatum following stimulation of the cerebral cortex, thalamus, and substantia nigra. Brain Res. Bull. 5, 725–733. Vandermaelen, C.P., and Kitai, S.T. (1980b). Intracellular analysis of synaptic potentials in rat neostriatum 65 following stimulation of the cerebral cortex, thalamus, and substantia nigra. Brain Res. Bull. 5, 725–733. Vazquez, Y., Salinas, E., and Romo, R. (2013). Transformation of the neural code for tactile detection from thalamus to cortex. Proc. Natl. Acad. Sci. 110, E2635–E2644. Wall, N.R., De La Parra, M., Callaway, E.M., and Kreitzer, A.C. (2013). Differential Innervation of Direct- and Indirect-Pathway Striatal Projection Neurons. Neuron 79, 347–360. Wallace, M.L., Saunders, A., Huang, K.W., Philson, A.C., Goldman, M., Macosko, E.Z., McCarroll, S.A., and Sabatini, B.L. (2017). Genetically Distinct Parallel Pathways in the Entopeduncular Nucleus for Limbic and Sensorimotor Output of the Basal Ganglia. Neuron. Wan, Y., Ade, K.K., Caffall, Z., Ilcim Ozlu, M., Eroglu, C., Feng, G., and Calakos, N. (2014). Circuit-selective striatal synaptic dysfunction in the sapap3 knockout mouse model of obsessive-compulsive disorder. Biol. Psychiatry 75, 623–630. Welch, J.M., Lu, J., Rodriguiz, R.M., Trotta, N.C., Peca, J., Ding, J.-D., Feliciano, C., Chen, M., Adams, J.P., Luo, J., et al. (2007). Cortico-striatal synaptic defects and OCD-like behaviours in Sapap3-mutant mice. Nature 448, 894–900. Wu, Y.-W., Kim, J.-I., Tawfik, V.L., Lalchandani, R.R., Scherrer, G., and Ding, J.B. (2015). Input- and Cell- Type-Specific Endocannabinoid-Dependent LTD in the Striatum. Cell Rep. 10, 75–87. Yamanaka, K., Hori, Y., Minamimoto, T., Yamada, H., Matsumoto, N., Enomoto, K., Aosaki, T., Graybiel, A.M., and Kimura, M. (2018). Roles of centromedian parafascicular nuclei of thalamus and cholinergic interneurons in the dorsal striatum in associative learning of environmental events. J. Neural Transm. 125, 501–513. Yin, H.H. (2010). The Sensorimotor Striatum Is Necessary for Serial Order Learning. J. Neurosci. 30, 14719– 14723. 66 Material Suplementario Figura S1. Neuronas tálamo-estriatales incrementan su tasa de respuesta al inicio y ejecución de secuencias de movimientos. A, ejemplo de neurona que muestra una regresión con el número de presiones alineada al final de la secuencia. B, regresión lineal del ejemplo en A. C, porcentaje de las unidades que hicieron regresión con las presiones alineados al final de la secuencia. D, gráfico de pastel con el promedio del porcentaje de unidades que presentaron regresión. E, mapa de la actividad de las unidades PFs-DMS alineadas a cada presión. F, similar a E, pero mostrando las unidades para VPs-DLS.G, gráfico de pastel clasificando a las unidades tálamo-estriatales según su pico de actividad. A Spikes sec e 00 • Presiones • Espigas ,. ... ...,,...,.,,,., 8 B D ~40 ~ ..IiI _ ~ o I + beta (PFs~DMS) + beta (VPs-DLS) Espigas/final-presiones -"! 81 R=-O.88 ~: lli 00 • \J,) 2 • ~ 2 4 6 8 Presiones(#) (PI.Str) 0 60 1 ' "# . ~ o -2 o bet¡ 'F!\~DMS) - beta ( V Ps~ DLS) (VPs ·Str) E 00 • ro 10 "O "§ 20 " F Presión 1 Press 1 2 Time (sec) ejecución final PFs -DMS Presión 2 penúltma última -1 TIme (sec) VPs-DLS Presión 2 penúltima última TIme (sec) G PFs .DMS VPs.DLS % inicio % ejecución % final % otras categorias UNAM – Dirección General de Bibliotecas Tesis Digitales Restricciones de uso DERECHOS RESERVADOS © PROHIBIDA SU REPRODUCCIÓN TOTAL O PARCIAL Todo el material contenido en esta tesis esta protegido por la Ley Federal del Derecho de Autor (LFDA) de los Estados Unidos Mexicanos (México). El uso de imágenes, fragmentos de videos, y demás material que sea objeto de protección de los derechos de autor, será exclusivamente para fines educativos e informativos y deberá citar la fuente donde la obtuvo mencionando el autor o autores. Cualquier uso distinto como el lucro, reproducción, edición o modificación, será perseguido y sancionado por el respectivo titular de los Derechos de Autor. 67 Figura S2. Tracks de manipulaciones optogenéticas. A, esquema del experimento. B, sitios de las puntas de las fibras y en la parte derecha se muestra un ejemplo para PFs-DMS. C, similar a B, solo para el circuito VPs-DLS. A VGluT2-Cre + inyección tálamica AAV DIO-Arch3.0-eYFP B ~~~(PFs , DMS) e (VPs-DLS) snio de inyección Fibra óptica Sitio de inyección Fibra óptica Article The Thalamostriatal Projections Contribute to the Initiation andExecution of a Sequence ofMovements Highlights d Parafascicular (PFs) and ventroposterior (VPs) thalamostriatal neurons were recorded d PFs and VPs thalamostriatal neuron activity correlates with initiation and execution d Inhibition of PFs or VPs thalamostriatal projections delays initiation d Inhibition of VPs thalamostriatal projection increases actions inside a sequence Authors Edgar Dı́az-Hernández, Rubén Contreras-López, Asai Sánchez-Fuentes, Luis Rodrı́guez-Sibrı́an, Josué O. Ramı́rez-Jarquı́n, Fatuel Tecuapetla Correspondence fatuel@ifc.unam.mx In Brief Diaz-Hernández et al. performed electrophysiological recordings of antidromically photo-identified thalamostriatal neurons and optogenetic inhibition of the thalamostriatal terminals to probe their contribution to the smooth initiation and execution of a sequence of movements. Their data support a model in which different thalamostriatal projections provide differential control of distinct phases of initiating/executing a sequence of movements. Dı́az-Hernández et al., 2018, Neuron 100, 739–752 November 7, 2018 ª 2018 Elsevier Inc. https://doi.org/10.1016/j.neuron.2018.09.052 Neuron Article The Thalamostriatal Projections Contribute to the Initiation and Execution of a Sequence of Movements Edgar Dı́az-Hernández,1 Rubén Contreras-López,1 Asai Sánchez-Fuentes,1 Luis Rodrı́guez-Sibrı́an,1 Josué O. Ramı́rez-Jarquı́n,1 and Fatuel Tecuapetla1,2,* 1Instituto de Fisiologı́a Celular, Universidad Nacional Autónoma de México, Ciudad Universitaria, Circuito exterior s/n, 04510 Ciudad de México, CDMX, Mexico 2Lead Contact *Correspondence: fatuel@ifc.unam.mx https://doi.org/10.1016/j.neuron.2018.09.052 SUMMARY One of the main inputs driving striatal activity is the thalamostriatal projection. While the hypothesis postulating that the different thalamostriatal projec- tions contribute differentially to shape the functions of the striatum is largely accepted, existing technical limitations have hampered efforts to prove it. Here, through the use of electrophysiological recordings of antidromically photo-identified thalamostriatal neurons and the optogenetic inhibition of thalamos- triatal terminals, we identify that the thalamostriatal projections from the parafascicular and the ventro- posterior regions of the thalamus contribute to the smooth initiation and the appropriate execution of a sequence of movements. Our results support a model in which both thalamostriatal projections have specific contributions to the initiation and execution of sequences, highlighting the specific contribution of the ventroposterior thalamostriatal connection for the repetition of actions. INTRODUCTION Achieving the proper initiation and execution of movement se- quences is necessary for the survival of organisms. The basal ganglia-thalamocortical loops have been shown to be essential for the initiation of movement (Romo and Schultz, 1992) or the smooth initiation and execution of sequences of movements (Haber et al., 2012; Jin et al., 2014; Tecuapetla et al., 2016). Two main glutamatergic inputs to the striatum, the largest nucleus of the basal ganglia, have been proposed to drive the striatal activity for the initiation and execution of actions: the corticostriatal and the thalamostriatal inputs (Wilson, 1989). While a vast amount of research has documented the contribution of the corticostriatal projections (Hikosaka et al., 1999; Kupferschmidt et al., 2017; Rothwell et al., 2015; ZhengandWilson, 2002), the thalamostriatal projection has received less attention, with functional studies mainly focusingon the thalamo/ventral striatal projection (Chris- toffel et al., 2015; Do-Monte et al., 2017; Matsumoto et al., 2001; Zhu et al., 2016) and few studies documenting the necessity of the thalamo/dorsal striatal projections to perform appropriate movements or actions (Bradfield et al., 2013; Parker et al., 2016). Recently, the notion that the motor cortex (Kawai et al., 2015) or the corticostriatal terminals (Kupferschmidt et al., 2017) act as the main driver for the striatal activity to allow the perfor- mance of a sequence has been challenged (Kawai et al., 2015). This has opened the question of whether the thalamostriatal pro- jectionmaybemore relevant thanpreviously thought in driving the dorsal striatum activity necessary to perform a sequence of ac- tions or a learned motor skill (Jin et al., 2014; Rueda-Orozco and Robbe, 2015). The dorsal striatum in rodent contains the dorso- medial (DMS) and the dorsolateral striatal (DLS) compartments, which are recognized as part of the cortico-basal ganglia associa- tive and sensorimotor loops (Yin and Knowlton, 2006). Since the dorsal striatal activity has been shown tobenecessary for the initi- ation and execution of learned sequences of actions (Tecuapetla et al., 2016), and it has been proposed that specific nuclei of the thalamus project to distinct areas of the striatum (Hunnicutt et al., 2016; Pan et al., 2010; Yamanaka et al., 2017) with specific contributions to the performance of learned actions (Ito andDoya, 2015; Yin and Knowlton, 2006), the general hypothesis of this study is that different thalamostriatal projections have differential contributions to the different phases of a sequence ofmovements (the initiation versus the execution), with neurons of the parafas- cicular complex contributing mainly to the initiation and thala- mostriatal neurons from the sensorimotor thalamostriatal loops contributing preferentially to the execution. To evaluate this hypothesis, we implemented electrophysio- logical recordings of thalamostriatal neurons and performed the optogenetic inhibition of the thalamostriatal terminals while animals initiated or executed a sequence of movements. The findings presented here support a model in which the thalamos- triatal projections from the parafascicular and the ventroposte- rior region of the thalamus both contribute to the initiation, but only the latter contributes to the execution of learned sequences of movements. RESULTS To identify the thalamostriatal contribution to the initiation and execution of a sequence of movements, we conducted three Neuron 100, 739–752, November 7, 2018 ª 2018 Elsevier Inc. 739 general experiments. First, we identified regions of the thalamus containing neurons innervating the dorsal striatum compart- ments (dorsomedial versus dorsolateral); second, we recorded the activity of thalamic and photo-identified thalamostriatal neu- rons while animals initiated and executed a sequence of move- ments; and third, we optogenetically inhibited the thalamostriatal terminals either before the initiation or during the execution of a sequence of movements. The Parafascicular and the Ventroposterior Regions of the Thalamus Innervate the Dorsomedial and the Dorsolateral Striatum Since one of the main goals of this study was to identify whether the thalamostriatal neurons showed activity related to the initiation/execution of a sequence of movements, we aimed to identify thalamic regions that innervate the dorsal striatum compartments (DMS versus DLS). To this purpose, we expressed eYFP in thalamic neurons and their striatal axonal projections (Figure 1). To achieve the specific expres- sion of eYFP in thalamic neurons, we used a mouse line previ- ously reported to express Cre recombinase upon the presence of the vesicular glutamate transporter type 2 (VGluT2Cre mice), a molecular marker identified to be specific for thalamic neu- rons (Raju and Smith, 2005; Wu et al., 2015). Using these mice, we injected an adeno-associated virus (AAV-DIO-eYFP) that has been modified to express eYFP only in the cells con- taining the Cre recombinase (Atasoy et al., 2008) into several thalamic regions, centering our injections on brain coordinates for the parafascicular (Pf), the centromedial (CM), and the ven- troposteriomedial (VPM) thalamic nuclei. In accordance to pre- vious studies (Hunnicutt et al., 2016; Pan et al., 2010; Parker et al., 2016; Smith et al., 2014), we observed that these AAV injections labeled cells in the parafascicular region (PFs: paraf- ascicular, posterior, and the centrolateral thalamic nucleus; Figures 1A1, 1A2, and S1D), the centromedial region (CMs: centromedial, anterior medial, ventrolateral, paraventricular; data not shown), and the ventroposterior region (VPs: ventro- posteromedial, ventromedial, and ventroposterolateral nucleus; Figures 1B1, 1B2, and S1D), with different patterns of striatal projections, respectively (Figures 1, panels 3 and 4, and S1D). Following this screening, we decided to use only AAV in- jections into the Pf or the VPM to target the PFs or the VPs thalamic regions in order to minimize the overlap of the regions of transfected cells (distance between the injections sites Pf versus the VPM 1.2 mm, radius of eYFP-expressing cells from the injection site 0.5 mm; Figure S1C). Remarkably, the quantification of the striatal fibers when expressing eYFP in the PFs showed a small but preferential distribution into the DMS over the DLS (55.7% and 44.3%, respectively; Figure 1C, right panels; n = 5 animals; PFs/DLS versus PFs/DMS, p < 0.05; c2 test; and Figure S1D, upper panel). The opposite was observed when expressing eYFP into the VPs of the thalamus (58.8% of the VPs fibers into the DLS and 41.3% in the DMS; Figure 1C, right panel; n = 5 animals; DLS versus DMS, p < 0.05, c2 test; Figure S1D, lower panel). To verify these preferential innervations, we performed retro- grade labeling of thalamostriatal neurons by injecting retro- beads into either the DMS or the DLS (Figures 1D–1F). Cell counting of the retrogradely labeled cells in the PFs and the VPs was consistent with the AAV injections. The PFs showed more cells labeled when injecting retrobeads in the DMS, and the VPs showed more cells when injecting retrobeads in the DLS (percentage of cells with retrobeads in the PFs: PFs/DMS = 78.5%, PFs/DLS = 60.3%; and in the VPs: VPs/DLS = 39.7%, VPs/DMS = 21.5%, n = 4 animals for DMS and 5 animals for DLS injections; p < 0.05, c2 test; Fig- ure 1F, right panels). Importantly, by performing dual retro- beads injections (two colors) into the same striatum, it was estimated that only the 12% and 2% of the PFs or VPs cells project to both the DMS and the DLS, respectively (Figure S2). Furthermore, to ascertain whether this anatomical evidence could be translated into functional connectivity, we expressed the light-sensitive protein channelrhodopsin (ChR2) (Boyden et al., 2005) in either the PFs or the VPs. Taking advantage of the fact that ChR2 was expressed along the thalamostriatal axonal innervation, we performed ex vivo whole-cell recordings of striatal neurons while activating the thalamostriatal terminals in the striatum (Figures 1G–1J). In these experiments, we guar- anteed the recording of monosynaptic thalamostriatal connec- tions by recording in the presence of 4-AP+TTX (Petreanu et al., 2009) and accounted for the variability of ChR2 expression by performing pairs of recordings per slice (one in the DMS and one in DLS per slice). Through these experiments, we observed that the activation by blue light (1 ms) of the thalamostriatal ter- minals from either the PFs or the VPs onto cells of the striatal compartments also reflected a preferential innervation of the PFs onto the DMS and the VPs onto the DLS, as evidenced by the amplitude of the excitatory postsynaptic currents evoked (Pf/DMS 175 ± 12 pA versus Pf/DLS 84 ± 8; VPs/DMS 26 ± 3 versus VPs/DLS 100 ± 11, p < 0.05, Mann-Whitney U test; paired comparison per site of transfection, p < 0.05, Wil- coxon test, Figures 1G and 1H; Table S1). This preference was consistent with the times a connection was detected (Figure 1I). Figure 1J shows an AMPA antagonist applied through bath perfusion (CNQX [10 mM]), in either the presence or the absence of 4-AP+TTX (n = 4, n = 8, respectively), that abolished the excit- atory postsynaptic currents (EPSCs) (p < 0.05, Wilcoxon test). Importantly, in most of these cells, we identified them as spiny projection neurons versus putative interneurons (Figures S1E and S1F). As a summary from this first set of experiments, it was concluded that the PFs and the VPs present a preferential anatomical and functional innervation for the DMS and DLS, respectively. This allowed us to select the PFs and the VPs as regions of interest to answer whether the thalamostriatal projec- tions contribute to the initiation and execution of a sequence of movements. Mice Self-Initiate and -Execute Sequences of Lever Presses To investigate in vivo the contribution of the thalamostriatal neurons to the shaping of sequences of movements, we trained mice to develop sequences of lever presses (Figure 2) (Jin and Costa, 2010; Tecuapetla et al., 2016). In short, mice were trained on a self-paced operant task until they pressed a lever in bouts or sequences of more than one press. After 1 day of 740 Neuron 100, 739–752, November 7, 2018 magazine training and 3 days of continuous reinforcement, an- imals were trained for 11 days in a fixed ratio 8 schedule (FR8; Figure 2A), where a drop of sucrose (10%) as reinforcer was dispensed every 8 presses (Figures 2B and 2C; note that the 8 presses do not have to be continuous). Following this training, animals gradually organized their presses in self-paced se- quences or bouts of presses (Figures 2C and 2D, filled sym- bols) and decreased the proportion of single lever presses DMS DLS Cx DMS DLS Cx μm μm PoDMS PFs VPL VPM Bregma 0.2 mm Bregma-1.7 mm Bregma -2.2 mm Fr DLS retrobeads injection B A D F H J PFs Striatum VPs Striatum μm 3 4 2 3 4 * e Y F P s tr ia ta l fi b e rs (% t o ta l s tr ia ta l fi b e rs ) DMS DLS * 0 DMS DLS * PFs VPs Site of injection PFs VPs DLS Cx VM 1 500 50 50 DMS DLS * n= 6 29 c el ls n= 6 42 c el ls Site of injection PFs PFs VPsVPs 0 100 DMS DLS * * R e tr o b e a d s la b e le d c e ll s (% o f P F s+ V P s T h c e ll s) I PFs (VGluT2Cre): AAV-DIO-eYFP PFs or VPs: AAV-DIO-ChR2-eYFP (VGluT2Cre) VPs (VGluT2Cre): AAV-DIO-eYFP G DMS DLS DMS DLS C 2 1 DMS PFs DLS VPs DMS retrobeads injection PFs VPs Pf PoFr VM VPL # # * * PFs DMS DLS VPs DMS DLS DMS DMS DLSDLS E P S C ( p A ) 200 400 0 (4-AP+TTX) 50 ms 100 pA 100 %100 % 70 % 62.5 % Connectivity (%) PFs DMS DLS VPs DMS DLS E P S C ( p A ) 200 0 400 CNQXControl PFs DMS VPs DLS Control CNQX TTX + 4-AP 50 ms 100 pA n= 1 38 0 fib er s n= 9 30 fi be rs VM 20 40 60 80 * * E Bregma -2.4 mm Bregma-1.7 mm μm500 μm500 μm500 μm500 μm500 μm500 DMS DLS VPM DLS DMS Cx μm500 Bregma 0.3 mm 0 60 100 0 100 PFs cells PFs cells VPs cells VPs cells DMS fibers DLS fibers DMS fibers DLS fibers Figure 1. The Parafascicular and Ventroposterior Regions of the Thalamus Differentially Innervate the Dorsal Striatum (A) (1) Diagram of AAV viral injection into the parafascicular (PFs) region to express eYFP depicting the anterograde thalamo/striatal projections. (2) Coronal confocal image of VGluT2 Cre animal treated to express eYFP into the PFs of the thalamus. (3) Confocal image from the dorsal striatum of the injection in 2. (4) Magnifications of the yellow squares from 3. (B) Similar to (A) for ventroposterior (VPs) thalamostriatal projections. See also Figure S1. (C) Left: thalamic fibers quantification as a percentage from samples obtained as in (A) and (B). Data represented asmean ± SEM. *p < 0.05, MannWhitney U test (n = 5). Right: cherry color, comparative proportion of fibers of the PFs versus the VPs as the percentage of the total number of fibers. *p < 0.05, c2 test. (D) Top diagram: representation of the retrobeads injection into the dorsomedial striatum to retrogradely label thalamo/striatal neurons. Bottom left: coronal confocal image from a C57BL/6Jmouse injected with retrobeads into the dorsomedial striatum (DMS). Bottom right: retrogradely labeled cells into the PFs. Scale bar: 500 mm. (E) Similar to (A) for a dorsolateral striatum injection. The bottom right panel shows retrogradely labeled cells into VPs of the thalamus. (F) Left: retrogradely thalamic cell quantifications obtained from samples as in (D) and (E) as the percentage from DMS and DLS injections. Data represented as mean ± SEM. *p < 0.05, Mann Whitney U test (n = 4 and 5 animals for DMS and DLS injections). Right: dark green, relative proportion of DMS versus DLS percentage of total number of cells labeled in the thalamus, *p < 0.05, c2 test. See also Figure S2. (G) Diagrams representing the AVV injection into the PFs or the VPs region of the thalamus to express channelrhodopsin (ChR2) along their striatal projections. After allowing for ChR2 expression (see Table S1 and Figure S1), whole-cell recordings were performed in pairs in the DMS and the DLS of the same striatum per slice. (H) Top: representative EPSCs recorded in pairs (one cell in DMS and one cell in DLS per slice). Bottom: EPSCs amplitudes recorded from striatal neurons evoked by light stimulation of thalamostriatal terminals in the striatum. Note that the recordings were performed in pairs per slice. Orange lateral symbols are mean ± SEM. *p < 0.05, Mann Whitney U test; n = 13 PFs cells and 8 VPs cells, paired comparison, #p < 0.05, Wilcoxon test. (I) Percentage of times that a connection (EPSC) was detected depending on the thalamic AAV injection site and the striatal compartment recorded. (J) Left: peak amplitude of EPSCs evoked by light stimulation of thalamo/striatal terminals from the PFs/DMS (black) or the VPs/DLS (red). Dashed lines are from experiments in the presence of 4-AP+TTX. Solid lines are experiments that were in the absence of 4-AP+TTX. Right: a representative example from the EPSCs recorded in one cell. Blue dot, light stimulation, 1 ms. Neuron 100, 739–752, November 7, 2018 741 (Figure 2D, empty symbols). The percentage of lever presses in sequences per session (where a sequence is defined as a bout withR2 lever presses) became stable after 6 days of training in the FR8 schedule (sessions 4–11; p > 0.05, both for C57BL6/6J and the cohort of Cre animals, n = 6 animals per group; Krus- kal-Wallis test; Figure 2D, filled symbols). The mean number of lever presses after 11 sessions in FR8 was 4.2 ± 0.4 for the C57BL6/6J (cohort of Cre animals: PFs/DMSArch3.0 = 3.7 ± 0.6; PFs/DMSeYFP = 3.3 ± 0.2; VPs/DLSArch3.0 = 3.7 ± 0.3; VPs/DLSeYFP = 3.5 ± 0.3; 6 animals per group; p > 0.05 Krus- kal-Wallis test; Figure 2F, left). To measure the latency to start a sequence of lever presses along training, we took advantage of the fact that, with training, animals developed not only stereo- typed sequences of lever pressing but also a preferred path from the food magazine to the lever press (Figure 2B; Video S1). We placed an infrared beam in this path and measured the latency of the animals to start a new sequence of lever presses (dashed green line in Figure 2B). After 11 days of training, animals showed a stable latency between crossing the infrared beam and performing the first lever press of a self-paced sequence of presses (C57BL6/6J animals = 1.0 ± 0.1 s; cohort of Cre animals: Pf/DMSArch3.0 = 1.2 ± 0.2; Pf/DMSeYFP = 1.1 ± 0.1; VPs/DLSArch3.0 = 1.2 ± 0.1; VPs/DLSeYFP = 1.3 ± 0.1; 6 animals per group; Figures 2E and 2F, right; p > 0.05, Kruskal-Wallis test). This training shows that animals developed self-initiation and -execution of sequences of lever presses, allowing us to ask whether the thalamostriatal neurons contribute to the initi- ation and execution of sequences of movements. Neurons in the Parafascicular and the Ventroposterior Regions of the Thalamus Show Activity Modulation during the Initiation and Execution of Sequences of Lever Presses Once we established that mice are capable of self initiating and -executing sequences of lever presses and identified the PFs and the VPs as regions of the thalamus that contain neurons projecting preferentially to the DMS and DLS, respectively, we aimed to determine whether neurons from these regions modu- late their activity during the initiation and execution of sequences of movements. To that end, we performed electrophysiological recordings in the PFs and the VPs regions of animals trained to self-initiate and -execute sequence of lever presses. Seven ani- mals were trained as explained in Figure 2A and, after six days in FR8, were implanted with a movable electrode array (see STAR Methods; Figure 3A) in the thalamic region of interest (Figure 3B). After 3 days of recovery, the FR8 training restarted. Following 11 and 12 sessions of FR8, animals were subjected to an electro- physiological recording session, and the recording of 86 and 122 units in the PFs or the VPs were collected, respectively CRF 0 3 10 20 30 Time (sec) Lever press Reward 0 1 2 C A Sequences >= 2 lever presses Single lever press D PFs-Arch3.0 VPs-eYFP VPs-Arch3.0 PFs-eYFP 3 5 7 9 11 0 50 100 Latency Head entry -1 C57BL/6J L e v e r p re s s e s ( % ) 0 2 4 6 ns. S e q u e n c e s ( # ) 0 2 4 6 1 3 5 7 9 11 Sessions (FR8) 1 Presses/sequence M e a n n u m b e r S e c o n d s F B E Initiation Execution Lever Latency Magazine PFs-Arch3.0 VPs-eYFP VPs-Arch3.0 PFs-eYFP C57BL/6J Sessions (FR8) IR beam -4 (Days) (Sessions in FR8) 1 11-12 FR8 ns. Test day L a te n c y t o s ta rt (s e c ) Latency to start Figure 2. Training Mice to Self-Initiate and -Execute Sequences of Lever Presses (A) Training protocol. CRF, continuous reinforce- ment; FR8, fix ratio 8; test day, electrophysiological recording or optogenetic session. (B) Representation of the experimental setup showing a diagram of the operant box and the position of an infrared beam in the path from the reward magazine to the lever. (C) A representative example of the behavioral timestamps from one C57BL/6J mouse during the 11th day in FR8. The timestamps are aligned to the first press in each sequence. Only presses in sequence (R2 without magazine checking) are depicted. (D) The proportion of presses as sequences of presses (filled symbols) or as individual presses (empty symbols) during training for a group of C57BL/6Jmice and the cohorts of Cre animals used in the study. n = 6 animals per group. Data repre- sented as mean ± SEM. (E) Mean latency for C57BL/6J mice between crossing the infrared beam and the first lever press of a sequence as training progressed. n = 6 animals per group. Data represented as mean ± SEM. No difference between the Arch3.0 groups and their corresponding controls (eYFP) was observed along training; the difference in the latency between the PFs-Arch3.0 and the C57BL/6J group disappeared with training. (F) Mean number of presses in sequence and latency to start a sequence of lever presses for C57BL/6J mice and the different cohorts of Cre animals expressing Arch3.0 used in this study. n = 6 animals per group. Data represented as mean ± SEM, n.s., not significant, Kruskal-Wallis test, p > 0.05. 742 Neuron 100, 739–752, November 7, 2018 (examples of recordings aligned to the first lever press of several sequences of presses are presented in the Figures 3C and 3D). Analysis of the recordings in these animals (by ROC curves, see STAR Methods) showed that neurons in the PFs and the VPs displayed activity modulation during the initiation and execution of the sequences of lever presses in both regions (Fig- ures 3C–3G). During the initiation of sequences, there was a higher proportion of modulated units in the PFs than in the VPs (PFinitiation = 65% versus VPsinitiation = 41%; p < 0.05; c2 test; Fig- ure 3G, pie chart 1 versus pie chart 3), while during the execution of sequences, there were more positively modulated units in the VPs than the PFs (VPsexecution (+) = 31% versus PFsexecution (+) = 23%; pie chart 2 versus 4, p < 0.05, c2 test) (3 and 4 animals for PFs and VPs, 8 and 6 positions recorded, respectively). Addi- tionally, to determine whether a recorded unit had a preferential activation along the sequence, we evaluated whether each of the units presented a higher modulation before the initiation versus the execution of the sequence (Figure S3). The PFs presented a higher proportion of units positively modulated before the initi- ation of the sequences (PFs = 42% versus VPs = 27%; p < 0.05, c2 test; Figure S3C); conversely, the VPs presented a higher pro- portion of units positively modulated during the execution of the sequence (PFs = 22% versus VPs = 55%; p < 0.05, c2 test; Fig- ure S3C). Overall, these recordings show that the PFs and the VPs contain neurons displaying activity modulation during the initiation and execution of sequences of lever presses. Activity of Antidromically Photo-Identified Thalamostriatal Neurons during the Initiation and Execution of Sequences of Lever Presses Although the recordings in the thalamus showed that the PFs and the VPs regions of the thalamus contain neurons that modulate their activity during the initiation and execution of the sequences of lever press, they do not show whether thalamos- triatal neurons show activity related to the sequences. To specifically investigate this, we performed antidromic photo- identification (Lima et al., 2009) of thalamostriatal neurons in vivo while animals initiated and executed sequences of lever presses. In a subset of animals, we expressed ChR2 into either the paraf- ascicular or the ventroposterior region of the thalamus of VgluT2- Cre animals followed by an electrode array implantation above PFs -2.18 Bregma μm Pf VPL VPM VPM VPL -2.06 Bregma -2.18 Bregma Initiation F Execution Time (sec) -1 0 1 2-2 # u n it s Z s c o re -1 0 1 2 3 4 10 30 50 70 90 0 20 40 60 % u n it s m o d u la te d Initiation E Execution # u n it s % u n it s m o d u la te d 10 30 50 -1 0 1 2-2 0 20 40 60 G Initiation Execution * Initiation Execution * * 1 2 3 4 -2 -1 0 1 2 0 10 -2 -1 0 1 2 Time (sec) Time (sec) C S p ik e s /s e c S p ik e s /s e c S p ik e s /s e c c e s/ s e ki p S Th Str A (PFs) (VPs) Electrode array B D 0 10 0 8 Initiation Execution Initiation Execution Cx + modulated - modulated Time (sec) First press (PFs) (VPs) First press -2 -1 0 1 2-2 -1 0 1 2 Time (sec) 0 10 Pf Pf Po Po Po Lever Electrode array μm 500 500 Figure 3. The Parafascicular and the Ventroposterior Thalamus Activity Is Modulated during the Initiation/Execution of Sequences of Lever Presses (A) Left: picture of an animal showing the recording head stage. Right: diagram representing the electrode array implantation into the thalamus. (B) Left: photomicrographs of thalamic coronal slices showing the track po- sition of the cannula conducting the electrode array used to record into the parafascicular region (PFs; upper left panel) or the ventroposterior region of the thalamus (VPs; bottom right panel). Right: schematic representation of the recording sites. (C and D) Representative peri-event histogram and raster plot of the activity of two units recorded in the PFs (C) and two in the VPs (D) of the thalamus, aligned to the first press of a sequence of lever presses. (E and F) Activity modulation of the units recorded in the PFs (E) or the VPs (F) region of the thalamus (presented as Z score) aligned to the first press of the sequence of lever presses. (G) Proportions of units recruited in the parafascicular or the ventroposterior regions of the thalamus throughout the initiation and execution of sequences. The pie charts above show the proportions of positive (orange), negative (blue), or not modulated units along the initiation (1 s before the first press) or during the execution (1 s after the first lever press in the sequence), *p < 0.05, c2 test. See also Figure S3. Neuron 100, 739–752, November 7, 2018 743 the expression site. In these animals, we took advantage of the fact that ChR2-expressing neurons can be antidromically acti- vated through light stimulation in their axons (Lima et al., 2009) and implanted a fiber optic into the DMS or DLS to stimulate the axons of PFs neurons projecting to the dorsomedial striatum (PFs/DMS) or the axons of VPs thalamic neurons projecting to the DLS (VPs/DLS) (Figure 4A). Following this procedure, we recorded the activity of thalamic units during the initiation and execution of sequences of lever presses, and immediately after the recording session, we assessed whether the recorded units responded antidromically to pulses of blue light delivered into the striatum (Figure 4B). To classify a unit as thalamostriatal, we considered latencies of %8 ms (Vandermaelen and Kitai, 1980) (mean PFs/DMS = 5.4 ± 0.3 ms; VPs/DMS = 5.4 ± 0.5 ms; Figure 4B, bottom right panels) and waveform correla- tions of >0.9 (behavioral session versus antidromic light stimula- tion) (Figure 4B, right upper panels). Using these criteria, we identified that 34% of the units (30 out of 86) recorded in the PFs and 12% (15 out of 122) recorded in the VPs were respon- sive to the antidromic striatal stimulation (Figures 4E and 4F). Although a smaller number of neurons in the VP versus the PFs were antidromically photo-identified, a larger proportion of VPs/DLS units were modulated both during the initiation and the execution of sequences (PFs/DMSinitation = 51% of units versus VPs/DLSinitation = 88% p < 0.05; c2 test; PFs/DMSexecution = 24% versus VPs/DLSexecution = 58%, p < 0.05; c2 test; Figures 4E–4G). Remarkably, when we exam- ined whether the antidromically identified units were preferen- tially modulated during the initiation versus the execution or the end of the sequences of presses, we observed that the PFs contained a bigger proportion of units modulated during the initiation, while more units in the VPs were modulated during the execution (PFs/DMSinitiation = 43% versus VPs/DLSinitiation = 33% of the units; PFs/DMSexecution = 13% versus VPs/DLSexecution = 20%; p < 0.05, c2 test; Fig- ure S4G). These results show that the thalamostriatal neurons from the PFs and the VPs are differentially modulated during the initiation and execution of sequences of movements. Next, to investigate the relationship between the activity of the thalamostriatal neurons recorded and the parameters measured during the initiation and execution of sequences (latency to start a sequence and the number of presses in the sequence), we per- formed linear regressions between the activity of these units and these parameters (Figure 5). A summary of the proportion of units with a significant regression between their activity and the different parameters measured is presented in the Tables S2 and S3. Remarkably, the quantification of the proportion of units presenting either positive or negative regressions between the latency to start the sequences and the activity showed more units positively and negatively modulated in the PFs than in the VPs before the start of the sequence (Figure 5D, pie charts on the left; p < 0.05, c2 test). Conversely, the regression between the activity and the number of lever presses showed that the VPs contained more cells showing positive regression during the execution (Figure 5H, pie charts on the right; p < 0.05, c2 test). A similar quantification of the regression between the num- ber of presses in the sequence and the activity aligned to the end of the sequence did not reach differences (Figure S4D). None- theless, when we compared per unit the activity along time, con- trasting the initiation versus the execution versus the end of the sequences, it was evident that the VPs/DLS units presented a higher level of activity during the end of the sequence (Figures S4E and S4F). In summary, the analysis of the activity from the antidromically photo-identified thalamostriatal units showed that the activity of thePFs/DMSandVPs/DLSneurons ismodulatedduringeither the initiation or the execution of sequences. The regression anal- ysis between the activity and the initiation (latency), or the activity and the execution (number of presses), revealed that the PFs/DMS contained more units correlated with the latency to initiate a sequence while the VPs contained more units related to the execution of presses (PF/DMSlatency-related units = 22%versus VPs/DLSlatency-related units = 0%; PFs/DMSpress-related units = 7% VPs/DLS press-related units = 33%; p < 0.05 c2; Figure 5I). Optogenetic Inhibition of the Thalamostriatal Projections before Sequence Initiation Delays the Initiation of Learned Sequences Up to here, we have shown that the thalamostriatal cells display activity modulations related to the initiation and execution of se- quences movements. However, we have not tested whether the activity of the thalamic neurons in the PFs or VPs contributes to the initiation/execution of actions. To address this, we performed optogenetic inhibitions of thalamic neurons during either the initi- ation or the execution of sequences of lever presses. The results from these inhibitions are presented in Figure S5 and confirm that the direct inhibition of the thalamic neurons in the PFs and VPs regions performed before the initiation of the sequence de- lays the initiation and affects the execution of sequences (Fig- ures S5B–S5J). However, since the direct inhibition of the cell bodies of the thalamic neurons affects the somas interfering with their output to the striatum but also to other brain areas, we asked whether the thalamostriatal projections were specif- ically necessary for the smooth initiation and execution of sequences, which required optogenetic inhibition of the thala- mostriatal terminals while animals initiated and executed se- quences. Although the feasibility of the optogenetic inhibition of neuronal terminals using Arch3.0 has been documented (Do-Monte et al., 2017; El-Gaby et al., 2016; Parker et al., 2016), even in the thalamostriatal projections (Parker et al., 2016), we further characterized the optogenetic inhibition of the thalamostriatal terminal in vivo and ex vivo (Figure S6). First, we verified in vivo that green light illumination of the thalamostria- tal terminals expressing Arch3.0 into the striatum decreased the firing rate of striatal units. For this verification, we expressed Arch3.0-eYFP in the PFs or the VPs (two animals each) and implanted an optrode into the DMS or DLS, respectively. This procedure allowed the examination of the striatal activity while delivering pulses of green light in vivo. Aligning the striatal activity to the onset of a 5 s pulse of green light showed unitary activity (Figures S6A–S6D) and local field potential (LFP) modulation from the inhibition of both the PFs/DMS and the VPs/DLS (Figure S6I). Additionally, we tested the same parameter for the inhibition of Arch3.0 in an animal expressing only eYFP (Figures S6E, S6F, and S6I). Additionally, we also verified in slice record- ings that green light illumination of the thalamostriatal terminals 744 Neuron 100, 739–752, November 7, 2018 decreased the probability to evoke postsynaptic spikes (Figures S6J–S6M). Verifications both in vivo and ex vivo confirmed the feasibility of using Arch3.0 to perform the optogenetic inhibition of the thalamostriatal terminals. Th Str A VGluT2-Cre 0.3 ms 50 μV Wf-sequences (Thalamic unit Str) S p ik e s /s e c C (PFs DMS) (VPs DLS)D PFs->DMS VPs->DLS 0 4 80 4 8 0 5 10 sti n U Latency (ms) Optic Fiber AAV + Electrode array (ChR2) Time (sec) Z s c o re 0 20 40 0 20 40 60 80 * * Initiation InitiationExecution Execution % u n it s m o d u la te d % u n it s m o d u la te d -2 -1 0 1 2 0 2 4 -2 -1 0 1 2 0 2 Time (sec) Time (sec) S p ik e s /s e c S p ik e s /s e c G -1 0 1 2 0 20 B E F # u n it s (PFs DMS) (VPs DLS) + modulated - modulated -1 0 1 2-2-1 0 1 2-2 Time (sec) 1 2 3 4 2 6 10 14 10 20 30 -1 1 2 3 -1 0 1 2-2-1 0 1 2-2 Time (sec) 0 Press 1 Press 1 Time (sec) Time (sec) PC 1 PC 2 PC 3 Light Wf-stimulation Figure 4. The Parafascicular and Ventro- posterior Thalamostriatal Neurons Are Modulated during the Initiation/Execution of Sequences of Lever Presses (A) Diagram of a sagittal brain slice representing the viral injection into the thalamic region of interest to express channelrhodopsin (ChR2) in the cell bodies and their axonal projections and the fiber optic im- planted into the striatum to antidromically photo- activate the thalamostriatal neurons. (B) Left: raster plot and histogram peri-stimulus of the activity of a photo-identified thalamo/ striatal unit. Its activity is aligned to a train of light stimulation (10 Hz; upper blue squares). Top right: mean waveform from the behavioral session (Wf- sequence) superimposed to the waveform of the same unit during the antidromic photo-identification (Wf-stimulation). PC 1–3, the three first principal components of the individual traces composing the two waveforms, presented on the left. Bottom right: latency of antidromic spikes evoked by the light stimulation of parafascicular/dorsomedial striatal cells (PFs/DMS) or ventroposterior/dorsolateral striatal cells (VPs/DLS) from the striatum. (C) Raster plot and peri-event histogram of the ac- tivity of a PFs/DMS unit aligned to the first press in the sequence of lever press. (D) Similar to (C) but for a VPs/DLS unit. (E and F) Activity of the photo-identified PFs/DMS (E) and VPs/DLS (F) units (presented as Z score) aligned to the first press in the sequence. (G) The proportion of units antidromically photo- identified as PFs/DMS (left) or VPs/DLS (right) along the initiation and execution of sequences of presses. The pie charts above represent the average of 1 s before or after the first lever press in the sequence *p < 0.05, c2 square test. See also Figure S4. Next, we performed the optogenetic inhibition of the PFs/DMS or the VPs/DLS terminals in vivo while ani- mals initiated and executed sequences of lever presses. For this purpose, we performed surgery in VgluT2-Cre mice to bilaterally express the inhibitory opsin Arch3.0-eYFP (or only the eYFP for con- trol experiments) into either the PFs or the VPs region of the thalamus followed by fiber optic implantation into the DMS or the DLS, respectively (Figure S7). After surgery, animals recovered for 3 days before training was started. Following 11–12 days of training in FR8, animals were subjected to an optogenetic manip- ulation test session. During this optoge- netic test session, trials of inhibition were randomly intercalated (30% of the trials along the ses- sion), allowing us to compare the effects of light inhibition of the thalamostriatal terminals in each animal within the same session (Figure 6). Neuron 100, 739–752, November 7, 2018 745 A E + beta - beta * * 0 40 0 60 0 40 0 60 0 20 40 0 4 -2 0 2 Presses(#) -2 0 2 -2 0 2 Initiation Execution + beta - beta S p ik e s /s e c Spikes-Presses 2 3 4 Time (sec) # seq B F 60 % u n it s % u n it s Lever press Spike Latency + beta - beta + beta - beta Initiation Execution (PFs DMS) (VPs DLS) Time (sec) I (PFs DMS) (VPs DLS) (PFs DMS) (VPs DLS) (PFs DMS) (VPs DLS) S p ik e s /s e c Latency (s) -2 0 2 Time (sec) # seq Spikes sec Latency Latency & Presses Presses 33 %67 %7.5 %22 % 70.5% (PFs DMS) (VPs DLS)* 1 3 5 R=-0.82 2 4 86 5 8 P re s s e s (# ) 4 8 12 R=0.70 1 2 3 4 5 6 L a te n c y (s e c ) 2 3 4 6 Spikes-Latency 5 Spikes sec 10 20 30 40 50 0 5 10 15 Lever press Spike Latency (PFs DMS) (VPs DLS) (PFs DMS) (VPs DLS) C D G H Figure 5. The Parafascicular and Ventro- posterior Thalamostriatal Neurons Are Correlated with the Initiation/Execution of Sequences of Lever Presses (A) Raster plot and the peri-event histogram from the activity of the unit aligned to the first press in the sequence of lever presses. (B) Positive linear regression of the activity and the latency to start a sequence of the thalamostriatal unit shown in (A). (C) The proportions of the thalamostriatal units that showed significant positive (blue) or negative (pink) regression between the activity and the latency to start a sequence. (D) The pie charts show the proportion of units that presented significant regressions 2 s before (left) or 2 s after (right) the first press in the sequence from (C). *p < 0.05, c2 test. (E) As in (A) but for a unit showing a negative regression between its activity and the number of lever presses. (F) Similar to (B), but, in this case, for regressions between the activity of the units and the number of lever presses in the sequences. (G and H) The proportions of the thalamostriatal units that showed significant positive (blue) or negative (pink) regression between the activity and the number of lever press in the sequences (G). The pie charts (H) show the proportion of units that presented significant regressions 2 s before (left) or 2 s after (right) the first press in the sequence from (G). *p < 0.05, c2 test. (I) Venn diagram representing the proportion of neurons with regression only to the latency (gray color), to the latency and lever press (black color), and only to the number of lever presses (red color), *p < 0.05, c2 test. See also Figure S4. 746 Neuron 100, 739–752, November 7, 2018 To evaluate the contribution of thalamic neurons during the initiation of individual sequences of lever presses, we inhibited their activity before the first lever press of each bout of presses (Tecuapetla et al., 2016). We took advantage of the fact that, with training, animals developed a preferred path from the food magazine to the lever press (see Video S1). We placed an infrared beam in the path between the magazine and the lever press (dashed red line in Figure 6A, lower panel). The breaking of this beam allowed us to trigger the optogenetic inhibition before the first lever press, ensuring that the thalamostriatal pro- jections were inhibited before the first lever press (on average 0.94 ± 0.1 s before the first lever press in the sequence) for 5 s. Following this procedure, we observed that the optogenetic in- hibition of the PFs/DMS or the VPs/DLS terminals, before the start of the sequence of lever presses, increased the latency to initiate the sequences (PFs/DMSlatencyon = 2.9 ± 0.7 s versus PFs/DMSlatencyoff = 0.8 ± 0.1; VPs/DLSlatencyon = 2.7 ± 0.5 versus VPs/DLSlatencyoff = 0.9 ± 0.9, p < 0.05, Wilcoxon test, paired plots Figures 6B and 6C, upper panels; Mann-Whitney U test for the comparison to a previous day: boxplots in the same figures; also see Video S2). Intriguingly, only the inhibition of the VPs/DLS terminals before the start of the sequence increased the number of presses (VPs/DLSleveron = 5.4 ± 0.5 lever presses versus VPs/DLSleveroff = 3.8 ± 0.5, p < 0.05, Wilcoxon test, paired plots Figure 6E; Mann-Whitney U test for the comparison to a previous day: boxplots in the same figure). This increase in the number of presses did not affect the sequence length (p > 0.05, Figure 6G), implying that animals performed more lever presses in the same time. All effects were compared to VgluT2-Cre animals expressing only eYFP under the same treatments and data of the same animals in a previous session without light stimulation (orange paired data and boxplots, Figure 6). Consistently, the effects detected were only present when animals expressed Arch3.0 and were exposed to the light inhibition of the thalamos- triatal fibers. S e c o n d s 0 2 4 6 8 # p re s s e s onoff onoff 0 1 2 onoff onoff onoff S e c o n d s 0 2 4 6 8 onoff onoff0 2 4 6 8 0 2 4 6 8 # p re s s e s 0 2 4 6 8 1 0 0 1 2 S e c o n d s 0 5 1 0 1 5 2 0 onoff onoff 0 2 4 6 8 Number of lever presses Number of lever presses Sequence length Sequence length * * * O n / O ff te s t d a y / p re v. d a y O n / O ff te s t d a y / p re v. d a y O n / O ff te s t d a y / p re v. d a y O n / O ff te s t d a y / p re v. d a y O n / O ff te s t d a y / p re v. d a y O n / O ff te s t d a y / p re v. d a y PFs DMS VPs DLS 0 5 1 0 1 5 2 0 Arch3.0 eYFP eYFP onoff onoff S e c o n d s B C D E F G Arch3.0 eYFP eYFP Inhibition - initiation A Th Str VGluT2-Cre + thalamic injection AAV DIO-Arch3.0-eYFP * * * 0 2 4 6 8 The light is activated when crossing the sensor of the latency Latency Lever Magazine Optical fibers onoff IR beam Latency to start Latency to start 0 2 4 6 8 1 0 Figure 6. Inhibiting of the Thalamostriatal Projection before the Sequence Is Initiated Delays the Initiation (A) Top: diagram of a sagittal brain slice representing the viral injection into the thalamic region of interest to express archaerhodopsin 3.0-eYFP in the cell bodies and their axonal projections and the fiber optic implanted in the striatum to inhibit the thalamostriatal terminal in vivo. The fiber optic implantation is bilateral. Bottom: experimental setup showing a diagram of the operant box and the position of an infrared beam in the path from the reward magazine to the lever (red dashed line) used to trigger the optogenetic inhibition of the thalamostriatal terminal before the initiation of the sequence of lever presses. See also Figure S6. (B and C) Effect on the latency to start a sequence by the optogenetic inhibition of the parafascicular region/dorsomedial striatal terminals (PFs/DMS) (B) or the ventroposterior region/dorsolateral striatal terminals (VPs/DLS) (C). Paired plots panels: latency to initiate a sequence of lever presses for animals ex- pressing either Arch3.0-YFP (green) or eYFP (orange); each point is the median latency per animal during trials of optogenetic inhibition (on) versus trials without optogenetic inhibition (off) from the same session, *p < 0.05, paired plots: Wilcoxon test. Right: to further verify the effect during the optogenetic session, we calculated the ratio of the optogenetic session (On) against the ratio from the same animals in a previous session without optogenetic inhibition (Off), *p < 0.05, boxplots, Mann-Whitney U test. (D and E) Same as in (B) and (C), but, in this case, for the quantification of the number of lever presses per sequence. (F and G) Same as in (B) and (C), but, in this case, for the quantification of the sequence length (time from the first to the last press per sequence). See also Figure S7. Neuron 100, 739–752, November 7, 2018 747 Importantly, to test the possibility that it is the region of the striatum itself that is necessary for the observed behavioral ef- fects rather than the thalamic input specifically, we performed an additional control. For this control, we inhibited the PFs termi- nals into the DLS (PFs/DLS; as from the ex vivo recording data, these connections showed to be stronger than the VPs/DMS [Pf/DLS 84 ± 8 pA; VPs/DMS pA 26 ± 3; p < 0.05, Mann- Whitney U test, Figure 1H]) and compared this to the effects of inhibiting the VPs/DLS on the initiation and execution of se- quences. Consistent with the idea that it is the specific thalamic input that is relevant for proper initiation and execution, the inhi- bition of the PFs/DLS did not affect either the latency or the number of presses in the sequences (Figure S8). The Inhibition of the Thalamic Ventroposterior Dorsolateral Striatal Projections during the Execution Increased the Number of Actions inside the Sequence Only Late in Training The finding showing that inhibition of the VPs/DLS, but not the PFs/DMS, terminals increased the number of lever presses (Figures 6D and 6E), in combination with the results from the re- cordings of theVPs/DLSneurons showing that abigger propor- tion of units in the VPs presented significant regressions between the activity and the number of lever presses in the sequences (Figure 4), raised the question of whether the observed increase in the number of presses was specific to the inhibition of the VPs/DLS terminal before the initiation. Hence, to evaluate whether intact activity of the thalamostriatal projection was required during the execution of sequences. We performed a second optogenetic session in the same animals. As in the previ- ous experiment, we took advantage of the ability of animals to self-initiate and -execute sequences while this time performing the optogenetic inhibition of the thalamostriatal terminals only during the execution. This was achieved by triggering the light in- hibition with the first press in the sequence (Figure 7A, bottom). As result of this last manipulation, we observed that only the inhi- bition of the VPs/DLS terminals increased the number of lever press in the sequences (VPs/DLSlever-presson = 5.1 ± 0.4 versus VPs/DLSlever-pressoff = 3.2 ± 0.3, p < 0.05,Wilcoxon test, paired plot Figure 7E; Mann-Whitney U test for the comparison to the previous day: boxplots same figure), without affecting the sequence length (VPs/DLSSequence_lengthon = 3.9 ± 1.5 s versus VPs/DLSSequence_lengthoff = 2.7 ± 0.9, p > 0.05, Wilcoxon test Figure 7G). This increase in pressingwasconfirmed tobe specific to the optogenetic inhibition of the VPs/DLS terminals, as we did not observe it when inhibiting the PFs/DMS terminals (Fig- ure 7D) or in animals expressing only eYFP (orange color data, Figure 7E). Next, to investigate whether the increase in number of actions inside the sequence as a consequence of the photo-inhibition of VPs/DLS terminals was specific to learned sequences, we per- formed another experiment, this time inhibiting the VPs/DLS terminal early in training (matching the expression of Arch3.0 to be comparable to the inhibition late in training; Figures S9A– S9D). In this experiment, we did not observe an increase in pressing (Figure S9E), favoring the idea that the VPs/DLS con- tributes to the execution of the sequence of movements only for a learned sequence. Finally, to evaluate whether the striatal inhibition of the thalamic fibers may affect passing fibers, we performed two additional experiments. In one, we recorded the activity in a cortical region while photo-inhibiting the thalamostriatal fibers in the striatum. In the second, we expressed Arch3.0 into the thalamus and its projections to the striatum and cortex but im- planted fiber optics in the cortex to inhibit the thalamocortical terminals. To select a cortical region from which to record, and the region in the thalamus to express Arch3.0, we used the fact that the PFs mainly projects to the striatum sending few collaterals to cortex, while the VPs mainly projects to the cortex, leaving few collaterals in the striatum (Smith et al., 2014) (also see Figure S10). Therefore, we infected an animal in the VPs and implanted a movable recording electrode into the somatosensory cortex (S1), the primary recipient of the VPs terminals in the cortex (Hunnicutt et al., 2016). Figure S6G shows the activity from S1 recorded in different deeps (five ses- sions, moving the electrode array 100 mm a day before each recording session). Neither the unitary activity (Figures S6G and S6H) nor the LFP recorded in S1 (Figure S6I) were signifi- cantly modulated by the inhibition of the VPs/DLS terminal in the DLS. On the other hand, in a separate group of animals, we bilaterally expressed Arch3.0-eYFP into the VPs region of the thalamus and bilaterally implanted optic fibers on the VPs pro- jections into S1 (Figure S11). Consistent with the previous experiment, we observed that the optogenetic inhibition of the VPs/S1 terminals before initiation or during execution of the sequences of lever presses did not affect the initiation or the execution (Figures S11B–S11I). Nonetheless, in this last experiment, the same inhibition from the operant box increased the horizontal displacement of animals in the open field (Figures S11J and S12; see Mathis et al., 2017). Together, the results from the optogenetic manipulations of the thalamostriatal terminals indicate that both the PFs/DMS and the VPs/DLS thalamostriatal projections contribute to proper initiation, and the latter additionally contributes to the proper execution of the learned sequence of movements. DISCUSSION The results presented here establish that the thalamostriatal pro- jections contribute to the smooth initiation and execution of se- quences of movements. First, we described the innervations from the parafascicular and from the ventroposterior regions of the thalamus to the dorsal striatum (Figure 1). Second, we showed that both the PFs and the VPs region of the thalamus, and thalamostriatal neurons from these regions, presented activ- itymodulated during and related to the self-initiation and -execu- tion of sequences of movements (Figures 3, 4, and 5). Third, we showed that the thalamostriatal projections from the PFs and the VPs regions of the thalamus contribute to the smooth initiation of a sequence of movements (Figure 6), with the latter specifically contributing to the proper execution of the learned sequences (Figures 6 and 7). The thalamostriatal innervations presented here from the PFs and the VPs are consistent with previous reports of these inner- vations in both primates and rodents (Eckert et al., 2012; Elena Erro et al., 2002; Hunnicutt et al., 2016; Pan et al., 2010). 748 Neuron 100, 739–752, November 7, 2018 A preferential innervation from these thalamic structures to specific striatal compartments had been previously suggested (Bradfield et al., 2013; Eckert et al., 2012), but no detailed at- tempts to characterize it regarding DMS versus DLS innervation had been made. Here, through the use of three different ap- proaches (anterograde labeling of the axons with a fluorescent protein using a viral vector, retrograde labeling with retrobeads, and testing the functional connectivity by ex vivo electrophysi- ological recordings), we described a preferential (not exclusive) innervation from the PFs into the DMS and from the VPs into the DLS (Figure 1). Importantly, since the anatomical presence of axons may not always reflect functional connectivity, partic- ularly for the VPs/DLS axons, as this connection had not been functionally demonstrated (the PFs/DMS had been pre- viously characterized; Ding et al., 2008), we showed that the differential anatomical axonal projections were mirrored in functional electrophysiological measurements (Figures 1G– 1J). In summary, from this set of experiments, we conclude that the PFs and VPs thalamic regions have a preferential inner- vation on the DMS and DLS compartments. Note that the PFs region comprised cells from the parafascicular, posterior, and the centrolateral while the VPs are from the ventroposter- omedial, ventromedial, and ventroposterolateral nucleus of the thalamus. Next, with the aim to identify whether the thalamostriatal pro- jections are necessary to shape the striatal activity observed during the initiation and execution of an action sequence (Jin and Costa, 2010; Jin et al., 2014), it became necessary not just to ascertain that the thalamic neurons showed activity modula- tion during the initiation and execution of sequences (Figure 3) but, in particular, that it was shown by thalamostriatal neurons. This was done through antidromic photo-identification (Lima et al., 2009), and in doing so, it became evident that the propor- tion of recruited units was different when comparing the total number of thalamic units recorded versus the photo-identified units, with more units in the VPs being modulated during the 0 2 4 6 8 0 2 4 6 8 10 0 5 10 15 20 0 2 4 6 8 0 2 4 6 8 10 0 5 10 15 20 0 2 4 6 8 0 2 4 6 8 10 0 2 4 6 8 0 2 4 6 8 10 * S e c o n d s # p re s s e s onoff onoff onoff onoff onoff onoff S e c o n d s onoff onoff # p re s s e s S e c o n d s onoff onoff Number of lever presses Number of lever presses Sequence length Sequence length O n / O ff te s t d a y / p re v . d a y O n / O ff te s t d a y / p re v . d a y O n / O ff te s t d a y / p re v . d a y O n / O ff te s t d a y / p re v . d a y O n / O ff te s t d a y / p re v . d a y O n / O ff te s t d a y / p re v . d a y onoff onoff S e c o n d s B C D E F G Inhibition - execution A Th Str VGluT2-Cre + thalamic injection AAV DIO-Arch3.0-eYFP * 0 2 4 6 8 10 0 2 4 6 8 10 Lever Magazine Optical fibers The light is activated on the first lever press of the sequence Arch3.0 eYFP eYFP Arch3.0 eYFP eYFP VPs DLSPFs DMS Latency to start Latency to start Figure 7. Inhibition of the Ventroposterior/Dorsolateral Striatal Terminals during Execution Increases the Actions inside the Sequence (A) Top: diagram of a sagittal brain slice representing the viral injection into the thalamic region of interest to express archaerhodopsin 3.0-eYFP in the cell bodies and their axonal projections and the fiber optic implanted in the striatum to inhibit the thalamostriatal terminal in vivo. The fiber optic implantation is bilateral. Bottom: experimental setup showing a diagram of the operant box to achieve the optogenetic inhibition during the execution; the light inhibition was turned on with the first press in the sequence of presses (red lever). (B and C) Effect on the latency to start a sequence by the optogenetic inhibition of the parafascicular/dorsomedial striatal terminals (PFs/DMS) (B) or the ventroposterior/dorsolateral striatal terminals (VPs/DLS) (C). Paired plots panels: latency to initiate a sequence of lever presses for animals expressing either Arch3.0-YFP (green) or eYFP (orange); each point is the median latency per animal during trials of optogenetic inhibition (on) versus trials without optogenetic inhibition (off), *p < 0.05, paired plots: Wilcoxon test. Right: to further verify the effect during the optogenetic session, we calculated the ratio of the optogenetic session (On) against the ratio from the same animals in a previous session without optogenetic inhibition (Off), *p < 0.05, boxplots, Mann-Whitney U test. Note that in this case the inhibition was triggered by the first press in the sequence of presses. (D and E) Same as in (B) and (C), but, in this case, for the quantification of the number of lever presses per sequence. (F and G) Same as in (B) and (C), but, in this case, for the quantification of the sequence length (time from the first to the last press per sequence). See also Figure S7. Neuron 100, 739–752, November 7, 2018 749 initiation and execution of sequences of movements (Figures 4E–4G). These differences in activity modulation may be ex- plained by the heterogeneity of neuronal populations in the same thalamic nucleus (Jahnsen and Llinás, 1984) or by taking into account that the thalamic neurons receive a diversity of in- puts (Cornwall and Phillipson, 1988), which in turn may provide a different temporal drive during the performance of sequences of movements. Importantly, the fact the thalamostriatal anti- dromically photo-identified units were positively modulated during the initiation and execution of sequences of movements (Figures 4 and 5) allowed us to consider that these unitsmay pro- vide an excitatory drive to shape the striatal activity. It is probable that they convey signals of sudden changes in the behavioral context (Yamanaka et al., 2017), such as when self-initiating an action or when sensory information is necessary to appropriately execute a sequence of movements (it is known that putamen neurons respond to the speed of a vibrotactile stimuli and that in- formation may come from the sensory thalamus; Merchant et al., 1997; Vázquez et al., 2012). Although we showed the functional connectivity of the PFs and the VPs into the striatum (Figure 1G), and a differential modulation of PFs and VPs thalamostriatal neurons during the initiation and execution of sequences (Figures 4A–4G), reason- able concern exists that the antidromic photo-identification from the striatum may also activate axons from the thalamus traversing through the striatum reaching a different final target (Hunnicutt et al., 2016). Therefore, in the third set of experi- ments, we addressed whether the thalamostriatal projections activity was specifically required for the initiation and execution of sequences (Figures 6 and 7). From these experiments, we observed that both the undisturbed PFs/DMS and the VPs/DLS projections were required for the smooth initiation of sequences while only the undisturbed VPs/DLS projection was required for the proper execution. However, to conclude that there is a specific contribution from the thalamic inputs to the striatal compartments, we tested the possibility that it may be the region of the striatum itself that is necessary for the observed behavioral effects, not the thalamic input. This possibility was evaluated by the inhibition of the PFs/DLS versus the VPs/DLS (Figures 6 and 7 versus Figure S8), where we observed that the inhibition of the PFs/DLS had no behav- ioral effects as opposed to the inhibition from the VPs/DLS. These results strengthen the idea that it is indeed specific tha- lamostriatal inputs that provide a specific temporal input driving the striatum, which contributes to initiating and executing a sequence of movements properly. Also, to investigate the pos- sibility that the optogenetic inhibition of the thalamostriatal ter- minals may affect passing fibers to cortical targets, we measured the activity in S1 while inhibiting the VPs/DLS ter- minals in the DLS. From these experiments, no modulation in S1 was observed either at the extracellular unitary activity level or in the LFP (Figures S6G–S6I). Furthermore, we also per- formed optogenetic inhibition of the VPs/cortical projections into S1 (Figure S11). Surprisingly, when we inhibited the VPs/S1 projection in the cortex, no deficits in the perfor- mance of the action sequence were observed. With no attempt to claim that we inhibited all VPs/S1 projections, this experi- ment showed that while inhibition of the VPs terminals into the striatum was enough to delay the initiation and increase the number of repetitions in the sequence of actions, the same area of inhibition of VPs terminals into S1 was not sufficient to affect the action sequence. These findings highlight that a specific drive is provided from the ventroposterior thalamic re- gion to the dorsolateral striatum, probably to provide online feedback of the sequence of movements being performed. Note that we focus our evaluation of affecting passing fibers on the VPs/DLS versus VPs/S1, a test to evaluate the fibers from the PFs/DMS reaching cortex remains to be tested. The finding that thalamostriatal terminals from both the PFs and the VPs contribute to the initiation is consistent with the idea that the thalamus provides inputs driving the phasic depo- larization of the striatum during the initiation of sequences (Jin and Costa, 2010; Jin et al., 2014), probably conveying signals that support the selection of movements or actions (Thorn and Graybiel, 2010). Intriguingly, the fact that the selective inhi- bition of the VPs/DLS connection increased the number of repetitions of actions inside a sequence (Figures 6 and 7) lead to the hypothesis that this inhibition decreases sensory feedback arriving from the VPs (Van Der Loos, 1976; Waite, 1973) necessary for the proper execution of sequences. Although this is a plausible hypothesis, it has to be considered that the execution was affected only when inhibiting the VPs/DLS terminals late in training (Figures 6 and 7), not early in training (Figure S9). Furthermore, since no modulation of locomotion was observed when optogenetically inhibiting the VPs/DLS projections in the open field (Figure S12), the idea supported here is that the VPs/DLS terminals contribute spe- cifically to the execution of sequences once these sequences have been learned. The mechanism through which the PFs/DMS and VPs/DLS projection contribute to the proper initiation and execution of actions is still to be discovered. Nonetheless, from the estimation of the striatal cell types receiving the thalamic inputs in our ex vivo recordings (Figures S1E and S1F), two options are present themselves. One mechanism would involve the direct modulation of the striatal projection neurons. A secondmechanismmay involve the classical feedfor- ward inhibition of the spiny projection neurons (Koós and Tep- per, 1999) or the recent finding of facilitation of the activity of SPN through the control of GABAergic interneurons controlling other striatal interneurons (Assous et al., 2017; Lee et al., 2017). Importantly, with regards to the VPs/DLS synapses that contribute to the execution, either of these possibilities would involve synaptic strengthening with learning, as these pro- jections were shown to be relevant only later in training. In conclusion, this study highlights the thalamostriatal contri- bution to the delineation of a sequence of movements and sup- ports a model in which specific thalamostriatal projections contribute to shaping the smooth initiation and proper execution of sequences of movements. Also, our findings support the idea that the thalamostriatal projections constitute parallel loops serving temporally defined striatal compartment-dependent functions (Parent and Hazrati, 1995; Yin and Knowlton, 2006) and may have important implications for pathological conditions that produce excessive repetitive behaviors or excessive behav- ioral switching. 750 Neuron 100, 739–752, November 7, 2018 STAR+METHODS Detailed methods are provided in the online version of this paper and include the following: d KEY RESOURCES TABLE d CONTACT FOR REAGENT AND RESOURCE SHARING d EXPERIMENTAL MODEL B Animals d METHODS DETAILS B Training B Single-Unit Recording, Antidromic Photo-Identifica- tion, and Thalamostriatal Inhibition Verification B Stereotaxic Virus Injections and Fiber Implantation B Retrobeads Injections B Temporally Defined Optogenetic Striatal Manipula- tions In Vivo B Thalamostriatal Fibers Quantification B Quantification of Neurons Expressing Opsins B Ex Vivo Whole-Cell Recordings B Horizontal Distance B Behavior and Optogenetic Inhibition B Analysis of the Electrophysiological Recordings In Vivo d QUANTIFICATION AND STATISTICAL ANALYSIS d DATA AND SOFTWARE AVAILABILITY SUPPLEMENTAL INFORMATION Supplemental Information includes 12 figures, four tables, and two videos and can be found with this article online at https://doi.org/10.1016/j.neuron.2018. 09.052. ACKNOWLEDGMENTS We thank Dr. Rui Costa for the VGluT2-Cre mice; M. Arias-Garcia, E. Gal- arraga, and J. Bargas for electrophysiology ex vivo support; S. Jimenez- Cristobal for help training animals; Professor R. Romo for comments on the manuscript; Dr. Anna Hobbiss and MD Anil Verma for proofreading the manuscript; and Gabriela X. Ayala Méndez and Ariadna Aparicio for help in the reproduction of the VGlutT2 Cre mice. Edgar Dı́az-Hernández is a doctoral student of Programa de Doctorado en Ciencias Biomédicas, Universidad Nacional Autónoma de México (UNAM) and a receiver the fellowship 574086 from Consejo Nacional de Ciencia y Tecnologia (CONACyT). This work was supported by the Ciencia Basica CONACyT grant 220412, Fronteras de la Ciencia CONACyT grant 2022, and the DGAPA-PAPIIT-UNAM grants IA200815 and IN226517 to F.T. AUTHOR CONTRIBUTIONS E.D.-H. and F.T. designed and wrote the study. R.C.-L. performed the ex vivo electrophysiology experiments. A.S.-F. helped with the analyses of the in vivo recordings. L.R.-S. performed the experiments of Figure S10. J.O.R.-J. pro- vided the genotyping and technical support for the development of the study. DECLARATION OF INTERESTS The authors declare no competing interests. Received: March 12, 2018 Revised: July 16, 2018 Accepted: September 27, 2018 Published: October 18, 2018 REFERENCES Assous, M., Kaminer, J., Shah, F., Garg, A., Koós, T., and Tepper, J.M. (2017). Differential processing of thalamic information via distinct striatal interneuron circuits. Nat. Commun. 8, 15860. Atasoy, D., Aponte, Y., Su, H.H., and Sternson, S.M. (2008). A FLEX switch tar- gets channelrhodopsin-2 to multiple cell types for imaging and long-range cir- cuit mapping. J. Neurosci. 28, 7025–7030. Boyden, E.S., Zhang, F., Bamberg, E., Nagel, G., and Deisseroth, K. (2005). Millisecond-timescale, genetically targeted optical control of neural activity. Nat. Neurosci. 8, 1263–1268. Bradfield, L.A., Hart, G., and Balleine, B.W. (2013). The role of the anterior, me- diodorsal, and parafascicular thalamus in instrumental conditioning. Front. Syst. Neurosci. 7, 51. Christoffel, D.J., Golden, S.A., Walsh, J.J., Guise, K.G., Heshmati, M., Friedman, A.K., Dey, A., Smith, M., Rebusi, N., Pfau, M., et al. (2015). Excitatory transmission at thalamo-striatal synapses mediates susceptibility to social stress. Nat. Neurosci. 18, 962–964. Cornwall, J., and Phillipson, O.T. (1988). Afferent projections to the parafascic- ular thalamic nucleus of the rat, as shown by the retrograde transport of wheat germ agglutinin. Brain Res. Bull. 20, 139–150. Ding, J., Peterson, J.D., and Surmeier, D.J. (2008). Corticostriatal and thala- mostriatal synapses have distinctive properties. J. Neurosci. 28, 6483–6492. Do-Monte, F.H., Minier-Toribio, A., Quiñones-Laracuente, K., Medina-Colón, E.M., andQuirk, G.J. (2017). Thalamic regulation of sucrose seeking during un- expected reward omission. Neuron 94, 388–400.e4. Eckert, U., Metzger, C.D., Buchmann, J.E., Kaufmann, J., Osoba, A., Li, M., Safron, A., Liao, W., Steiner, J., Bogerts, B., and Walter, M. (2012). Preferential networks of the mediodorsal nucleus and centromedian-parafas- cicular complex of the thalamus–a DTI tractography study. Hum. Brain Mapp. 33, 2627–2637. El-Gaby, M., Zhang, Y., Wolf, K., Schwiening, C.J., Paulsen, O., and Shipton, O.A. (2016). Archaerhodopsin selectively and reversibly silences synaptic transmission through altered pH. Cell Rep. 16, 2259–2268. Elena Erro, M., Lanciego, J.L., and Giménez-Amaya, J.M. (2002). Re-examina- tion of the thalamostriatal projections in the rat with retrograde tracers. Neurosci. Res. 42, 45–55. Grider, M.H., Chen, Q., and Shine, H.D. (2006). Semi-automated quantification of axonal densities in labeled CNS tissue. J. Neurosci. Methods 155, 172–179. Haber, S.N., Adler, A., and Bergman, H. (2012). The Basal Ganglia. In The Human Nervous System, 3rd Edition, J.K. Mai and G. Paxinos, eds. (Elsevier), pp. 678–738. Hikosaka, O., Nakahara, H., Rand, M.K., Sakai, K., Lu, X., Nakamura, K., Miyachi, S., and Doya, K. (1999). Parallel neural networks for learning sequen- tial procedures. Trends Neurosci. 22, 464–471. Hunnicutt, B.J., Jongbloets, B.C., Birdsong, W.T., Gertz, K.J., Zhong, H., and Mao, T. (2016). A comprehensive excitatory input map of the striatum reveals novel functional organization. eLife 5, e19103. Ito, M., and Doya, K. (2015). Distinct neural representation in the dorsolateral, dorsomedial, and ventral parts of the striatum during fixed- and free-choice tasks. J. Neurosci. 35, 3499–3514. Jahnsen, H., and Llinás, R. (1984). Electrophysiological properties of guinea- pig thalamic neurones: an in vitro study. J. Physiol. 349, 205–226. Jin, X., and Costa, R.M. (2010). Start/stop signals emerge in nigrostriatal circuits during sequence learning. Nature 466, 457–462. Jin, X., Tecuapetla, F., and Costa, R.M. (2014). Basal ganglia subcircuits distinctively encode the parsing and concatenation of action sequences. Nat. Neurosci. 17, 423–430. Kawai, R., Markman, T., Poddar, R., Ko, R., Fantana, A.L., Dhawale, A.K., Kampff, A.R., and Ölveczky, B.P. (2015). Motor cortex is required for learning but not for executing a motor skill. Neuron 86, 800–812. Neuron 100, 739–752, November 7, 2018 751 Koós, T., and Tepper, J.M. (1999). Inhibitory control of neostriatal projection neurons by GABAergic interneurons. Nat. Neurosci. 2, 467–472. Kupferschmidt, D.A., Juczewski, K., Cui, G., Johnson, K.A., and Lovinger, D.M. (2017). Parallel, but dissociable, processing in discrete corticostriatal in- puts encodes skill learning. Neuron 96, 476–489.e5. Lee, K., Holley, S.M., Shobe, J.L., Chong, N.C., Cepeda, C., Levine, M.S., and Masmanidis, S.C. (2017). Parvalbumin interneurons modulate striatal output and enhance performance during associative learning. Neuron 93, 1451– 1463.e4. Lima, S.Q., Hromádka, T., Znamenskiy, P., and Zador, A.M. (2009). PINP: a new method of tagging neuronal populations for identification during in vivo electrophysiological recording. PLoS ONE 4, e6099. Mathis, M.W., Mathis, A., and Uchida, N. (2017). Somatosensory cortex plays an essential role in forelimb motor adaptation in mice. Neuron 93, 1493– 1503.e6. Matsumoto, N., Minamimoto, T., Graybiel, A.M., and Kimura, M. (2001). Neurons in the thalamic CM-Pf complex supply striatal neurons with informa- tion about behaviorally significant sensory events. J. Neurophysiol. 85, 960–976. Merchant, H., Zainos, A., Hernández, A., Salinas, E., and Romo, R. (1997). Functional properties of primate putamen neurons during the categorization of tactile stimuli. J. Neurophysiol. 77, 1132–1154. Pan, W.X., Mao, T., and Dudman, J.T. (2010). Inputs to the dorsal striatum of the mouse reflect the parallel circuit architecture of the forebrain. Front. Neuroanat. 4, 147. Parent, A., and Hazrati, L.N. (1995). Functional anatomy of the basal ganglia. I. The cortico-basal ganglia-thalamo-cortical loop. Brain Res. Brain Res. Rev. 20, 91–127. Parker, P.R., Lalive, A.L., and Kreitzer, A.C. (2016). Pathway-specific remodel- ing of thalamostriatal synapses in Parkinsonian mice. Neuron 89, 734–740. Petreanu, L., Mao, T., Sternson, S.M., and Svoboda, K. (2009). The subcellular organization of neocortical excitatory connections. Nature 457, 1142–1145. Raju, D.V., and Smith, Y. (2005). Differential localization of vesicular glutamate transporters 1 and 2 in the rat striatum. In The Basal Ganglia VIII, J.P. Bolam, C.A. Ingham, and P.J. Magill, eds. (Springer), pp. 601–610. Romo, R., and Schultz, W. (1992). Role of primate basal ganglia and frontal cortex in the internal generation of movements. III. Neuronal activity in the supplementary motor area. Exp. Brain Res. 91, 396–407. Rothwell, P.E., Hayton, S.J., Sun, G.L., Fuccillo, M.V., Lim, B.K., and Malenka, R.C. (2015). Input- and output-specific regulation of serial order performance by corticostriatal circuits. Neuron 88, 345–356. Rueda-Orozco, P.E., and Robbe, D. (2015). The striatum multiplexes contex- tual and kinematic information to constrain motor habits execution. Nat. Neurosci. 18, 453–460. Smith, Y., Galvan, A., Ellender, T.J., Doig, N., Villalba, R.M., Huerta-Ocampo, I., Wichmann, T., and Bolam, J.P. (2014). The thalamostriatal system in normal and diseased states. Front. Syst. Neurosci. 8, 5. Tecuapetla, F., Matias, S., Dugue, G.P., Mainen, Z.F., and Costa, R.M. (2014). Balanced activity in basal ganglia projection pathways is critical for contraver- sive movements. Nat. Commun. 5, 4315. Tecuapetla, F., Jin, X., Lima, S.Q., and Costa, R.M. (2016). Complementary contributions of striatal projection pathways to action initiation and execution. Cell 166, 703–715. Thorn, C.A., and Graybiel, A.M. (2010). Pausing to regroup: thalamic gating of cortico-basal ganglia networks. Neuron 67, 175–178. Van Der Loos, H. (1976). Barreloids in mouse somatosensory thalamus. Neurosci. Lett. 2, 1–6. Vandermaelen, C.P., and Kitai, S.T. (1980). Intracellular analysis of synaptic potentials in rat neostriatum following stimulation of the cerebral cortex, thal- amus, and substantia nigra. Brain Res. Bull. 5, 725–733. Vázquez, Y., Zainos, A., Alvarez, M., Salinas, E., and Romo, R. (2012). Neural coding and perceptual detection in the primate somatosensory thalamus. Proc. Natl. Acad. Sci. USA 109, 15006–15011. Vong, L., Ye, C., Yang, Z., Choi, B., Chua, S., Jr., and Lowell, B.B. (2011). Leptin action on GABAergic neurons prevents obesity and reduces inhibitory tone to POMC neurons. Neuron 71, 142–154. Waite, P.M. (1973). Somatotopic organization of vibrissal responses in the ven- tro-basal complex of the rat thalamus. J. Physiol. 228, 527–540. Wilson, C.J. (1989). The basal ganglia. In The Synaptic Organization of the Brain, Third Edition, G.M. Shepherd, ed. (Oxford Press), pp. 279–316. Wu, Y.W., Kim, J.I., Tawfik, V.L., Lalchandani, R.R., Scherrer, G., and Ding, J.B. (2015). Input- and cell-type-specific endocannabinoid-dependent LTD in the striatum. Cell Rep. 10, 75–87. Yamanaka, K., Hori, Y., Minamimoto, T., Yamada, H., Matsumoto, N., Enomoto, K., Aosaki, T., Graybiel, A.M., and Kimura, M. (2017). Roles of cen- tromedian parafascicular nuclei of thalamus and cholinergic interneurons in the dorsal striatum in associative learning of environmental events. J. Neural Transm. (Vienna) 125, 501–513. Yin, H.H., and Knowlton, B.J. (2006). The role of the basal ganglia in habit for- mation. Nat. Rev. Neurosci. 7, 464–476. Zheng, T., and Wilson, C.J. (2002). Corticostriatal combinatorics: the implica- tions of corticostriatal axonal arborizations. J. Neurophysiol. 87, 1007–1017. Zhu, Y., Wienecke, C.F., Nachtrab, G., and Chen, X. (2016). A thalamic input to the nucleus accumbens mediates opiate dependence. Nature 530, 219–222. 752 Neuron 100, 739–752, November 7, 2018 STAR+METHODS KEY RESOURCES TABLE REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies Mouse anti- NeuN Monoclonal Millipore Cat# AB-177621; RRID: AB_177621 Alexa Fluor-594 Goat Anti-Mouse Thermo Fisher Scientific Cat# A-11032; RRID: AB_2534091 Bacterial and Virus Strains AAV1.EF1a.DIO.hChR2(H134R)-eYFP.WPRE.hGH UPENN Vector Core Cat# AV-1-20298P rAAV5/Ef1a-DIO-eARCH3.0- EYFP UNC Vector Core N/A AAV1.EF1a.DIO.eYFP.WPRE.hG UPENN Vector Core Cat# AV-1-27056 Chemicals, Peptides, and Recombinant Proteins Retrobeads (Green and Red) Lumafluor, USA N/A NaCl JT Baker Cat# 3624-01 12*H2O* Na2HPO4 JT Baker Cat# 3822-01 H2O* NaH2PO4 JT Baker Cat# 3818-01 HEPES Sigma-Aldrich Cat# H2275 Sucrose Sigma-Aldrich Cat# S9378 D (+) Glucose Sigma-Aldrich Cat# S9378 CaCl2 Sigma-Aldrich Cat# C3881 L- Ascorbic acid Sigma-Aldrich Cat# P2256 H2O*ATPNa2 Sigma-Aldrich Cat# A2383 GTPNa Sigma-Aldrich Cat# G8877 MgCl2 JT Baker Cat# 2444-01 Biocytin Sigma-Aldrich Cat# 576-19-2 EGTA Sigma-Aldrich Cat# E9884 KMeSO4 Sigma-Aldrich Cat# 2386-56-3 NaHCO3 Sigma-Aldrich Cat# S5761 KCl Sigma-Aldrich Cat# P9541 4-AP Sigma-Aldrich Cat# A-0152 TTX Alomone Labs Cat# T-550 CNQX Sigma-Aldrich Cat# C-239 Experimental Models: Organisms/Strains Mouse/VGluT2-cre Champalimaud Centre for the Unknown RRID: MGI: 5141283 C57BL/6J Jackson Laboratory RRID: IMSR_JAX:000664 Software and Algorithms MATLAB 2016 MathWorks https://www.mathworks.com/products/ matlab.html GraphPad Prism 6 GraphPad Software https://www.graphpad.com/ ImageJ NIH https://imagej.nih.gov/ij/ Cerebral Plex Blackrock Microsystems http://blackrockmicro.com/ Offline sorter 2.8.8 Plexon https://plexon.com/products/offline-sorter/ Neuroexplorer 3.6 Plexon https://plexon.com/products/neuroexplorer/ Med PC Med-Associates https://www.med-associates.com/ Im-Patch Open access software designed in LabView http://www.im-patch.com/download.html (Continued on next page) Neuron 100, 739–752.e1–e5, November 7, 2018 e1 CONTACT FOR REAGENT AND RESOURCE SHARING Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Fatuel Tecuapetla (fatuel@.ifc.unam.mx). EXPERIMENTAL MODEL Animals All procedures were approved by the institutional committee for the care and use of laboratory animals of the Cell Physiology Institute (Protocol number FTA91-16), National Autonomous University of México and the National Norm for the use of Animals (NOM-062- ZOO-1999). Males and females (2-4 months of age) from the C57BL/6J line or resulting from the backcrossing of VGluT2 Cre mice (Vong et al., 2011) (donated by Dr. Rui M. Costa, from the Champalimaud Centre for the Unknown) into Black C57BL/6J for at least 6 generations were used. VGlut2 Cre animals were obtained from our breading colony in our institutional bioterium (the VGlut2 Cre line was conserved in heterozygosis). Animals were housed under a cycle of 12 h light/dark (lights light at 6:00 am) with access ad libitum to food and water before beginning behavioral experiments. METHODS DETAILS Training We trained mice to develop self-paced sequences of lever presses as previously described (Jin and Costa, 2010; Tecuapetla et al., 2016). Animals were allowed to recover for 3-4 days after surgery (for viral injection and fiber optic implantation) before training started. All training was run in the same chamber, and consisted of 1 day of magazine training, 3 days of continuous reinforcement 5, 15 and 30 respectively and 11 days of fix ratio 8 (FR8), before the test sessions started. Each chamber (21.6 cm L 3 17.8 cm W 3 12.7 cm H; Med-Associates, St. Albans, VT) was equipped with one retractable lever on the left side of the food magazine and a house light (3W, 24V) mounted on the left side of the chamber (see Video S1). Sucrose solution (10%)was delivered into ametal cup in the magazine through a syringe pump. Magazine entries were recorded using an infrared beam and licks using a contact lick- ometer. Mice were placed on food restriction throughout training and fed daily after the training sessions with enough pellets tomain- tain 80%–85% of normal body weight depending on performance. Training started with a 30-min session of reward delivery in a random time schedule (on average one every 60 s) 30 times. In the second day of training, the animals were presented with a lever on the left side of the magazine and received a reward every time they pressed, allowing this 5 times (CRF5), followed by two days more of CRF15 and CRF30. After these three days on CRF the reward schedule was changed, so that a reward was delivered only after every eight lever presses (fix ratio eight; FR8). We kept the animals in FR8 for 11-12 days before the experimental test session. The behavior was self-paced and throughout training there was no explicit signal for the animal to initiate or terminate a sequence of lever presses. All timestamps of lever presses, magazine entries, and licks for each animal were recorded with 10ms resolution. A sequence of lever presses was defined as a bout of consecutive lever presses with no licks. Single-Unit Recording, Antidromic Photo-Identification, and Thalamostriatal Inhibition Verification To investigate thalamic activity, we recorded from the thalamus (4 animals in the VPs and 3 for the PFs) using either fixed or movable arrays of 16 electrodes [tungsten (35mm) (Innovative Neurophysiology, Durham, DC)]. The spike activity was initially sorted using an online algorithm (Central software, Blackrock Microsystems, UT), and only units with a clearly identified waveform and relatively high signal-to-noise ratio were used for further analysis using an offline-sorting algorithm (Offline Sorter, Plexon). In order to define if the recorded unit projected to the striatum, we used in vivo antidromic photo-identification (Lima et al., 2009). In short, we expressed ChR2 into the thalamic region of interest using VgluT2-Cre animals (a line of mice expressing Cre recombinase in the thalamic neu- rons). During the same surgery a movable array was implanted 200 microns above the ChR2 expression site and an optic fiber was implanted into the striatum. These procedures allowed us to record the activity of the thalamic neurons during the initiation and execution of actions and at the end of the session, we verified (if the neuron recorded responded to the light stimulation into Continued REAGENT or RESOURCE SOURCE IDENTIFIER Other Confocal microscope Carl Zeiss LSM 710 473 nm laser Laserglow http://www.laserglow.com 556 nm laser CNI Lasers http://www.cnilaser.com/ Movable arrays of 16 electrodes Innovative Neurophysiology, Durham http://www.inphysiology.com/ movable-arrays/ e2 Neuron 100, 739–752.e1–e5, November 7, 2018 the striatum (10 Hz, 1 s, 1-10ms pulses). Themovable array allowed us to search for responding cells in at least 4 sessions per animal, advancing the array 100 mm24 hr before the recording session. To determine whether a unit could be antidromically photo-identified, the spikes during the behavioral sessionwere collected, and immediately after the spikes of antidromic striatal ChR2 stimulationwere collected. Only the units that responded significantly to the train of light stimulation, had a correlation > 0.9 between the behavioral spike and the antidromically evoked spike and were below a latency of 8 ms were considered photo-identified. To record the inhi- bition of the thalamostriatal pathway, AAV-Arch3.0 injectionswere performed into the PFs or VPs followed by an optrode implantation into the DMS or DLS respectively (n = 2 animals for each condition). After 2-3 weeks of Arch3.0 expression, recordings from the stria- tumof animals in an open field and receiving 5 s of green light illumination into the striatum (0.05Hz, 30 times) were acquired. From the same recordings, the data were filtered to acquire the unitary activity or the LFP. Stereotaxic Virus Injections and Fiber Implantation To perform surgeries animals were anesthetized using a mix of oxygen (1liter/min) with 1% isoflurane (1%–2% for interventional pro- cedures). For the optogenetic experiments after anesthesia, each animal was bilaterally injected using glass pipettes with 500 nL of viral stock solution [rAAV5-EF1a-DIO-eArch3.0-EYFP (Vector core, University of North Caroline), AAV1.EF1a.DIO.eYFP.WPRE; AAV1.EF1a.DIO.hChR2(H134R)-eYFP.WPRE titer > 1x1012 (Vector core UPENN University Pennsylvania)] by pressure into either the parafascicular (PFs) or the ventroposterior regions (VPs) of the thalamus; coordinates from Bregma, PFs: AP 2.20mm, ML 0.70mm and DV 3.2mm. VPs: AP 1.80 mm, ML 1.65 mm, DV 3.65 mm, below the surface of the brain. After the injections [23 nL every 5 s (Nanoject II, Drummond Scientific), animals were left for >15 min to allow time for virus spreading, and a fiber optic (300mm) (NA 0.37) was implanted into each hemisphere of the striatum [dorsomedial striatum for PFs/DMS (n = 8 mice for Arch.3.0 and n = 6 mice for eYFP); coordinates from Bregma: AP: 0.20 mm, ML1.80, and DV 2.35 mm; or dorsolateral striatum for VPs/DLS (n = 8 mice for Arch.3.0 and n = 6 mice for eYFP); AP: 0.20 mm, ML: 2.50 mm, and DV 2.35 mm]. For the inhibitions in the thalamus the following coordinates were used: Anterior nuclei of the thalamus (ANT), AP: 0.4mm, ML: 0.75mm, DV: 3.35mm (n = 3 mice); for the centromedial nuclei (CM), AP: 1.5, ML = 0, DV = 3.35 (n = 4 mice); dorsal medial lateral nuclei of the thalamus (DML), AP:1.7,ML: 0.8, DV: 3.1 (n = 2mice). The Parafascicular (Pf), AP:2.2,ML: 1.0, DV: 3.2 (n = 4mice); and the ventral region of the thalamus (VPs), AP: 1.9, ML: 1.65, DV: 3.6 (n = 3 mice). The optical fibers were fixed to the skull using acrylic cement (Lang Dental Manufacturing). Retrobeads Injections For the retrograde labeling experiments, 300 nL of retrobeads (Lumafluor, USA), were injected into the dorsolateral (n = 5mice) or the dorsomedial striatum (n = 4mice). Coronal sections (50microns) of the thalamuswere obtained to determine the total number of cells labeled in the PFs or VPs. The quantification was done in one slice every 300 microns covering these regions. Temporally Defined Optogenetic Striatal Manipulations In Vivo Optical stimuli were delivered via 300 mm diameter implantable fibers (Doric lenses), coupled to a single longitudinal mode laser (MSL-FN-556, CNI lasers). To deliver light for the optogenetic inhibition a free launching system controlled by an AOM (aaoptoelec- tronics) and a fast speed shutter (Thorlabs) triggered by the TTL output from theMEDPCbehavioral boxwas used. For the antidromic photo-identification experiments a blue laser (Laserglow) coupled to a fiber optic (Doric lenses) was used. Measures of the power at the tip of the fiber (similar to the one implanted) were verified for every experiment using a powermeter (Thorlabs) and the power adjusted at the tip of the fiber to be z30 mW for the green light or 2-4 mW for the blue light. To achieve the optogenetic inhibition of the thalamostriatal projection before the initiation of a sequence of lever presses, we took advantage of the fact that animals devel- oped stereotypical sequences of lever presses (Figure 2; Video S1). Thus when the animal moved from the magazine to the lever, the infrared beam was broken, setting a timestamp to trigger light on and quantify the latency to initiate the sequence of lever presses (Figures 2B and 6A). To achieve the light inhibition during the performance of the sequence, we used the timestamp of the first lever press in the sequence of lever presses (Figures 7A). During the session of optogenetic inhibition there were control trials and stim- ulation trials. The stimulation trials were randomly presented throughout the session (30% of total trials). Thalamostriatal Fibers Quantification After extracting the brains of the experimental mice, 50-micron sections of the striatum (coronal sections) were mounted and sealed. Z stacks were acquired at 63x magnification (192x192x10 microns; 1 microns interslice) from a randomly selected quadrant, using a randomly positioned grid covering either the dorsolateral or dorsomedial striatum, see Table S4 for the ranges sampled (n = 5 mice per group) (ZEN lite software, Zeiss). These Z stackswere imported into ImageJ, then amaximumprojection imagewas used to apply a filter (Hessian filter) allowing for the quantification of fibers as defined by the number of fibers crossing a randomly generated line spaced approximately 20 microns (Grider et al., 2006). Quantification of Neurons Expressing Opsins Brains were sectioned in sagittal or coronal 50 mm slices [using a vibratome (S1000 Ted Pella)] and kept in PBS 1% solution before mounting or immunostaining treatment]. After washing with PBS, the tissue sections were permeabilized with 0.3% Triton X-100 in PBS, for 10 min at room temperature. After a blocking step (incubation for 30 min, at room temperature with 10% FBS:0.2% Triton Neuron 100, 739–752.e1–e5, November 7, 2018 e3 X-100: PBS), sections were incubated overnight at room temperature, with the antibody against NeuN. After washing with PBS, a secondary antibody conjugatedwith the Alexa 594 fluorochrome (1:1000, Molecular Probes) was added. Sections were then washed and DNA was counterstained with 4,6-diamidino-2-phenylindole (DAPI, Sigma). Once the sections were mounted, Z stacks of either the PFs or the VPs region of the thalamuswere acquired from the slice in the center of the injection and slices 300microns aside, from the upper right quadrant using a randomly positioned grid (square grid 200 microns; ZEN lite software, Zeiss). The Z stacks were imported into ImageJ and quantification of NeuN-positive and eYFP-positive cells was done. Ex Vivo Whole-Cell Recordings To express ChR2 in the thalamic neurons and their striatal projection we injected 300 nanoliters of AAV2.1 EF1a-DIO-ChR2 (Vector core) into the PFs (n = 7 mice) or the VPs (n = 6 mice) region of the thalamus (see details Table S1). Allowing from 8 to 16 days of expression (Table S1), the animals were deeply anesthetized with ketamine (120 mg/kg, i.p.; Anesket) and xylazine (30 mg/kg, i.p.; Bayer) and perfused transcardially with an ice-cold perfusion solution (in mM): 242 sucrose, 2.5 KCl, 7 MgCl2, 28 NaHCO3, 1.25 NaH2PO4, 10 glucose, 1 ascorbic acid, and 3 of sodium pyruvate (pH 7.4); saturated with 95/5% O2/CO2. Coronal slices were obtained at the striatal level. Slices were then transferred to a saline solution containing (in mM): 125 NaCl, 3 KCl, 1.3 MgCl2, 2.6 CaCl2, 26 NaHCO3, 1.25 NaH2PO4, 15 glucose (pH 7.4), saturated with 95/5% O2/CO2, 310 mOsmol/L and left to equilibrate in this saline solution for at least 1 h at room temperature. Single slices were transferred to a submerged recording cham- ber and superfused continuously with oxygenated saline (3–5ml/min). Postsynaptic currents in whole cell configuration were evoked by pulses of blue light (1ms), with a pair of cells (one in the DLS and one in the DMS) recorded per slice. To guarantee monosynaptic thalamostriatal connections these recordings were performed in the presence of 4-aminopyridine (120 mM) and TTX (1mM) (Petreanu et al., 2009), although the first cell recorded in every slice started in the absence of 4-AP-TTX to allow their voltage characterization (Figure S1E). The recordings were acquired with micropipettes made with borosilicate glass (Harvard apparatus 30-0057) and fire polished for DC resistances in the bath of about 3–6 MU. Internal solution was (in mM): 120 KMeSO3, 2 MgCl2, 10 HEPES, 10 EGTA/KOH, 1 CaCl2, 2.41 Na2ATP, 1.14 NaGTP, 3.18 biocytin, 10 NaCl, and 290 mOsmol/L. Neurons were visualized using an infrared filter with an upright microscope (Scientifica electrophysiology) and a digital camera (Evolution VF FAST). Whole cell record- ings in voltage clamp were acquired using an amplifier (PC-501A Warner Instrument), a NIDAQ (CB 68LP, National Instruments) and the Im-Patch (open access software designed in the LabView) for the acquisition of the data. Horizontal Distance To evaluate if optogenetic inhibition has an effect on the locomotor activity, the mouse was placed in an open field (40x40cm and height 30cm) and videos from the animals were acquired into a custom developed software in LabVIEW, at a rate of 15 frames per second. The light stimulation periods are signaled in the videos. The animal received a light stimulation of 5 s everyminute. Offline analysis tracking the position of the animal to determine the trajectory before and during the stimulation was performed (Tecuapetla et al., 2014). Behavior and Optogenetic Inhibition A sequence of lever presses was defined as a bout of consecutive lever presses (R2) with no head entry and no licking. The latency to initiate a sequence was determined by the time between the animal breaking an infrared sensor positioned between the magazine and the first press (Figure 2B). The latency was calculated for control trials and stimulation trials. The stimulation trials were presented randomly (30% of total trials per session). The median of all tests performed was compared, unless otherwise specified in the text. Analysis of the Electrophysiological Recordings In Vivo Once a putative unit was isolated (through on-line and offline sorting), the timestamps of the spikes were aligned to the first press in a sequence of lever presses using custom developed scripts inMATLAB. To determine the percentage ofmodulated units along time in each epoch we performed an ROC curve analysis to ask if the spike frequency in each bin of time was different to that of the baseline time. For this purpose, we first aligned the spikes to each epoch (e.g first press) using 2 s before and 2 after each epoch. We calcu- lated the spike frequency using a 200 ms time window. We used as a baseline from 3 to 2.5 s before the first press in the se- quences of lever presses. We compared the spike frequency in each bin of time against the baseline, using a sliding window of 10 ms. To resolve statistically if the area under the curve (AUROC) was significant we made 1000 permutations and divided the sum of AUROC values that fall in either >0.5 or <0.5, between number of permutations. The AUROC value was significant if the outcome was (p < 0.05). Furthermore, in each epoch we obtained a binary matrix comparing each bin to base line. This binary matrix was used to find the number of units modulated in each bin time. (Figures 3, 4, S3, and S4). To compare the modulation of a unit throughout the presses (when comparing Z-score), the activity before the first press was compared with the activity before each press. The units that showed the highest activity before the first press were designated as initiation only, and those that showed the highest activity before the last press were designated end. The units that were classified as execution had higher activity after the first press (Figures S3 and S4). For the linear regression analysis (Figures 5 and S4) the firing rate was determined in spikes/ sec per trial. The trials were sorted by the number of lever presses (2 to 10, groups every increment of one) or by the latency to start (1 to 10 s, in groups of 1 s). Only regressions with b different from zero (p < 0.05) were accepted; mobile windows of 500ms displacing every 20 ms were used. e4 Neuron 100, 739–752.e1–e5, November 7, 2018 QUANTIFICATION AND STATISTICAL ANALYSIS All data were presented as median ± SEM unless otherwise specified in the text. The significance level was p < 0.05. For the propor- tion comparison, a Chi square test was used; for the paired comparisons the Wilcoxon test and for the non-paired comparisons Mann-Whitney U test were used. All statistics analyses were done using Graph Pad and MATLAB. DATA AND SOFTWARE AVAILABILITY All data and MATLAB scripts will be available upon contact to the lead, corresponding author. Neuron 100, 739–752.e1–e5, November 7, 2018 e5 Neuron, Volume 100 Supplemental Information The Thalamostriatal Projections Contribute to the Initiation and Execution of a Sequence of Movements Edgar Díaz-Hernández, Rubén Contreras-López, Asai Sánchez-Fuentes, Luis Rodríguez- Sibrían, Josué O. Ramírez-Jarquín, and Fatuel Tecuapetla 1 Figure S1. The parafascicular (PFs) and ventroposterior (VPs) regions of the thalamus project to the dorsomedial (DMS) and dorsolateral (DLS) striatum and impinge on striatal projection neurons or interneurons. Related to Figure 1. (A) Representation of AAV viral injection into the thalamus to express eYFP in the PFs or VPs. (B) Cumulative coronal confocal images of VGluT2 Cre animals treated to express eYFP in the parafascicular (left panel) or the ventroposterior region of the thalamus (right panel) (n= 6 animals per group). (C) Percentage of neurons eYFP/NeuN+ from the viral infection in the parafascicular (left panel) or the ventroposterior region (right panel) of the thalamus (n= 3 animal per group). (D) Cumulative coronal confocal images showing the projections of the PFs (upper panel) or from the VPs (lower panel) of the thalamus into the striatum. (E) Voltage responses (recorded in whole cell) to current injection from a striatal projection cell (SPN) and an interneuron (Int). These cells are from the recordings presented in Figure 1H (before the bath application of 4-AP+TTX). (F) Percentage of SPNs versus interneurons that received EPSCs from the PFs or the VPs of the thalamus. Scale bar in B and D: 500µm. 2 Figure S2. Most of the parafascicular (PFs) and ventroposterior (VPs) thalamic cells retrogradely labeled from the striatum project to the dorsolateral (DLS) or to the dorsomedial (DMS) striatum. Related to Figure 1. (A) Dual retrobeads injections into the dorsolateral (green) and the dorsomedial (red) striatum (DLS and DMS respectively). (B) Retrogradely labeled cells in the PFs region of the thalamus (upper panels) or the VPs region of the thalamus (bottom panels), at two anteroposterior levels. (C) Proportion of cells retrogradely labeled in the PFs and VPs from dual retrobeads injections as in A (n=4 animals). Scale bar represents in A and B: 500µm. 3 Figure S3. The parafascicular (PFs) and the ventroposterior (VPs) regions of the thalamus display activity modulation during the initiation/execution of sequences of lever presses. Related to Figure 3. (A) Activity of the units recorded in the PFs (z-score), aligned to the first, second, last-1 and last press in the sequence of lever press. (B) Similar to A, for the activity recorded in the VPs. (C) Proportion of the units recorded in the PFs (left pie chart) or in the VPs of the thalamus (right pie chart). Comparison between pie charts, *p<0.05, x 2 test. 4 Figure S4. The parafascicular (PFs) and ventroposterior (VPs) thalamostriatal neurons display activity modulation during the initiation/execution of sequences of lever presses. Related to Figures 4 and 5. (A) Raster-plot and the perievent histogram from the activity of a unit aligned to the last press in the sequence of lever presses. (B) Negative linear regression of the activity of a thalamostriatal unit and the number of presses in the sequences. (C) Proportions of the thalamostriatal units that showed significant positive (blue panel) or negative (pink panel) regression between the activity and the number of presses. (D) Pie charts show the proportion of units presenting regressions 2 seconds before (left pies) and 2 seconds after the last press in the sequence (right pies). No difference in the proportion of these units was detected. *p>0.05, x 2 test. (E) Activity in Z score of the photo-identified parafascicular thalamostriatal units (PFsàDMS), aligned to the first, second, last-1 and last press in the sequences of presses. (F) Similar to E for the photo-identified ventroposterior thalamo-striatal units (VPsàDLS). (G) Proportion of PFsàDMS units (left pie charts) or the VPsàDLS (right pie charts) showing a preferential activation along the initiation/execution of the sequences. Comparison between pie charts, *p<0.05, x 2 test. 5 6 Figure S5. Inhibition of the thalamic neurons delays the initiation and increases the performance of learned sequences of lever presses. Related to main Figures 6 and 7. (A) Sagittal brain slice representing the viral injection into the thalamic region of interest to express archaerhodopsin (Arch3.0) and optic fiber implantation. The animals were bilaterally implanted. (B) Experimental setup showing a diagram of the operant box and the position of an infrared beam in the path (red dashed line) from the reward magazine to the lever used to trigger the optogenetic inhibition in the thalamic cells before the initiation of the sequence. (C) Example of the behavior aligned to the first press. Trials without light (control off) and with light inhibition (on) before the first press. Note the increased latency during the on trials. (D) Left panel, latency to start a sequence of lever presses for animals expressing Arch3.0-YFP [latency without light inhibition = 0.7±0.2 seconds versus latency with light inhibition = 1.4±0.3 seconds; n=7 animals (4PFs+3VPs), p < 0.05; Wilcoxon test); each point is the median latency for one animal]. The inhibition before the first lever press increased the latency to initiate an actions sequence, specifically the inhibition of the parafascicular or the ventroposterior region of the thalamus (the boxplots show the ratio from individual trials normalized to the median of no light trials (off) in the same session: Pf ration on/off = 1.9±0.2; VPs ratio on/off = 1.8±0.2, p < 0.05; Mann Whitney U test). Orange color, animals expressing only eYFP (n=3 animals); ANT: Anterior nuclei (n=3 animals); Pf: Parafascicular (n=4 animals); VPs: Ventroposterior (n=3 animals); CM: Central medial (n=4 animals) and DML: Dorsal medial lateral thalamus (n=2 animals). In these experiments the opsin, or eYFP, and the fiber optics were injected and implanted into the mentioned nucleus. The spreading of light above the threshold to inhibit cells is around 500 microns (Tecuapetla et al., 2014). (E) Similar to D, for the quantification of the number of lever presses per sequence. (F) Similar to D, for the length of each sequence (time first press-last press). (G) Experimental setup showing a diagram of the operant box to illustrate when the optogenetic inhibition was done during the execution. In this case the first press in the sequence triggered the light inhibition. (H) Example of the behavior aligned to the first press; as in C in this case the inhibition was performed during the execution of the sequences. (I) J and (K) Latency to start the sequence, number of lever presses and sequence length (time first press-last press) respectively. In this case the inhibition was performed during the execution of the sequences. The data is plotted as explained in D. 7 Figure S6. Inhibition of the thalamostriatal projections in vivo and ex vivo. Related to Figures 6 and 7. (A) Sagittal brain slice representing the viral injection of the thalamic region of interest to express archaerhodopsin 3.0 (Arch3.0) in the cell bodies and their axonal projections. In the same diagram the optrode implantation to record the striatal activity during the optogenetic inhibitions of the thalamostriatal fibers is presented (configuration used for C-F). (B) Representative peri-event histogram and raster plot of the activity of a unit aligned to the green light illumination, the insert at the top shows the waveform of the recorded unit. A high frequency firing unit was selected to better depict the inhibitory effect of light illumination. 8 (C) Activity of the units recorded in the striatum (z-score) while switching on the green light illumination to inhibit the thalamostriatal terminals in the striatum. (D) Mean of the activity (Z score) for the units that decreased their activity during green light illumination. The black line shows the average of all units. Insert, pie chart shows the proportion of the units modulated and no modulated (comparing baseline versus light on). The average latency to decrease their firing rate was 80±37 ms. (E) and (F) Similar to C and D, in this case for an animal expressing only eYFP. (G) and (H) Similar to C and D, in this case for an animal recorded in somatosensory cortex (S1) while optogenetically inhibiting the VPsàDLS terminals in the DLS. (I) Local field potential (average of at least 20 trials) recorded while illuminating with a pulse of green light for 5 seconds, each trace is from a different animal. (J) Diagram for the ex vivo whole cell recording in a parahorizontal brain slice, representing the recording site, the electrical stimulation site and the green light illumination. (K) Photomicrograph of a cell recorded in the striatum labeled with biocytin-Cy3 (red), and the thalamo-striatal fibers (green: Arch3.0-eYFP). (L) Left panel, action potentials generated for the electrical stimulation (train) of thalamo-striatal fiber. The size and shape of the postsynaptic responses are the combination of action potentials and EPSC recorded in voltage clamp (inverted for a better visualization). Right panel, same electrical stimulation + light illumination of the thalamostriatal fibers in the same experiment. Horizontal black bar = 50 ms and vertical black bar = 0.5 normalized voltage to max amplitude. (M) Quantification of the probability to evoke action potentials during the electrical stimulations versus electrical stimulation + light stimulation. *p<0.05, Mann Whitney U test. 9 Figure S7. Optic fiber tips position from the thalamostriatal optogenetic inhibition. Related Figures 6 and 7. (A) Sagittal brain slice representing the viral injection into the thalamic region of interest to express Arch3.0- eYFP in the cell bodies and their axonal projections. In the same diagram the fiber optic implanted into the striatum to inhibit the thalamo-striatal projections in vivo is presented. (B) Left panels, Coronal brain slices representing the fiber optic tips from the experiments inhibiting the parafascicular thalamo-dorsomedial striatal projection (PFsàDMS). Animals expressing Arch3.0-eYFP (green) and controls animals (containing only eYFP: orange). Right panels, representative photomicrograph of the expression of Arch3.0-eYFP into the parafascicular region (PFs; Top panel) and the fiber track into the dorsolateral striatum (DLS; bottom panel). (C) Similar to B, in this case for the experiments inhibiting the ventroposterior thalamo-striatal projection (VPàDLS). Scale bar represents in B and C: 500µm. 10 Figure S8. Inhibition of the parafascicularàdorsolateral striatal projections (PFsàDLS) does not increase the latency to start or the number of lever presses inside the sequence. Related to Figures 6 and 7. (A) AAV injection into the PFs to express Arch 3.0-eYFP in the cell bodies and their axonal projections. Each animal receives bilateral expression of Arch3.0 and optic fiber implantation in the dorsolateral striatum (DLS). (B) Operant box diagram showing the position of an infrared beam (red dot) in the path from the magazine to the lever used to trigger the inhibition of the PFsàDLS terminals before the initiation of the sequence of lever presses. (C) Diagram to inhibit during the execution, in this case the first lever press in the sequence trigger the inhibition of the PFsàDLS terminals. (D) Representative photomicrograph showing the expression of Arch3.0-eYFP in the site of injection (left panel) and an optic fiber track in DLS (yellow arrow, right panel). Scale bar 500µm. (E) Latency to initiate a sequence of lever presses for animals expressing Arch3.0-YFP; each point is the median latency for one animal during the trials of optogenetic inhibition (on) vs. the trials without optogenetic inhibition (off) [Arch3.0latency_on = 0.9±0.1, Arch3.0latency_off = 0.6±0.1sec, n= 6 animals; eYFPlatency_on = 1.1±0.1, eYFPlatency_off =0.6±0.1, sec, n= 3 animals On/Off]. Box plots: ratio of the optogenetic session (On) divided by the ratio from the 11 same animals in a session without optogenetic inhibition [Arch testOn/Off =0.9±0.1 vs eYFPOn/Off = 1.2±0.1 p>0.05 U Mann Whitney test]. (F) Similar to D, in this case presenting the optogenetic inhibition during the execution. (G) Diagram representing the optical fiber track tips from Arch3.0 (green; n=6) and eYFP (orange; n=3) animals used for E-F and H-I. (H) and (I) Number of lever presses for animals as explained in E and F respectively. Inhibition before initiation: Arch3.0lever_on= 2.9±0.4 vs. Arch3.0lever_off = 2.6±0.3 presses, n= 6 animals; eYFPlever_on= 4.5±0.3 versus eYFPlever_off = 3.6±1, n= 3 animals; ratio against a previous session with no light inhibition: Arch3.0 = 0.9±0.1 vs eYFP=0.7±0.1. Inhibition during the execution: Arch3.0lever_on= 2.8±0.3 vs. Arch3.0lever_off = 2.5±0.3 presses, n= 6 animals; eYFPlever_on=3.5±0.8 vs. eYFPlever_off =4.8±0.4, n= 3 animals; ratio against a previous session: Arch3.0=1±0.2 vs eYFP =1.4±0.2. 12 Figure S9. Inhibition of the ventroposterior thalamoàdorsolateral striatum projection early in training does not increase the number of lever presses inside the sequence. Related to Figure 6 and 7. (A) Training protocol for the optogenetic inhibition early in training. CRF: continuous reinforcement; FR8: Fix ratio 8; test day: optogenetic session. (B) Sagittal brain slice representing the viral injection into the ventroposterior thalamic region to express archaerhodopsine 3.0-eYFP in the cell bodies and their axonal projections. In the same diagram the fiber optic implantation, to inhibit the ventroposterior thalamostriatal projections in vivo is presented. Note that the optogenetic manipulation was bilateral. (C) Diagram of the operant box to show that in this experiment first lever press in the sequence was used as trigger for inhibition of the thalamostriatal projection during the execution. (D) Coronal brain slices representing the fiber optic track tips from the animals used in this experiment. (E) Number of lever presses in the sequence for animals expressing Arch3.0-YFP; each point is the median of lever presses per animal during the trials of optogenetic inhibition (on) versus the trials without optogenetic inhibition (off) [off: 2.9±0.2 vs. on: 2.9±0.3; p>0.05 n=6 animals, Wilcoxon test]. Box plot: to further verify the effect during the optogenetic session we calculated the ratio of the trials on/off of the optogenetic session (On) against the same ratio from the same animals in a session without optogenetic inhibition (Off): VPsàDLS = 1.0±0.3. 13 Figure S10.The ventroposterior region (VPs) of the thalamus has a preferential innervation to cortex while the parafascicular region (PFs) to the striatum (for a review see Smith et al., 2014). Related to Figures 6 and 7. (A) and (C) Retrobeads injections into primary motor cortex (M1) or the somatosensory cortex (S1/S2). (B) and (D) Retrogradely labeled cells from A or C in the parafascicular region of the thalamus (PFs; upper panels) or the ventroposterior region of the thalamus (VPs; bottom panels). (E) Quantification of cells retrogradely labeled in the PFs or the VPs [n=4 animals for DMS, DLS (injections shown in Figure S1), M1 n=3 animals and n=2 animals for S1/S2). Scale bar represents 500µm in A-D. 14 Figure S11. The optogenetic inhibition of the ventroposterior thalamoàsomatosensory cortex projections (VPsàS1) does not increase the latency to start or the number of lever presses inside the sequence. Related to Figure 6 and 7. (A) AAV injection in the ventroposterior thalamic region to express Arch 3.0-eYFP, fiber optic implantation in S1. (B) Operant box diagram showing the position of an infrared beam (red dot) in the path from the magazine to the lever used to trigger the inhibition of the VPsàS1 terminals before the initiation of the sequence of lever presses. (C) As B, in this case the first lever press in the sequence was used to trigger the inhibition during the execution. (D) Representative photomicrograph showing the optical fiber track in S1 (yellow arrow, left panel) and the expression of Arch 3.0 eYFP both in S1 and the VPs thalamic region (yellow arrow, right panel). (E) Latency to initiate a sequence of lever press for animals expressing Arch3.0-YFP; each point is the median latency per animal during the trials of optogenetic inhibition (on) vs. the trials without optogenetic inhibition (off) from the same session [VPsàS1latency_on = 2.1±0.4 seconds, VPsàS1latency_off = 1.3±0.1, n= 6 animal, p>0.05, Wilcoxon test, paired plots]. The ratio of the optogenetic session (On) against the ratio from the same animals in a previous session without optogenetic inhibition (Off) is presented in the box plot [VPsàS1On/Off =1.1±0.1]. (F) Similar to E, in this case presenting data from a session of optogenetic inhibition during the execution. (G) Diagram representing the optical fiber track tips from the 6 animals used for this experiment. 15 (H) and (I) Number of lever presses for animals as explained in E and F respectively (VPsàS1lever_on= 5.4±0.4 vs. VPsàS1lever_off = 5.0±0.3, n= 6 animals, p>0.05, Wilcoxon test, paired plot) (VPsàS1lever_on= 6.2±0.7 vs. VPsàS1lever_off = 5.3±0.7, n= 6 animals p>0.05, Wilcoxon test, paired plot) and their ratio against a previous session (box plot: VPsàS1On/Off =1.1±0.1). (J) As positive control of the VPsàS1 inhibition, an increase in horizontal distance was identified when performing the same inhibition in the open field. The inhibition of S1 has been shown to contribute to the motor adaptation in mice (Mathis et al., 2017). (K) and (L) To evaluate if the lack of effect in F and I is due to a ceiling effect, a linear regression between the magnitude of the effect and the baseline number of lever press was performed both in the inhibition of the VPsàS1 and on the inhibition of the VPsàDLS projections. Note that only the VPsàDLS inhibition showed a significant regression. 16 Figure S12. The inhibition of the parafascicular thalamostriatal or the ventroposterior thalamostriatal projections does not present changes on the horizontal displacement in the open field. Related to Figures 6 and 7. (A) Example of a mouse in the open field for the tracking of the displacement of the animal (red dot) during light inhibition (green square). (B) Trajectory of the animal with optogenetic inhibition (green) and without (gray). (C) Normalized distance displacement aligned to light stimulation in the parafascicular thalamo-striatal projections (green and orange, mean from animals expressing Arch3.0-eYFP or eYFP respectively, n=6 animals per group). (D) Normalized distance displacement aligned to light stimulation in the ventroposterior projections. Green and orange, mean from animals expressing Arch3.0-eYFP or eYFP respectively, n=6 animals per group. Black, mean from animals inhibiting the VPs-S1 projections, n=6 animals. (E) Normalized distance displacement during the 5 second of inhibition. * p<0.05 Kruskal Wallis test. 17 PFsàDMS or DLS (AAV ChR2 injection à recording) TTX + 4AP VPsàDMS or DLS (AAV ChR2 injection à recording) TTX + 4AP ChR2 Expression (days) # Slice per animal Recorded site (cell type) EPSC (pA) Latency (ms) ChR2 Expression (days) # Slice per animal Recorded site (cell type) EPSC (pA) Latency (ms) Mouse 1 (14 days) slice 1 DMS (SPN) 48,0 5,5 Mouse 1 (15 days) slice 1 DMS (SPN) 54,8 5,7 DLS (SPN) 56,4 7.0 DLS (SPN) 300,7 5,3 slice 2 DMS (SPN) 90,8 5,2 slice 2 DMS 36,4 4,7 DLS 37,5 5,2 DLS (Int) 145,2 4,2 slice 3 DMS 164,5 5,5 Mouse 2 (8 days) slice 1 DMS 59,9 8,0 DLS (Int) 196,9 4,2 DLS (SPN) 52,4 4,8 Mouse 2 (16 days) slice1 DMS (SPN) 115,1 5,5 Mouse 3 (9 days) slice 1 DMS 0,0 DLS 196,8 5,1 DLS (SPN) 18,1 6.0 slice2 DMS (SPN) 273,9 5.0 Mouse 4 (10 days) slice 1 DMS 16,0 7,5 DLS (SPN) 105,8 4,7 DLS (SPN) 7,1 6,7 slice3 DMS (Int) 248,8 3,7 Mouse 5 (16 days) slice 1 DMS (SPN) 31,0 6,7 DLS (Int) 13,4 7,2 DLS (SPN) 57,8 7,1 Mouse 3 (10 days) slice1 DMS (SPN) 23,6 9.0 Mouse 6 (12 days) slice 1 DMS 0,0 DLS 0,0 DLS (SPN) 31,6 5,8 Mouse 4 (11 days) slice 1 DMS (SPN) 16,4 6.2 slice 2 DMS (SPN) 0,0 DLS 0,0 DLS (SPN) 54,8 4,8 slice 2 DMS 177,3 6.0 DLS (SPN) 121,7 6,6 Mouse 5 (15 days) slice1 DMS 135.0 5,6 DLS (SPN) 29.0 7,7 Mouse 6 (13 days) slice 1 DMS (SPN) 367,6 3,2 DLS 0,0 slice 2 DMS (SPN) 59,6 4,2 DMS (SPN) 82,5 3,9 Mouse 7 (15 days) slice 1 DMS 58,1 5,5 DLS (Int) 0,0 slice 2 DMS (SPN) 453,3 3,3 DLS 225,6 4,1 slice 3 DMS (SPN) 87,7 3,6 Table S1. Details for the ex vivo recordings. Related to Figure 1. Parafascicular (PFs) or ventroposterior region (VPs) of the thalamus. Channelrhodopsin (ChR2), dorsomedial striatum (DMS), dorsolateral striatum (DLS), spiny projection neuron (SPN), striatal interneuron (Int). 18 Regression (aligning the activity) Spikes-Latency (start of sequence) Spikes –Presses (start of sequence) Spikes-Presses (end of sequence) PFsàDMS 34 % (30 of 86) 83 % (25 of 30) 70 % (21 of 30) 50 % (15 of 30) VPsàDLS 12 % (15 of 122) 67 % (10 of 15) 100 % (15 of 15) 66 % (10 of 15) Time related to event (-2 to 0 seconds) (-2 to 2 seconds) (-2 to 0 seconds) Fisher test P=0.2629 P= 0.0195 * P=0.3517 Table S2. Proportion of photo identified thalamoàstriatal units presenting significative regressions between the activity and the latency to start a sequence or the number of presses in the sequence (p<0.05). Related to Figure 5. Regression (aligning the activity) Spikes-Latency (start of sequence) Spikes-Latency & Spikes-Presses (start of sequence) Spikes –Presses (start of sequence) PFsàDMS 34 % (30 of 86) 22 % (6 of 27) 70.5 % (19 of 27) 7.5 % (2 of 27) VPsàDLS 12 % (15 of 122) 0 % (0 of 15) 67 % (10 of 15) 33 % (5 of 15) Time related to event (-2 to 0 seconds) (-2 to 0 seconds) (-2 to 2 seconds) (-2 to 2 seconds) Fisher test P=0.07 P=0.73 P=0.04 * Table S3. Proportion of photo identified thalamoàstriatal units presenting selective regressions to latency or presses or both during the initiation/execution of sequences. Related to Figure 5. 19 PFs àDMS PFsàDLS VPsàDMS VPsàDLS Fibers AP: 0.0 -0.5 mm ML: 1.0-1.8 mm DV: 2.2-2.8 mm (0.3mm 3 ) AP:0.0-0.5 mm ML: 2.2-2.9 mm DV: 2.2-2.7 mm 0.2mm 3 ) AP:0.0-0.5mm ML: 1.1-1.8 mm DV: 2.2-2.8mm (0.2mm 3 ) AP: 0.0-0.5 mm ML: 2.0-2.8 mm DV: 2.2-2.7 mm (0.2mm 3 ) Ex vivo AP: -0.2-0.9 mm ML: 1.0:1.8 mm DV: 2.2-2.6 mm (0.3mm 3 ) AP: -0.2-0.9 mm ML: 2.0:2.7 mm DV: 2.2-2.7 mm (0.3 mm 3 ) AP: -0.2-0.7 mm ML: 1.0:1.8 mm DV: 2.0-2.6 mm (0.3 mm 3 ) AP: -0.2-0.7 mm ML: 2.0-2.7 mm DV: 2.2-2.7 mm (0.3 mm 3 ) Optogenetic inhibitions AP: -0.2-0.7 mm ML: 1.5-1.9 mm DV: 2.2-2.7 mm (0.5 mm 3 ) AP: 0.0 -0.8 mm ML: 2.3-3.0 mm DV: 2.2-2.6 mm (0.5 mm 3 ) AP: -0.2-0.8 mm ML: 2.3-3,0 mm DV: 2.2-2.6 mm (0.5 mm 3 ) Table S4. Ranges for the position of sampling fibers the cells recorded in the ex vivo experiments and the optic fiber tips in the optogenetic inhibition. Related to Figure 1, 6 and 7. In parenthesis are the volumes calculated from these ranges.