The rostral intralaminar nuclear complex of the thalamus supports striatally mediated action reinforcement
Curation statements for this article:-
Curated by eLife
eLife assessment
Cover et al., examine the pathway from the intralaminar nucleus of the thalamus (rILN) to the dorsal striatum (DS) in the reinforcement of behavior/actions. The rILN sends a large glutamatergic projection to the DS, but its role in action selection was unknown. The authors found that the rILN neurons that project to the DS were activated at both action initiation and with the reward. Activation and inhibition of this pathway increased the success or decreased the success of reward acquisition, respectively. The findings are an important advance our understanding of the function of rILN to DS projection in reward-based behavior. The manuscript has provided convincing evidence with the appropriate methodologies to support these claims.
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
- Neuroscience (eLife)
Abstract
The dorsal striatum (DS) mediates the selection of actions for reward acquisition necessary for survival. Striatal pathology contributes to several neuropsychiatric conditions, including aberrant selection of actions for specific rewards in addiction. A major source of glutamate driving striatal activity is the rostral intralaminar nuclei (rILN) of the thalamus. Yet, the information that is relayed to the striatum to support action selection is unknown. Here, we discovered that rILN neurons projecting to the DS are innervated by a range of cortical and subcortical afferents and that rILN→DS neurons stably signaled at two time points in mice performing an action sequence task reinforced by sucrose reward: action initiation and reward acquisition. In vivo activation of this pathway increased the number of successful trials, whereas inhibition decreased the number of successful trials. These findings illuminate a role for the rostral intralaminar nuclear complex in reinforcing actions.
Article activity feed
-
-
Author Response
Reviewer #1 (Public Review):
In this manuscript, Cover et al. examine the role of thalamic neurons of the rostral intralaminar nuclei (rILN) that project to the dorsal striatum (DS) in mice performing a reinforced action sequence task. Using patch-clamp electrophysiology, they find that neurons from the three rILN (CM, PC, and CL) have similar electrophysiological properties. Using fiber photometry recordings of calcium activity from rILN neurons that project to DS, they show that these neurons increase in activity at the first lever press and reward acquisition in mice performing a lever pressing operant task. They additionally demonstrate that this action initiation and reward-related activity exists more generally in mice performing other movements or rewarded tasks. Building on their lab's previous work, the …
Author Response
Reviewer #1 (Public Review):
In this manuscript, Cover et al. examine the role of thalamic neurons of the rostral intralaminar nuclei (rILN) that project to the dorsal striatum (DS) in mice performing a reinforced action sequence task. Using patch-clamp electrophysiology, they find that neurons from the three rILN (CM, PC, and CL) have similar electrophysiological properties. Using fiber photometry recordings of calcium activity from rILN neurons that project to DS, they show that these neurons increase in activity at the first lever press and reward acquisition in mice performing a lever pressing operant task. They additionally demonstrate that this action initiation and reward-related activity exists more generally in mice performing other movements or rewarded tasks. Building on their lab's previous work, the authors further find that by optogenetically activating or inhibiting these rILN-DS neurons, mice will increase or decrease task performance, respectively. Lastly, the authors show that a variety of cortical and subcortical areas have input to rILN-DS neurons suggesting that these neurons might act as an integrator of signals from such areas during task performance.
• The authors beautifully show that the electrophysiological properties of CM, PC, and CL neurons are similar and go on to treat the rILN as one homogenous nucleus for functional fiber photometry recordings and optogenetic stimulations. It seems that these recordings and stimulations were only performed in CL, as indicated in the images (Fig. 2A, 4A). Is this the case, or were CM, PC, and CL neurons sampled? It would be helpful to clarify if DS projecting neurons from all rILN nuclei show the reported action initiation and reward acquisition activity or only CL neurons.
The arrangement of the rILN nuclei presents a technical challenge for experiments attempting to selectively record from or manipulate a single nucleus in this grouping. Based on our findings that the three nuclei do not differ in electrophysiological properties, we approached the in vivo experiments with the intent to target the rILN as a unit. As the reviewer points out, the medial-lateral coordinate for optic fiber placement tended to align above the CL and PC nuclei. However, variability in fiber placement and spread of light within tissue resulted in inclusion of CM activity as well. Given the spread of light through tissue (Shin, et al., 2016; PMID: 27895987), it would be very difficult to confidently determine from histology which photometry recordings were primarily obtained from CL vs PC vs CM neuronal activity. We agree with the reviewer that these three nuclei may differently signal during reward-driven behavior. Our di-synaptic tracing study supports this possibility as it revealed unique afferent connectivity to rILNDS projecting neurons. We now mention this limitation of our approach in the discussion (lines 324 - 330).
• Along similar lines, to what extent of rILN was targeted for optogenetic activation and inhibition? It seems that the authors implanted a total of 4 optic fibers, two on each side (please clarify in methods). What was the reasoning behind this? Please show that only rILN and not PF was activated/inhibited.
We apologize for the confusion in our description of this method. For our optogenetic experiments, we infused viruses at four locations (bilateral striatum and rILN) and implanted only two fibers (bilateral rILN) to selectively target striatally-projecting rILN neurons. We have added clarification on this detail to the methods section.
To prevent inadvertent modulation of Pf neurons, we used virus injection coordinates and volumes that prevented viral spread to the Pf and furthermore implanted the optic fibers in the more rostral regions of the rILN. We histologically confirmed viral expression and fiber placement for all mice and excluded any mice with viral spread to the Pf or off-target fiber placement. We include these criteria for post-hoc exclusion in the methods.
• While AAV1 is becoming a popular tool for transsynaptic labeling, performing confirmatory patch-clamp recordings with optogenetic activation of inputs, would provide better evidence for the synaptic connection between upstream regions, such as ACC and OFC, and rILN neurons.
We agree that electrophysiological confirmation of these inputs to the rILN would complement our tracing study. As our focus for this experiment was to specifically identify inputs that synapse on striatally-projecting rILN neurons, we interrogated putative afferents that were already established to project to the rILN. There are several studies that demonstrate the physiological circuits from some of these afferent projections to the rILN (without di-synaptic specificity), such as the SNr rILN projection (Rizzi & Tan, 2019; PMID: 31091455).
• In addition, the transsynaptic tracing experiments would benefit from showing the cell count quantifications in CM, PC, and CL. It seems that the authors have already performed this quantification for constructing their diagrams on the right. To make any point about the relative strength of afferent innervation to rILN-DS neurons showing such quantification would be necessary.
Thank you for this suggestion, we now include cell counts for 2 cases per investigated afferent (Supplemental Table S2).
• Why is the injection site for the retrograde cre-dependent tdTomato AAV (Fig. 5 middle left panels) showing expression? Is the cre coming through transsynaptic AAV1 from direct projections of each AAV1 injection site (AAV1 is not supposed to spread across a second synapse)? The diagrams suggest that not all regions (e.g. SUM or SC) have direct projections to DS.
We apologize for this confusion. The tdTomato fluorophore expression observed in the striatum may arise from several possible circuit configurations. To survey just a couple: 1) tdTomato expression in the DS arises from direct projections from the afferent bypassing the thalamus (e.g. ipsilateral ACC→Striatum), which would result in labeled striatal somata (ACC pyramidal neurons delivering AAV1-cre to an MSN, and those local MSN collaterals retrogradely picking up rAAV-DIO-tdtomato) and ACC labeled axon terminals in the DS (ACC interneurons delivering AAV1-cre to DS-projecting ACC pyramidal neurons that pick up rAAV-DIO-tdtomato); 2) terminal projections arising from the labeled rILN neurons shown in the middle-right panels (i.e. ACC→rILN→Striatum).
Reviewer #2 (Public Review):
This manuscript details the role of the rILN to the DS pathway in the onset of operant behavior that promotes the delivery of a reward and in the ultimate acquisition of that reward. The strengths of the paper are in the detailed fiber photometry study that encompasses several behavioral domains that correlate to the signal observed in the rILN to DS pathway. I am especially interested in how the "encoding" shifts across time as the animals refine their behavior both in a temporal sense and in the magnitude of the signal. Further, the authors demonstrate then that this is dependent on action, as they do not observe signals in a Pavlovian behavioral task, but do observe reward-based signals in a "free consumption" task (the strawberry milk). The examination into devaluation also enhances the understanding of this pathway, even though there were no differences between a valued and devalued task. Finally, the authors examine bi-directional optogenetic manipulation of the pathway, and its impact on how the trials are completed, omitted, or incomplete. They find that manipulation alters the % completed trials and regulates trial omission. This paper really does not have any glaring weaknesses to point out, however, the physiological assessment does seem to have a few strong trends and even though the studies are well powered, and included both sexes, sex as a biological variable was not commented on that I could find. My estimation of the data doesn't suggest strong sex differences in any metric measured. Additionally, the data that included projections to the rILN were very interesting, and future studies looking into the physiology of these neurons, and/or how the physiology of these neurons adapt after operant training may be very interesting to understand plasticity within the adaptation across the training from FR1 to FR5 with time limits.
Thank you for your review. We analyzed our data for sex differences but did not identify any significant differences between male and female subjects for any of the experiments.
-
eLife assessment
Cover et al., examine the pathway from the intralaminar nucleus of the thalamus (rILN) to the dorsal striatum (DS) in the reinforcement of behavior/actions. The rILN sends a large glutamatergic projection to the DS, but its role in action selection was unknown. The authors found that the rILN neurons that project to the DS were activated at both action initiation and with the reward. Activation and inhibition of this pathway increased the success or decreased the success of reward acquisition, respectively. The findings are an important advance our understanding of the function of rILN to DS projection in reward-based behavior. The manuscript has provided convincing evidence with the appropriate methodologies to support these claims.
-
Reviewer #1 (Public Review):
In this manuscript, Cover et al. examine the role of thalamic neurons of the rostral intralaminar nuclei (rILN) that project to the dorsal striatum (DS) in mice performing a reinforced action sequence task. Using patch-clamp electrophysiology, they find that neurons from the three rILN (CM, PC, and CL) have similar electrophysiological properties. Using fiber photometry recordings of calcium activity from rILN neurons that project to DS, they show that these neurons increase in activity at the first lever press and reward acquisition in mice performing a lever pressing operant task. They additionally demonstrate that this action initiation and reward-related activity exists more generally in mice performing other movements or rewarded tasks. Building on their lab's previous work, the authors further find …
Reviewer #1 (Public Review):
In this manuscript, Cover et al. examine the role of thalamic neurons of the rostral intralaminar nuclei (rILN) that project to the dorsal striatum (DS) in mice performing a reinforced action sequence task. Using patch-clamp electrophysiology, they find that neurons from the three rILN (CM, PC, and CL) have similar electrophysiological properties. Using fiber photometry recordings of calcium activity from rILN neurons that project to DS, they show that these neurons increase in activity at the first lever press and reward acquisition in mice performing a lever pressing operant task. They additionally demonstrate that this action initiation and reward-related activity exists more generally in mice performing other movements or rewarded tasks. Building on their lab's previous work, the authors further find that by optogenetically activating or inhibiting these rILN-DS neurons, mice will increase or decrease task performance, respectively. Lastly, the authors show that a variety of cortical and subcortical areas have input to rILN-DS neurons suggesting that these neurons might act as an integrator of signals from such areas during task performance.
• The authors beautifully show that the electrophysiological properties of CM, PC, and CL neurons are similar and go on to treat the rILN as one homogenous nucleus for functional fiber photometry recordings and optogenetic stimulations. It seems that these recordings and stimulations were only performed in CL, as indicated in the images (Fig. 2A, 4A). Is this the case, or were CM, PC, and CL neurons sampled? It would be helpful to clarify if DS projecting neurons from all rILN nuclei show the reported action initiation and reward acquisition activity or only CL neurons.
• Along similar lines, to what extent of rILN was targeted for optogenetic activation and inhibition? It seems that the authors implanted a total of 4 optic fibers, two on each side (please clarify in methods). What was the reasoning behind this? Please show that only rILN and not PF was activated/inhibited.
• While AAV1 is becoming a popular tool for transsynaptic labeling, performing confirmatory patch-clamp recordings with optogenetic activation of inputs, would provide better evidence for the synaptic connection between upstream regions, such as ACC and OFC, and rILN neurons.
• In addition, the transsynaptic tracing experiments would benefit from showing the cell count quantifications in CM, PC, and CL. It seems that the authors have already performed this quantification for constructing their diagrams on the right. To make any point about the relative strength of afferent innervation to rILN-DS neurons showing such quantification would be necessary.
• Why is the injection site for the retrograde cre-dependent tdTomato AAV (Fig. 5 middle left panels) showing expression? Is the cre coming through transsynaptic AAV1 from direct projections of each AAV1 injection site (AAV1 is not supposed to spread across a second synapse)? The diagrams suggest that not all regions (e.g. SUM or SC) have direct projections to DS.
-
Reviewer #2 (Public Review):
This manuscript details the role of the rILN to the DS pathway in the onset of operant behavior that promotes the delivery of a reward and in the ultimate acquisition of that reward. The strengths of the paper are in the detailed fiber photometry study that encompasses several behavioral domains that correlate to the signal observed in the rILN to DS pathway. I am especially interested in how the "encoding" shifts across time as the animals refine their behavior both in a temporal sense and in the magnitude of the signal. Further, the authors demonstrate then that this is dependent on action, as they do not observe signals in a Pavlovian behavioral task, but do observe reward-based signals in a "free consumption" task (the strawberry milk). The examination into devaluation also enhances the understanding of …
Reviewer #2 (Public Review):
This manuscript details the role of the rILN to the DS pathway in the onset of operant behavior that promotes the delivery of a reward and in the ultimate acquisition of that reward. The strengths of the paper are in the detailed fiber photometry study that encompasses several behavioral domains that correlate to the signal observed in the rILN to DS pathway. I am especially interested in how the "encoding" shifts across time as the animals refine their behavior both in a temporal sense and in the magnitude of the signal. Further, the authors demonstrate then that this is dependent on action, as they do not observe signals in a Pavlovian behavioral task, but do observe reward-based signals in a "free consumption" task (the strawberry milk). The examination into devaluation also enhances the understanding of this pathway, even though there were no differences between a valued and devalued task. Finally, the authors examine bi-directional optogenetic manipulation of the pathway, and its impact on how the trials are completed, omitted, or incomplete. They find that manipulation alters the % completed trials and regulates trial omission. This paper really does not have any glaring weaknesses to point out, however, the physiological assessment does seem to have a few strong trends and even though the studies are well powered, and included both sexes, sex as a biological variable was not commented on that I could find. My estimation of the data doesn't suggest strong sex differences in any metric measured. Additionally, the data that included projections to the rILN were very interesting, and future studies looking into the physiology of these neurons, and/or how the physiology of these neurons adapt after operant training may be very interesting to understand plasticity within the adaptation across the training from FR1 to FR5 with time limits.
-
Reviewer #3 (Public Review):
The manuscript by Cover et al. follows up on their recent work examining a poorly characterized connection from nuclei in the rostral intralaminar thalamus to the dorsal striatum. Their previous work demonstrated that mice self-administer optogenetic activation of this pathway, which promotes dopamine release in the striatum (in a multi-synaptic fashion).
In terms of thalamostriatal connectivity, there has been a greater focus on the more robust striatal inputs from the center median and parafascicular thalamic nuclei. Notably, the rostral intralaminar thalamic inputs are thought to be morphologically distinct from their parafascicular counterparts in that they have stronger thalamocortical projections, their axons preferentially synapse on the spines of striatal output neurons (as opposed to the dendritic …
Reviewer #3 (Public Review):
The manuscript by Cover et al. follows up on their recent work examining a poorly characterized connection from nuclei in the rostral intralaminar thalamus to the dorsal striatum. Their previous work demonstrated that mice self-administer optogenetic activation of this pathway, which promotes dopamine release in the striatum (in a multi-synaptic fashion).
In terms of thalamostriatal connectivity, there has been a greater focus on the more robust striatal inputs from the center median and parafascicular thalamic nuclei. Notably, the rostral intralaminar thalamic inputs are thought to be morphologically distinct from their parafascicular counterparts in that they have stronger thalamocortical projections, their axons preferentially synapse on the spines of striatal output neurons (as opposed to the dendritic shafts of these neurons or cholinergic interneurons), and they may relay information from the cerebellum to striatum. As such, the author's functional characterization of the striatal projection from the less understood intralaminar thalamic nuclei is an important conceptual advance.
By using projection-specific calcium imaging, the authors show that these projections activate during lever pressing or the initiation of well-learned lever-pressing sequences and during the receipt of reward. Notably, the authors found no correspondence between the expected value of the lever presses, since devaluing the rewards or extinguishing their delivery altogether had no effect on the magnitude of this pathway's activation at the time of lever pressing. Devaluation also had no impact on the magnitude of activation at the time of reward delivery.
By contrast, the magnitude of activation in this pathway did inversely correlate with the animal's latency to initiate pressing and retrieving the reward. Moreover, activity in this pathway was positively correlated to spontaneous movement in an open-field arena. In conjunction with the author's earlier study, these findings suggested this pathway could be important for goal-directed action selection. In agreement with this idea, optogenetically manipulating this pathway bi-directionally modulated performance in their lever-pressing task.
The data presented overall support the claim that this pathway is important for operant conditioning. One weakness is that the optogenetic inhibition experiments produced very small effect sizes. This could be related to the technical difficulty of inhibiting enough of these sparse projections to the striatum. Another potential drawback (related to this weakness) is an over-interpretation of the importance of this projection and the underemphasis on the importance of somatically driven dopamine release, ideas that could be better addressed in the abstract and discussion.
-
