The research in the laboratory is directed at issues of thalamic functional organization and thalamocortical relationships. We use a broad interdisciplinary approach, attempting to answer the same or closely related questions with several different techniques. More specifically, we use neuroanatomical techniques, including electron microscopy aimed at detailed circuit reconstruction (e.g., “connectomics”), to explore various circuits; we use in vitro recordings from brain slices to study circuit, cell, and synaptic properties; and we record from in vivo preparations to evaluate these circuits in awake, behaving preparations of whole animals to determine the relationship between behavioral and cognitive parameters and thalamocortical functioning.
Stimulation techniques in slices include electrical activation and laser photostimulation involving both uncaging of glutamate (or GABA) and optogenetics; recording involves mainly patching of single neurons and imaging via flavoprotein autofluorescence.
For in vivo recording, we also plan to use electrical stimulation plus newer technology involving DREADDs and optogenetics (both ChR2 to activate and NpHR or JAWS to inactivate) based on Cre lines to control specific pathways. Recording of responses involves both electrophysiology (single cell and current source density analysis) plus imaging using two photon Ca2+ signals, which allows us to monitor hundreds of cell responses simultaneously. This also enables us to investigate circuit properties of the recorded areas.
Our laboratory is engaged in two active collaborations: one with the laboratory of Jason MacLean in studies involving two photon Ca2+ imaging; and the other with the laboratory of Bobby Kasthuri for the electron microscopy.
Recent reviews are available on another part of this website (Halassa and Sherman 2019;Sherman 2016;Sherman and Guillery 2013).
Drivers and Modulators
We have pointed out that not all afferents to thalamic relay cells are equal, and that it is important to identify which input carries the information to be relayed. Examples of the information-bearing afferents are the retinal input to the lateral geniculate nucleus and medial lemniscal input to the ventral posterior nucleus. These we call the driver inputs. All other inputs to any given thalamic relay, such as those from layer 6 of cortex and the brainstem (e.g., cholinergic, noradrenergic, etc.), are modulator inputs and determine how driver inputs are processed. Included in this modulatory role is the control of firing mode and its switching between tonic and burst firing plus gain control of the driver input. We have generated a list of features that distinguish glutamatergic drivers from modulators and have further shown that this duality of inputs types can be applied to other pathways, such as those in cortex.
We have further developed this classification for glutamatergic afferents, because these are often considered to be the information bearing inputs to circuits and are generally dealt with as if they are equal participants in some sort of anatomical democracy. However, the example from thalamus makes clear that this is incorrect: the driver input (e.g., retinal) is functionally quite distinct from the layer 6 modulatory feedback, and yet both are glutamatergic. Thus many (indeed, it appears that most) glutamatergic afferents in thalamus and cortex are modulatory, much like cholinergic or noradrenergic afferents. However, there is an important distinction: these other, “classic” modulatory systems tend to have poor topography and generally underlie global effects related to overall behavioral state; glutamatergic modulators are highly topographic and thus provide the sort of local modulation needed for such processes as spatial attention, adaptation, etc.
Now that we have extended this classification for glutamatergic inputs to cortex, we find that, based on a range of physiological and anatomical criteria, the same two synaptic input types are found in both thalamus and cortex. We hypothesize the same function in cortex for these two glutamatergic input types as in thalamus: driver inputs carry the basic information in circuits, whereas modulator inputs provide highly topographical modulation of driver processing.
First and Higher Order Order Relays
Because driver input determines the nature of a thalamic relay (i.e., the lateral geniculate nucleus can be characterized as a relay of retinal input), identifying the driver input to a thalamic relay is a key first step in understanding the function of that relay. In the process of doing so, we realized that the driver input to many thalamic relays originates in layer 5 of cortex. This is different from the corticothalamic pathway emanating from layer 6: all thalamic relays receive a layer 6 input, and this is largely feedback and modulatory, while only some receive an additional layer 5 input, and this is often feedforward and driver (although we recently have identified feedback configurations for layer 5 corticothalamic input as well; see below). Thus those with layer 5 input are the thalamic limb of a cortico-thalamo-cortical or transthalamic pathway, being a key link in a chain of corticocortical communication.
Those thalamic relays receiving driver input from a subcortical source are first order relays, because they represent the first relay of a particular form of information (e.g., visual or somatosensory) to cortex. Those that receive a driver input from layer 5 of cortex are higher order relays, because they relay information already in cortex back into cortex. Examples of first order relays for vision, somesthesis, and hearing are, respectively, the lateral geniculate nucleus, the ventral posterior nucleus, and the ventral part of the medial geniculate nucleus; their higher order partners are the pulvinar, posterior medial nucleus, and dorsal part of the medial geniculate nucleus. We have also identified other first and higher order relays, and it appears that most of thalamus by volume is higher order. This provides a straightforward function of relaying information between cortical areas for most of thalamus that hitherto seemed quite mysterious in terms of their roles.
This idea that higher order relays play a key role in corticocortical communication challenges the dogma that this communication is solely the result of direct corticocortical connections. A curious feature that has emerged is that often, and perhaps always, cortical areas connected directly also have a transthalamic connection organized in parallel. This raises a series of questions:
- Are direct connections always paralleled by a transthalamic one, or, are some cortical areas connected only by a direct or transthalamic input?
- What is the difference in terms of information content between the direct and transthalamic pathways?
- Why is one path relayed via thalamus?
- Since we now have evidence for both feedforward and feedback transthalamic pathways (see bel;ow), how do they differ functionally?
The Relationship of Thalamus to Efference Copies
We have noted that many and perhaps all driver inputs to thalamus, whether to first order or higher order targets, are branches of axons that also project to extrathalamic, subcortical motor centers, sometimes even reaching the spinal cord. Thus, for example, many or all retinal axons innervating the lateral geniculate nucleus branch to also innervate midbrain centers involved in the control of eye movements, pupil size, accommodation, etc. Furthermore, the cortical layer 5 axons innervating higher order thalamic relays branch to innervate brainstem and sometimes spinal motor centers. Because of the understanding that branching axons in mammals means that the exact same message is transported down all branches of an axon, it follows that the message sent to thalamus for relay to cortex is an exact copy of that sent to extrathalamic motor centers. Because a copy of a motor message is a neat definition of an efference copy, we conclude that, in many or most cases, the message sent to thalamus for relay to cortex can be read out as an efference copy, notifying the target cortical area of a motor message initiated by the afferent cortical area. We thus hypothesize that driver inputs to thalamus play a dual role: they provide information about the rest of the body and environment and also act as efference copies of messages sent to motor centers. In this scheme, one of the roles of transthalamic circuits is to continuously upgrade these commands and simultaneously inform higher order cortical areas of messages sent to motor centers. Other possible roles are also under study.
Extension of Examples of Transthalamic Pathways
We have recently extended the concept of transthalamic pathways in two ways. First, these had been limited to sensory pathways: primary visual cortex through pulvinar to higher visual cortical areas; primary somatosensory cortex through the posterior medial nucleus to higher somatosensory cortical areas; and primary auditory cortex through the dorsal division of the medial geniculate nucleus to higher auditory cortical areas. We recently identified parallel direct and transthalamic pathways from the primary somatosensory cortex to the primary motor cortex, the transthalamic portion relayed by the posterior medial nucleus.
Second we have identified feedback transthalamic pathways of two sorts in the somatosensory system. In one, we have identified a pathways from layer 5 of the primary somatosensory cortex to the posterior medial nucleus back to the primary somatosensory cortex. In the other, we have identified a pathway from layer 5 of the secondary somatosensory cortex to the posterior medial nucleus back to the primary somatosensory cortex. Because we have also identified the projection from the posterior medial nucleus to the primary somatosensory cortex as strictly modulatory, these feedback transthalamic circuits seem to subserve a modulatory function. This contrasts with the feedforward transthalamic circuits, which are organized as driver pathways.
We have preliminary evidence for the same feedback transthalamic pathways in the visual system.
Potential Clinical Correlates
The role of higher order thalamic relays indicated above may relate to aspects of awareness of self and, along with the idea that messages relayed by these and other thalamic nuclei constitute efference copies, contribute to forward models of self and the environment. There is a long history of cognitive defects that are associated with higher order thalamic relays, and recent evidence based on MRI and postmortem anatomy in schizophrenic patients indicates that first order nuclei appear normal but higher order nuclei (e.g., the medial dorsal nucleus and pulvinar) are severely shrunken with extensive neuronal loss. Thus pathology of higher order thalamic relays may contribute to self-generated actions not being registered as such, producing inaccurate forward models. Failures in the production of normal forward models may play a significant role in producing some of the abnormal experiences encountered in schizophrenia and perhaps other cognitive disorders.
Halassa, M.M. & Sherman, S.M. 2019. Thalamocortical Circuit Motifs: A General Framework. Neuron., 103(5), 762-770
Sherman, S.M. 2016. Thalamus plays a central role in ongoing cortical functioning. Nat Neurosci, 19, 533-541
Sherman, S.M. & Guillery, R.W. 2013. Functional Connections of Cortical Areas: A New View from the Thalamus Cambridge, MA, MIT Press.