Thalamic subnetworks as units of function

Abstract

The thalamus engages in various functions including sensory processing, attention, decision making and memory. Classically, this diversity of function has been attributed to the nuclear organization of the thalamus, with each nucleus performing a well-defined function. Here, we highlight recent studies that used state-of-the-art expression profiling, which have revealed gene expression gradients at the single-cell level within and across thalamic nuclei. These gradients, combined with anatomical tracing and physiological analyses, point to previously unappreciated heterogeneity and redefine thalamic units of function on the basis of unique input-output connectivity patterns and gene expression. We propose that thalamic subnetworks, defined by the intersection of genetics, connectivity and computation, provide a more appropriate level of functional description; this notion is supported by behavioral phenotypes resulting from appropriately tailored perturbations. We provide several examples of thalamic subnetworks and suggest how this new perspective may both propel progress in basic neuroscience and reveal unique targets with therapeutic potential.

Fig. 1:. Distinct thalamic nuclei in rodents.

Excitatory nuclei can be grouped into seven major nuclear divisions, namely anterior, medial, lateral, ventral, intralaminar, midline, and posterior groups, based on Jones. TRN represents the major inhibitory nucleus. Anterior to posterior (AP) coronal sections.

Fig. 2:. From thalamic nuclei to subnetworks.

Relationship between individual nuclei, cell types, subpopulations, and subnetworks.

Fig. 3:. Diversity of thalamic neurons.

a, Golgi staining led to the identification of three putative types of thalamic neurons based on cellular morphology^,, two excitatory populations (bushy, radiate) and one inhibitory population (interneuron). Drawings are taken from Jones. b-c, Slice recordings found that membrane (Property 1) and synaptic properties (Property 2) vary along a systematic gradient from medial to lateral thalamic nuclei, based on data from Phillips et al. (b). These medial to lateral electrophysiological differences were linked to graded transcriptional profiles, based on data from Phillips et al. (c).

Fig. 4:. Six genes can be used combinatorially to define thalamic nuclei.

Analysis of the Allen Brain Institute in situ hybridization database suggested that thalamic nuclei can be divided into nine subgroups based on similarities in gene expression, based on data from Nagalski et al.. Six transcription factors were identified that combinatorially defines most thalamic nuclei. Anterior to posterior (AP) coronal sections.

Fig. 5:. Single-cell heterogeneity of thalamic neurons.

a, Using the Drop-seq method, two major cell types (Rora⁺ or Gad2⁺) were found in mouse thalamus, based on data from Saunders et al.. Each of these cell types can be subdivided into eleven putative subpopulations, with distinct marker genes. b, Projection-based scRNA-seq showed that thalamic nuclei form five major subpopulations, based on data from Phillips et al.. AD and RE each represented one subpopulation, while other nuclei followed a topographical arrangement (medial, intermediate, and lateral groups). Anterior to posterior (AP) coronal sections.

Fig. 6:. Subnetworks in the thalamus.

a-b, Distinct subpopulations in PVT have been identified along the anterior to posterior axis, based on data from Gao et al.. Gal is a marker for anterior PVT (aPVT) neurons, whereas D2 is a marker for posterior PVT (pPVT) neurons (a). These PVT subpopulations have different functional roles, stimulus salience (aPVT) vs. stimulus valence (pPVT) and form independent thalamo-corticothalamic loops with prefrontal cortex (b). c-e, scRNA-seq of mouse TRN neurons identified heterogeneity, which exhibits a transcriptomic gradient of two negatively correlated profiles, based on data from Li et al.. Neurons in the extremes of this gradient express distinct molecular markers (Spp1, Ecel1), and these subpopulations show distinct spatial localization within TRN in terms of core vs. shell-like patterns (c). These TRN subpopulations have distinct single-cell electrophysiological features, such as rebound bursts (d), and projections to first order vs. higher-order thalamic nuclei (e).

Fig. 7:. From scRNA-seq to three distinct PF subpopulations.

a-c, Based on projections to different striatal regions, three PF subpopulations were identified (medial/m, central/c, lateral/l), based on data from Mandelbaum et al.. scRNA-seq revealed distinct transcriptional profiles for these PF subpopulations (a). Single-cell electrophysiological properties varied in a systematic manner from medial to lateral subpopulations (Property 1 corresponds to membrane capacitance and Property 2 corresponds to excitability), based on data from Mandelbaum et al. (b). These PF subpopulations have different input-output connectivity patterns, suggestive of distinct functional roles, based on data from Mandelbaum et al. (c).