Brain-wide mapping of inputs to the mouse lateral posterior (LP/Pulvinar) thalamus-anterior cingulate cortex network

Abstract

The rodent homolog of the primate pulvinar, the lateral posterior (LP) thalamus, is extensively interconnected with multiple cortical areas. While these cortical interactions can span the entire LP, subdivisions of the LP are characterized by differential connections with specific cortical regions. In particular, the medial LP has reciprocal connections with frontoparietal cortical areas, including the anterior cingulate cortex (ACC). The ACC plays an integral role in top-down sensory processing and attentional regulation, likely exerting some of these functions via the LP. However, little is known about how ACC and LP interact, and about the information potentially integrated in this reciprocal network. Here, we address this gap by employing a projection-specific monosynaptic rabies tracing strategy to delineate brain-wide inputs to bottom-up LP→ACC and top-down ACC→LP neurons. We find that LP→ACC neurons receive inputs from widespread cortical regions, including primary and higher order sensory and motor cortical areas. LP→ACC neurons also receive extensive subcortical inputs, particularly from the intermediate and deep layers of the superior colliculus (SC). Sensory inputs to ACC→LP neurons largely arise from visual cortical areas. In addition, ACC→LP neurons integrate cross-hemispheric prefrontal cortex inputs as well as inputs from higher order medial cortex. Our brain-wide anatomical mapping of inputs to the reciprocal LP-ACC pathways provides a roadmap for understanding how LP and ACC communicate different sources of information to mediate attentional control and visuomotor functions.

Keywords: brain mapping; frontal cortex; lateral thalamic nuclei; neural pathways; pulvinar; superior colliculi.

FIGURE 1

Anatomical gradients of reciprocal LP→ACC circuits. (a1) Retrograde labeling from ACC with fluorophore‐conjugated CTB. (a2) Injection site in ACC. (a3 and a4) CTB retrogradely labeled neurons in LP found in medial subregions. (b1) Anterograde labeling of LP axons in ACC, with injection in medial LP. (b2–b4) LP axons in ACC spanning (b2) +0.15, (b3) +0.6, and (b4) +1.2 mm anterior from Bregma. (c1) Retrograde labeling of LP‐projecting ACC neurons, with retrograde AAV injection in medial LP. (c2–c4) ACC neurons projecting to LP spanning (c2) +0.15, (c3) +0.6, and (c4) +1.2 mm anterior from Bregma. (d1) Retrograde labeling of LP‐projecting ACC neurons, with CTB injection in medial LP. ACC neurons projecting to LP at (d2) +0 and (d3) +1.0 mm anterior from Bregma. (e) Quantification of ACC→LP projecting neurons along the rostrocaudal axis shows peak projection density at about 0.75–0.85 mm anterior to Bregma. Bars represent mean of n = 3 mice. Individual lines represent samples from each mouse

FIGURE 2

Whole‐brain input tracing to ACC→LP neurons and LP→ACC neurons. (a) Rabies viral tracing strategy for ACC→LP neurons involved retrograde transport of Cre recombinase injected in LP. Helper AAVs were injected in ACC for Cre‐dependent expression of rabies helper proteins. After a week, G‐deleted rabies virus was injected in ACC. (b) Starter cell region (ipsilateral ACC) for ACC→LP input mapping. (c) Rabies viral tracing strategy for LP→ACC neurons. Retrograde‐AAV carrying Cre recombinase was injected in ACC, while helper viruses were injected in LP. G‐deleted rabies virus was injected in LP a week after AAV injections. (d) Starter cell region (ipsilateral LP) for LP→ACC input mapping. (e) Overview of brain regions projecting to ACC→LP (blue) and LP→ACC (magenta) neurons. Bars represent mean of n = 3 mice, error bars show the standard deviation, and circles represent samples from individual mice

FIGURE 3

Control experiments lacking G complementation or lacking cre recombinase for rabies tracing of inputs. (a–c) No G‐complementation control for mapping inputs to ACC→LP neurons. (a) Quantification of rabies‐infected (mCherry‐positive cells) in ACA region in the experiment (+ G, n = 3) and control (– G, n = 2). Error bars indicate standard deviation. (b1) Labeled cells found in ACA and MOs but not in adjacent cortical regions. Scale bar = 250 μm. (b2) Magnified region in (b1). Scale bar = 125 μm. Magnified region of (b1) and (b2) with individual fluorescent channels: (b3) mCherry (rabies‐infected) (b4) BFP, and (b5) GFP (TVA). Without G protein complementation, there were no cells found in regions other than ACA and MOs, including (c1) LP, (c2) RSP, and (c3) SC. (d–f) No retroAAV‐Cre Control for mapping inputs to ACC→LP neurons. (d) Quantification of rabies‐infected (mCherry‐positive cells) in ACA region in the experiment (+ retro‐cre, n = 3) and control (– retro‐cre, n = 2). Error bars indicate standard deviation. Very few cells were found to have cre‐independent labeling with low if any leak expression of cre‐dependent transgenes. (e1) Micrographs showing a labeled cell found in ACAd. Scale bar = 500 μm. (e2) Magnified region in (e1). Scale bar = 250 μm. (f1 and f2) No labeled cells found in adjacent cortical regions and ACC regions. (g–i) No G‐complementation control for mapping inputs to LP→ACC neurons. (g) Quantification of rabies‐infected (mCherry‐positive cells) in LP in the experiment (+ G, n = 3) and control (– G, n = 2). Error bars indicate standard deviation. (h1) Labeled cells found in LP but not in any other regions. Scale bar = 500 μm. (h2) Magnified region in (h1). Scale bar = 250 μm. Magnified region of (h1) and (h2) with individual fluorescent channels: (h3) mCherry (rabies‐infected), (h4) BFP, and (h5) GFP (TVA). Without G protein complementation, there were no cells found in regions other than LP, including (i1) ACA and MOs, (i2) SC and (i3) RSP. (j–l) No retroAAV‐Cre Control for mapping inputs to LP→ACC neurons. (j) Quantification of rabies‐infected (mCherry‐positive cells) in LP in the experiment (+ retro‐cre, n = 3) and control (– retro‐cre, n = 2). Error bars indicate standard deviation. No cells were found to have cre‐independent labeling with low if any leak expression of cre‐dependent transgenes, (k) no rabies‐infected cells in all regions shown in section. Scale bar = 500 μm. (l1) No rabies‐infected cells found in mLP. Scale bar = 500 μm. (l2) Magnified region in (l1). Scale bar = 250 μm

FIGURE 4

Frontal, prefrontal, and retrosplenial inputs to ACC→LP neurons. (a) Micrographs showing inputs from dorsal ACC, PL, ORB areas, and the agranular insula cortices. (b and c) Inputs to ACC→LP neurons include contralateral ACC, particularly ventral ACC. (d) Comparison of proportions of cells found in each hemisphere in ACA subregions and MOs, the values are taken as a percentage to all cells counted in each ACC→LP experimental animal. Bars represent mean of n = 3 mice, error bars indicate the standard deviation, and circles represent samples from individual mice. (E and f) Retrosplenial cortex inputs to ACC→LP neurons can arise from all retrosplenial subregions. Scale bars = 250 μm

FIGURE 5

Cortical inputs to ACC→LP and LP→ACC projectors. (a and b) Distributions of the functional domains of cortical inputs to ACC→LP (A) and LP→ACC (B) neurons. (c and d) Cortical inputs to ACC→LP (c) and LP→ACC (d) neurons sorted by functional domains. These inputs are represented with a circle with radius proportional to the percentage of total inputs it represents for each layer, and also as a whole region with all layers summed (mean of n = 3 mice)

FIGURE 6

Cortical inputs to LP→ACC neurons. LP→ACC neurons receive cortical inputs from (a–d) prefrontal areas ACC, PL, and MOs, scale bars = 250 μm. (e) Temporal association cortical areas TeA and ectorhinal cortex, scale bar = 125 μm. (f) Retrosplenial areas, scale bar = 125 μm. (g) Auditory cortex and (h) somatosensory cortex, scale bars = 250 μm

FIGURE 7

Visual cortical inputs to ACC→LP and LP→ACC projectors. (a–c) Micrographs showing inputs to ACC→LP neurons (a) VISp, (b) VISa and VISam, (c) VISp and VISpm. (d–f) Micrographs showing inputs to LP→ACC neurons (a) VISp and VISrl, (b) VISa, and (c) VISam. Scale bars = 125 μm. (g and h) Distribution of inputs from different visual cortical areas to ACC→LP (g) and LP→ACC (h) projectors

FIGURE 8

Subcortical inputs to ACC→LP neurons. (a and b) Inputs from the lateral septal area. (c1, c2, and c3) Claustral inputs span the anterior–posterior axis. Scale bars = 100 μm. (d–g) Thalamic inputs to ACC→LP neurons are largely from the anterior thalamic nuclear group, submedial thalamus (SMT), and lateral group. Scale bars (except (c)) = 250 μm

FIGURE 9

Tectal and pretectal midbrain inputs to LP→ACC neurons. (a) Distribution of midbrain and hypothalamic inputs to LP→ACC neurons. (b) Micrograph showing laminar distribution of superior colliculus inputs to LP→ACC neurons and (c) its quantification. Bars represent mean of n = 3 mice, and circles represent samples from individual mice. Scale bars = 250 μm

FIGURE 10

Midbrain and hindbrain inputs to LP→ACC neurons. Inputs to LP→ACC neurons from (a) ZI, LGv, and the APN, (b) pretectal areas, (c) SC and reticular formation. (d) Superior and inferior colliculi and the cholinergic laterodorsal tegmentum (LDT). Scale bars = 250 μm. (e) Periaqueductal gray (PAG), (f) midbrain reticular nucleus (MRN), (g) parabrachial nucleus (PB), and (h) pontine reticular nucleus (PRN). Scale bars = 100 μm

FIGURE 11

Inhibitory (Vgat1‐positive) inputs to LP. (a) Retrograde labeling of inhibitory inputs to LP, with injection of AAV carrying Cre‐dependent fluorescent tracer in LP of a Vgat1‐Cre mouse. (b) Vgat1‐positive axons in LP, and cell bodies in LGv. (c) Intrathalamic inhibition from the thalamic reticular nucleus (RT). (d) Vgat1‐positive cell bodies in ipsilateral APN and LGv. Sparse interneuronal labeling in LP and LGd. Vgat1‐positive axons in contralateral thalamus. (e) Vgat1‐positive inputs from LGv. (f) Sparse inhibitory inputs from both hemispheres of the nucleus of the posterior commisure (NPC). (g) Vgat1‐positive inputs from the anterior pretectal nucleus (APN). (h) Inhibitory inputs from the zona incerta. (i) Vgat1‐positive axons but not cell bodies in ipsilateral SC. Scale bars = 250 μm

FIGURE 12

Overview of inputs to ACC→LP and LP→ACC neurons. Blue arrows represent inputs to ACC→LP neurons. Magenta arrows represent inputs to LP→ACC neurons. Weight of the arrows reflects projection‐specific relative proportion of inputs (not to scale) and should only be compared within projections (blue for ACC →LP and magenta for LP→ACC)