A genetically defined tecto-thalamic pathway drives a system of superior-colliculus-dependent visual cortices

Abstract

Cortical responses to visual stimuli are believed to rely on the geniculo-striate pathway. However, recent work has challenged this notion by showing that responses in the postrhinal cortex (POR), a visual cortical area, instead depend on the tecto-thalamic pathway, which conveys visual information to the cortex via the superior colliculus (SC). Does POR's SC-dependence point to a wider system of tecto-thalamic cortical visual areas? What information might this system extract from the visual world? We discovered multiple mouse cortical areas whose visual responses rely on SC, with the most lateral showing the strongest SC-dependence. This system is driven by a genetically defined cell type that connects the SC to the pulvinar thalamic nucleus. Finally, we show that SC-dependent cortices distinguish self-generated from externally generated visual motion. Hence, lateral visual areas comprise a system that relies on the tecto-thalamic pathway and contributes to processing visual motion as animals move through the environment.

Keywords: calcium imaging; electrophysiology; extra-geniculate; higher visual areas; mouse; postrhinal cortex; pulvinar; superior colliculus; tecto-thalamic pathway; visual cortex.

Figure 1.. A Lateromedial Gradient of SC-Dependent HVAs.

(A) Top: Schematic of experimental configuration for wide-field calcium imaging of visual cortex (VCx) in response to stimuli consisting of moving dots under control condition (left) and tetrodotoxin (TTX) silencing of superior colliculus (SC; right). Bottom: Example ΔF/F (120 trial average) across visual areas before (left) and after (right) TTX injection in SC. Cortical visual areas delineated as shown in Figure S1. (POR: Postrhinal Cortex; LI: Laterointermediate Area; P: Posterior Area; LM: Lateromedial Area; AL: Anterolateral Area; AM: Anteromedial Area; PM: Posteromedial Area; V1: Primary Visual Cortex). Orientation: A: anterior; M: medial. (B) Time courses of calcium response to visual stimulus from the example experiment above. Traces are normalized to peak response in each area under control condition. Shading: SEM ;(C) Summary data for 4 mice. Y axis plots remaining visual response after TTX injection in SC, normalized by the control response for each area. Gray bar shows 95% confidence interval above and below mean (red line). Gray lines between dots link areas from the same animal. P values are calculated with paired t-test. Green data are from the example mouse shown in A and B.

Figure 2.. Visual responses of isolated units in lateral HVAs depend on SC.

(A) Schematics of experimental configuration. An optrode is used to silence neuronal activity in SC by optogenetically activating GABAergic neurons conditionally expressing ChR2 in a GAD-Cre mouse. Electrophysiological recordings are performed simultaneously in visuotopically matching regions of V1 and lateral HVAs. The visual stimulus consists of a drifting grating (diameter: 45 degrees). The recordings in the lateral HVAs were centered around LI near its borders with POR and LM. The stereotactic injection of the anterograde tracer AAV1.CAG.tdTomato (red) in V1 enabled the identification of the areas LI, POR and LM. ;(B) Summary PSTH for 42 visually responsive isolated regular spiking (RS) units recorded in lateral HVAs centered around LI of awake head-fixed mice in response to visual stimulation under control conditions (black) and during SC optogenetic silencing (blue). The gray horizontal bar illustrates the duration of stimulus presentation and the blue horizontal bar indicates the period of SC silencing. Bottom: Scatter plot for the isolated units (Firing rate (FR) average Z-scores: Laser OFF: 1.98 ± 0.22; Laser ON: 0.21 ± 0.09; P<10⁻⁷, 42 RS units, 4 mice, Wilcoxon signed-rank test). (C) Top: As in (B), but for 36 visually responsive isolated RS units recorded in the V1 simultaneously with the recordings in lateral HVAs shown in (B). Bottom: Scatter plot for the isolated units (Firing rate average Z-scores: Laser OFF: 2.29 ± 0.23; Laser ON: 1.93 ± 0.58 Hz; P=0.005, 36 RS units, 3 mice, Wilcoxon signed-rank test).

Figure 3.. A Genetically Defined Cell Type in SC Mediates Visual Responses in Lateral Visual Areas.

(A) Left: Schematic of anterograde tracing with injection of AAV-tdTomato and AAV-GFP-flex in SC of NTSR1-GN209-Cre mouse to label projections from SC to thalamus. Right: Photomicrographs of the same two coronal sections (Bregma: −2.1 mm) showing tdTomato (left) and GFP (right) expression at injection site (top) and along thalamic projections (bottom). Note the restricted projection pattern of the NTSR1-GN209-Cre expressing SC neurons to pulvinar (GFP) as compared to the broader tecto-thalamic projection pattern captured by the nonspecific expression of tdTomato in SC. Inset: Photomicrograph from a more caudal section (Bregma −2.5 mm) illustrating the stronger tecto-pulvinar projection in more caudal regions of the pulvinar. Dotted lines delineate various brain structures according to the Allen Brain Atlas. (APN: Anterior Pretectal Nucleus; LGN: Lateral Geniculate Nucleus; MB: Midbrain; Pulv: Pulvinar Nucleus; IGL: Intergeniculate Leaflet). (B) Left: Schematic of retrograde labeling of pulvinar projecting SC neurons with CTB-Alexa-488 in NTSR1-GN209-Cre mouse crossed with tdTomato reporter. Actual injection site is 1.25 mm anterior to the plane depicted in the schematic. Right: Confocal images showing expression of tdTomato reporter in SC (left panel), retrograde uptake of CTB conjugate (middle panel), and composite of the two (right panel). © Top: Schematics of experimental configuration for wide-field calcium imaging of visual cortex from the hemisphere in which ipsilateral SC conditionally expresses TeLC in NTSR1-GN209+ neurons (left) and from the control hemisphere (right). Bottom: Example ΔF/F (240 trial average) across visual areas in TeLC (left) and control (right) hemispheres. Cortical visual areas delineated as shown in Figure S1. Orientation: A: anterior; M: medial. (D). Summary data: Response magnitude for each area in TeLC hemisphere relative to the response in the contralateral hemisphere. Within each hemisphere, responses were normalized by the response in their respective V1 (see methods). Gray bar shows 95% confidence interval above and below mean (red line). Gray lines between dots link areas from the same animal. P values are calculated with a t-test. Green dots represent data from the example mouse shown in A and B. Within each mouse, data on the impact of TeLC on a given area was only included if that area responded to the visual stimulus in the control hemisphere (see methods).

Figure 4.. A Lateromedial Gradient of Tecto-Pulvinar versus V1 Afferent Input

(A) Schematic of experimental configuration. V1 projections to HVAs are labeled with tdTomato, while SC projections to HVAs via pulvinar are labeled with GFP using transsynaptic anterograde tracing. (B) Top: Confocal fluorescent micrograph of coronal sections of three example HVAs (POR, LI and LM) illustrating tecto-recipient pulvinar afferents (GFP; top row), V1 afferents (TdTomato; middle row), and the two merged in a composite image (bottom row). Pial surface (Pia) and white matter (WM) are to the left and right of the micrographs, respectively. Bottom: Summary data. Average normalized depth profiles of fluorescent density across all mice (N = 4 mice). Each fluorophore is normalized to the maximum density across all areas of a given mouse. (C) Fraction of total GFP (left) and tdTomato (right) fluorescence measured with each area. Bars show 95% confidence intervals above and below mean (black line). Gray lines between dots link areas from the same animal. Insets present anatomical “heatmaps” with average fraction of total fluorescence in each area.

Figure 5.. SC Dependent HVAs Distinguish Self from Externally Generated Motion of Visual Stimuli

(A) Schematic of experimental configuration for widefield calcium imaging of visual cortex in a mouse running on a treadmill. The animal is presented with a dot which progresses along a naso-temporal trajectory either at a speed commensurate to the running speed of the mouse (coupled; i.e. self-generated), or, on alternate trials, with a dot that replays the previous coupled stimulus irrespective of the running speed of the mouse (uncoupled; i.e. externally-generated). (B) Example ΔF/F (averaged over 37 trials) across visual areas in alternated coupled (left) and uncoupled (right) presentations of visual stimulus presentations with positive visuomotor divergence (VMD). The visual stimulus moved at the same speed in both conditions. Note the increased response of lateral visual areas in the uncoupled trials with positive visuo-motor divergence. Cortical visual areas delineated as shown in Figure S1. Orientation: A: anterior; M: medial. (C) Example trial illustrating the calcium response (ΔF/F), running speed and the moving dot speed for a pair of coupled and uncoupled stimulus presentations from a trial with positive visuo-motor divergence. The horizontal dotted lines indicate baseline for ΔF/F and 0 for running and stimulus speed. The positive visuo-motor divergence elicits a larger calcium response compared to an identical yet self-generated visual stimulus in SC-dependent visual areas but not in V1. (D) Time courses of ΔF/F for three example areas (POR, LI and V1) averaged across all mice (N = 4 mice) for coupled (black) and uncoupled (aqua) trials with positive (top row) or negative (bottom row) visuo-motor divergence. Note the increased response of POR and LI in the uncoupled trials with positive visuo-motor divergence, but not in uncoupled trials with negative visuo-motor divergence. Also, note similar responses in V1 for coupled and uncoupled trials across both conditions. Responses are normalized to the average response in V1 across all coupled trials. P-values are calculated with Benjamini-Hochberg method to control false-discovery rate from multiple comparisons. (E) Average difference in ΔF/F between uncoupled and coupled. Zero on the x-axis indicates that the dot moved at the same speed during the coupled and uncoupled stimulus presentation, whereas positive and negative values represent positive visuo-motor divergence (aqua) and negative (orange) visuo-motor divergence trials, respectively. (F) Summary data: ΔF/F in each area across all animals for coupled (black) and uncoupled trials with positive visuo-motor divergence (aqua). Responses are normalized by maximal response across all areas in a given mouse. Bars show a 95% confidence interval above and below the mean (black line). Gray lines between dots link responses from the same animal between conditions. P-values are calculated from a paired t-test.

Figure 6.. NTSR1-GN209+ Cells in SC Distinguish between Self and Externally Generated Motion of Visual Stimuli.

(A) Schematic of experimental configuration for fiber photometry in SC in a mouse running on a treadmill. Two excitation wavelengths at 470 nm (blue; excitation wavelength) and 405 nm (violet; isosbestic hemodynamic control channel) are alternated while imaging visual responses of GCamp7f-expressing NTSR1-GN209+ cells. The animal is presented with a dot which progresses along a naso-temporal trajectory either at a speed commensurate to the running speed of the mouse (coupled; i.e self-generated), or, on alternate trials, with a dot that replays the previous coupled stimulus irrespective of the running speed of the mouse (uncoupled; i.e externally-generated). (B) Example trial illustrating the calcium response (ΔF/F), running speed and the moving dot speed for a pair of coupled and uncoupled stimulus presentations. Negative visuo-motor divergence (VMD;aqua) and positive visuo-motor divergence (orange) events are highlighted during the uncoupled stimulus presentation of the visual stimulus. The horizontal dotted lines indicate baseline for ΔF/F and 0 for running and stimulus speed. A positive visuo-motor divergence event elicits a larger calcium response compared to an identical yet self-generated visual stimulus event. A negative visuo-motor divergence event elicits a response even without movement of the visual stimulus. (C) Average calcium response in NTSR1-GN209+ cells in SC for coupled (black solid line) versus uncoupled trials (dashed line). Data plotted as ΔF/F for negative visuo-motor divergence (aqua), positive visuo-motor divergence (orange) and trials with little visuo-motor divergence (light gray). Time zero indicates the onset of the visual stimulus on the monitor. A gray heatmap of p-values is inset for the comparison of the average response on coupled and uncoupled trials, with all p-values <0.05 indicated. N = 3 mice. For positive, negative, and trials with little visuomotor divergence, n = 194, 114, and 387 trials respectively. P-values are calculated with Benjamini-Hochberg method to control false-discovery rate from multiple comparisons. (D) Difference in calcium response in NTSR1-GN209+ cells between pairs of coupled and uncoupled stimulus presentations. Zero on the x-axis indicates that the dot moved at the same speed during the coupled and uncoupled stimulus presentation, whereas positive and negative values represent positive visuo-motor divergence (aqua) and negative (orange) visuo-motor divergence trials, respectively. The calcium response in SC is stronger for positive visuo-motor divergence and negative visuo-motor divergence during uncoupled replays compared to stimulus presentations coupled to the running speed of the mouse.