Streamlined sensory motor communication through cortical reciprocal connectivity in a visually guided eye movement task

Abstract

Cortical computation is distributed across multiple areas of the cortex by networks of reciprocal connectivity. However, how such connectivity contributes to the communication between the connected areas is not clear. In this study, we examine the communication between sensory and motor cortices. We develop an eye movement task in mice and combine it with optogenetic suppression and two-photon calcium imaging techniques. We identify a small region in the secondary motor cortex (MO_s) that controls eye movements and reciprocally connects with a rostrolateral part of the higher visual areas (V_RL/A/AL). These two regions encode both motor signals and visual information; however, the information flow between the regions depends on the direction of the connectivity: motor information is conveyed preferentially from the MO_s to the V_RL/A/AL, and sensory information is transferred primarily in the opposite direction. We propose that reciprocal connectivity streamlines information flow, enhancing the computational capacity of a distributed network.

Fig. 1

A visually guided eye movement task in mice. a The experimental design. Three LEDs instructed when and where to move the left eye. b After the mice fixated the central LED, one of the two LED targets (nasal or temporal) was turned on, and mice were required to shift their left eye toward the target. c Representative eye traces recorded during one behavioral session. Traces are aligned to the target onset. Cyan traces, trials with nasal target; magenta traces, trials with temporal target. d The distributions of reaction times for one animal (cyan, nasal target condition; magenta, temporal target condition). Reaction time range <3 s (nasal 91.7%; temporal 81.3%) is shown because of the long tail of its distributions. e Average traces for eye movement for both eyes in one behavioral session (nasal, 32 trials, cyan; temporal, 39 trials, magenta). Traces for the left eye, solid lines; right eye, dotted lines. Thick and thin lines represent mean ± s.e.m. f The amplitude of the eye movements for the left and right eyes (cyan, nasal target; magenta, temporal). The same animal as in e. g The average eye movement amplitude is shown for the nasal (cyan) or temporal (magenta) target trials in 10 mice. Error bars indicate s.e.m

Fig. 2

Optogenetic suppression of the MO_s during the visually guided eye movement task. a The experimental design. b ChR2 was expressed virally in PV interneurons in Pvalb-IRES-Cre transgenic mice. AAV-DIO-ChR2-eYFP was injected into the MO_s (coronal section stained with Nissl), and ChR2-eYFP expression was confined to cells expressing PV (white arrows). c Firing was suppressed during blue light illumination. Left, two representative neurons are shown. Right, blue light illumination completely suppressed firing activity in all eight neurons (p < 0.01, Wilcoxon signed-rank test). The activity was monitored with blind cell-attached recording. d–g Optogenetic suppression of the MO_s during the task. The virus was injected into the MO_s either in the left (d, g) or right (e, f) hemisphere. Unilateral suppression of the MO_s severely impaired eye movements in the contraversive (d, f) but only mildly in the ipsiversive (e, g) direction. h, i Reaction times for the eye movements. PV activation delayed the onset of contraversive eye movements (h, PV activation, n = 96 trials; control, n = 88 trials; p < 10⁻¹¹), but showed only minor effects on ipsiversive movements (i, PV activation, n = 84 trial; control, n = 126 trials; p = 0.0501, Person’s chi-square test)

Fig. 3

Neural coding in the MO_s. a The experimental design. b Cranial window for two-photon imaging. Arrowheads indicate 700 µm rostral and 700 µm lateral from the bregma. Note a bridging vein leading to the superior sagittal sinus, which often runs around the coordinates. c Virally expressed GCaMP6f in the MO_s for two-photon calcium imaging during the visually guided eye movement task. d Averaged fluorescence change of two example cells (ROIs shown in c, aligned to the onset of the visual target or to the onset of the eye movement; cyan, the trial condition for ipsiversive eye movement; magenta, for contraversive). e, f Sorted normalized ∆F/F traces for 904 neurons showing significant response either to visual targets or to eye movements. Response was aligned to the onset of the visual target or the movement in the contraversive (e) or ipsiversive (f) trials. g–j Comparison between contraversive and ipsiversive movement trials for visual and motor activity (n = 519 or 471 neurons, showing significant visual or motor activity). i, j A histogram for difference between ipsiversive and contraversive conditions in visual or motor activity. Neurons with a larger response in the contraversive condition (p < 0.05, Mann–Whitney test) were labeled as magenta (n = 135 in visual activity, 122 in motor activity) and neurons with a larger response in the ipsiversive condition in cyan (n = 80, 45). The population preferred for the contraversive condition in visual activity (n = 519, p < 0.05, Wilcoxon signed-rank test) and in motor activity (n = 471, p < 10⁻¹²)

Fig. 4

Reciprocal connectivity between MO_s and V_RL/A/AL. a–d Connectivity from the MO_s to the V_RL/A/AL. a The experimental design for the dual-virus injections. AAV-tdTomato was injected into the MO_s in the right hemisphere and AAV-GFP was injected into the occipital regions in the left hemisphere. b A corresponding whole-brain image. c, d The tdTomato projection from MO_s overlapped with areas RL, A, and AL (V_RL/A/AL). GFP-labeled callosal projections delineated boundaries between the V1 and higher visual areas. The image c was taken with a fluorescent stereoscopic microscope from the dotted rectangle in b, and the image d was taken with a two-photon microscope from a region indicated in b. e–n The V_RL/A/AL projects back to the MO_s. e The experimental design for the sequential virus injections. First, AAV-tdTomato was injected into the MOs, and then a couple weeks later AAV-GFP was injected into the tdTomato-labeled axonal band (the V_RL/A/AL). f A corresponding whole-brain image. g At the injection site of AAV-GFP, the expression overlapped with the tdTomato-labeled axonal band (the location of V_RL/A/AL). h At the injection site of AAV-tdTomato (the MO_s), the expression overlapped with the GFP-labeled axonal band. The two black arrowheads indicate the midline, and the white dotted line is the expected areal border based on coordinates and Paxinos and Franklin’s atlas. i–k A coronal section that includes the MO_s. The injection site for AAV-tdTomato at MO_s (j, tdTomato) and the axon terminals from the V_RL/A/AL (k, GFP) overlapped well. l–n A coronal section that includes V_RL/A/AL. The axon terminals originating from the MO_s (m, tdTomato) and the inection site for AAV-GFP (n, GFP) overlapped well. The dashed lines in j, k, m, and n indicate the areal borders based on Nissl staining. o Projections from the MO_s and the V_RL/A/AL in thalamus. Note that projections from the V_RL/A/AL to the higher visual thalamus (LPLR) overlapped with those from the MO_s. The borders were drawn based on descriptions in refs. ^,

Fig. 5

Neural coding in the V_RL/A/AL and the V1. a–c Coding in the V_RL/A/AL. a GCaMP6f expression at layer 2/3 in the V_RL/A/AL. The position of the V_RL/A/AL was determined with fluorescent guidance of tdTomato-labeled projections from the MO_s. b Representative cells. Corresponding ROIs are shown in a. c Sorted normalized ∆F/F responses for neurons showing significant response (n = 913). Traces were aligned to either the onset of the visual target or the movement in the contraversive movement condition. d–f Coding in the V1. d GCaMP6f expression at layer 2/3 in the V1. e Representative cells. f Sorted normalized ∆F/F responses for neurons with a significant response (n = 306). g–l Comparison between visual and motor activity for neurons in the MO_s (g, h, n = 904), V_RL/A/AL (i, j n = 913), and V1 (k, l, n = 306). g, i, k Visual and motor activity was plotted for each neuron. h, j, l The angle in polar coordinates was computed for each neuron and plotted as a histogram. Degrees of 0 or 90 indicate pure visual or motor activity

Fig. 6

Biased neural coding in MO_s and V_RL/A/AL axons. a–c Coding in MO_s axon terminals projecting to the V_RL/A/AL. a GCaMP6f was expressed in MO_s neurons, and their axons were imaged in the V_RL/A/AL. b Calcium traces from two representative axons (mean ± s.e.m., cyan for the ipsiversive movement condition; magenta for the contraversive). The corresponding ROIs are shown in a. c–e Comparison between visual and motor activity for MO_s axons (c, n = 123), L5 somas (d, n = 395), and L2/3 somas (e, n = 904). f–h Histograms of the angle in polar coordinates for MO_s axons, L5 neurons, and L2/3 neurons. The distribution of the axons is more biased towards 90 degrees, indicating higher motor selectivity. Median, 25% quantile, and 75% quantile are represented as box plots, whisker length is 1.5 of the interquartile range. i–p Similar to a–h, but for V_RL/A/AL axon terminals projecting to the MO_s. V_RL/A/AL axons, n = 170; V_RL/A/AL L5 somas, n = 289; L2/3 somas, n = 913. Note skewed distribution of the axons towards 0 degrees, indicating higher visual selectivity. The histograms for L2/3 somas (h, p) are the same with those in Fig. 5h, j