Fig. 3: V1 DS cells projecting to area RL or PM respond differently to disruption of retinal direction selectivity.

a Target-unspecific V1 neurons were labeled by injecting AAV2/1-GCaMP6f into the V1. PM- (PM-p) and RL-projecting (RL-p) V1 neurons were labeled by injecting rAAV2-retro-GCaMP6m into PM and RL, respectively. b Two-photon calcium imaging from L2/3 in V1 of control (1087 target-unspecific, 109 PM-p, and 513 RL-p DS cells; 5 mice per group) and Frmd7^tm mice (954 target-unspecific, 93 PM-p, and 235 RL-p DS cells; five mice per group). Left: example PM- and RL-projecting neurons expressing GCaMP6m (scale bar, 10 µm). Right: trial-averaged fluorescence (ΔF/F₀) time courses for the same neurons. Shading: SEM. c Tuning curves for neurons shown in b. Error bars: SEM. Solid line: Gaussian fit. d Fraction of DS cells in target-unspecific, PM-p, and RL-p V1 neuronal populations (two-sided χ² test with Yates correction). e Preferred TF for target-unspecific (two-sided Mann–Whitney U-test), PM-p (two-sided Mann–Whitney U-test), and RL-p (two-sided Mann–Whitney U-test) V1 DS cells. Triangles: Medians. f Response amplitude as a function of motion direction and TF for target-unspecific, PM-p, and RL-p V1 DS cells. White asterisks: significantly decreased in Frmd7^tm mice, two-sided Mann–Whitney U-test. g Fractional distributions of preferred motion directions for target-unspecific, PM-p, and RL-p V1 DS cells at 0.3 and 1.2 Hz. The fractions are normalized to the largest fraction across genetic groups. h Distributions of preferred motion directions at 0.3 and 1.2 Hz for target-unspecific, PM-p, and RL-p V1 DS cells (two-sided Kolmogorov–Smirnov test). **P < 0.01, ***P < 0.001, n.s., not significant, in d, e, and h. Source data are provided as a Source Data file.