Fig. 7. Disulfiram treatment improves coding of naturalistic scenes in V1.
(A) Top: Example frames from the movie clip presented to rd10 mice. Bottom: Example responses from three different neurons to repeated presentations of the movie. Each row is the response to a single presentation. Different neurons show varying amounts of response reliability to repeated presentations. (B) Comparison of the response reliability of neurons from disulfiram-treated versus control-treated mice (experimental design similar to Fig. 6A). Neurons in disulfiram mice show significantly greater reliability compared to neurons from control mice. (C) Comparison of reliability segmented by deciles. Reliability is similar in lower deciles but becomes significantly greater for disulfiram-treated mice in higher deciles. (D) Performance of a population-level decoder across control (left) versus disulfiram-treated (right) mice. (E) Across all population sizes used for decoding, disulfiram treatment shows significantly improved performance (predicted frame within 10 frames of actual frame) compared to control. (B, C, and E) Values shown are means ± SEM. Control: n = 1006 neurons across seven mice; disulfiram: n = 2075 neurons across nine mice. Two-tailed F test in (B), F(1,2066) = 5.27, *P = 0.022; in (C), F(1,2064) = 58.08, ***P = 3.82 × 10−14; in (E), F(1,79) = 16.20, ***P = 1.30 × 10−4.