Fig. 2: Large-scale retinal receptive field mapping.

a, Normalized absorption spectra of mouse photoreceptors (purple, S-opsin; green, M-opsin). Normalized emission spectra of the UV and green light emitted by the DLP projector (filled purple, UV; filled green, green stimulus light), epifluorescence (orange) and two-photon (red) excitation are overlaid. b, Schematic of the epifluorescence imaging setup. c, Montage of five consecutively recorded fields (orange dashed box denotes one field) of a whole-mounted mouse retina from a Vglut2-ires-cre; TITL-R-CaMP1.07-D; ROSA26-ZtTA triple-transgenic mouse. Inset: black, imaged montage; red, retinal outline. d, Double-labeled immunostaining of RCamp1.07-expressing RGCs (red) and RBPMS (cyan; n = 3). e, As in d but labeling with SMI32 (cyan). Arrowheads depict double-labeled cells. f, Example Ca²⁺ signals (gray, five repetitions; black, mean) from DS (top) and non-DS (bottom) RGCs. g, Example distribution of preferred directions in one FOV. Inset shows a polar plot of DS preference. h, Example Ca²⁺ signals to chirp stimulus from three different RGCs (gray, five repetitions; black, mean). i, Recording Ca²⁺ signal stability across sequentially imaged FOVs for nine retinas (each session lasted ~25 min, 3–7 sessions per retina). White lines denote medians, and minima and maxima of the gray bars indicate the 25th and 75th percentile range of the dF/F distribution, respectively. j, Example RFs recorded using ‘shifting’ white noise (top) and their respective parametrizations (bottom). Blue and red ellipses correspond to 2 s.d. contours of the ON and OFF Gaussians, respectively. k, Histogram of goodness of fit for all recorded RFs.