Supplementary Figure 4: Additional receptive field (RF) properties.

(a) Receptive field examples of dLGN boutons, V1 somata and LP boutons with a single subfield. Scale bars, 10 deg. (b) Distributions of minor and major axis lengths of subfields from dLGN boutons, V1 somata and LP boutons. (c) Distributions of the major axis orientations of receptive fields. The subfields of LP receptive fields were more likely to be horizontally oriented, while the subfields of dLGN boutons were biased for vertical orientations. Subfields of V1 neurons were less orientation-biased. dLGN: n = 2317, LP: n = 1825, V1: n = 356 RF subfields. *, P < 0.05; ***, P < 10⁻¹⁰, Χ²-test for uniformity. (d) Examples of dLGN, LP and V1 receptive fields with significant ON and OFF subfields. About one third of LP and dLGN boutons and V1 neurons exhibited subfields of both signs (LP: 471 of 1354; dLGN: 590 of 1727; V1: 89 of 267). Scale bar, 10 deg. (e) Distributions of distances between the centroids of ON and OFF subfields for boutons or neurons with a significant ON and OFF subfield. The subfields of dLGN boutons were often arranged in a center-surround configuration, such that the distances between ON and OFF subfield centroids were relatively small. LP boutons showed a larger subfield separation while values for V1 neurons lay in between LP and dLGN. dLGN: n = 590, LP: n = 471, V1: n = 89 receptive fields. **, P < 0.0001; ***, P < 10⁻¹⁰, Wilcoxon rank-sum test. These differences in distance cannot be fully accounted for by differences in RF size. Calculating the RF subfield distance normalized to the average axis length of the subfields resulted in median distances between ON and OFF subfield centroids of 18% (dLGN), 22% (V1) and 25% (LP) of the average axis length. dLGN: 7 mice, LP: 13 mice, V1: 4 mice.