Supplementary Figure 1: Odor identity classification in dpDp (related to Fig. 1).

a. Classification of odor identity by template matching of activity vectors in naïve (NAV) fish using Pearson correlation distance (1 – Pearson correlation; cf. Fig. 1g). Fills show percentage of correctly decoded odors using a fixed population size of 95 neurons. White dashed line indicates chance level (25 %; McNemar test, n = 156 trials, d.f. = 1, Χ2 = 75.01, P < 10–16); black dashed line shows mean classification success, averaged over all four odors (McNemar test for comparison against 100% correct, Χ2 = 38.03, P = 7 x 10–10). Comparison against classification success using cosine distance (Fig. 1g; McNemar test): Χ2 = 0, P = 1. b. Same odor classification procedure using Euclidean distance as distance metric. Comparison against chance level (25 %; McNemar test): Χ2 = 69.01, P = 1 x 10–16. Comparison against 100% correct (McNemar test): Χ2 = 44.02, P = 3 x 10–11. Comparison against classification success using cosine distance (Fig. 1g; McNemar test): Χ2 = 4.17, P = 0.04. n and d.f. as in (a). ***P < 0.001.