Supplementary Figure 2: Innate odor preference and associative olfactory conditioning (related to Fig. 2).
a. Behavioral setup to analyze innate behavioral responses to odors. Tank water (control) or odor solution (His, Ser, Ala or Trp) were delivered to one side of the tank using a gravity-fed system. b. Experimental paradigm. Following acclimatization, tank water was applied to test for non-specific responses, followed 10 min later by odor application at the same location. Swimming speed was quantified before applications and for 40 s after each application. c. Left: Behavioral discrimination score (CS+ preference score, calculated as ζCS+ – ζCS– over the last nine trials) did not differ significantly between ALA, TRP, and HIS training groups (one-way ANOVA, d.f. = 42, F = 0.72, P = 0.49, N (ALA) = 12 animals, N (TRP) = 16, N (HIS) = 15). Open circles represent individual fish. Multiple comparisons between all groups (Tukey test, two-sided): ALA vs TRP, q = 0.21, P = 0.98; ALA vs HIS, q = 1.27, P = 0.65; TRP vs HIS, q = 1.59, P = 0.50. Right: behavioral preference for Ala vs Trp or His (Trp for ALA, TRP, UNC; His for HIS; ζAla – ζTrp or His; 0: no preference; > 0: preference for Ala; < 0: preference for Trp or His). Ala preference score (calculated over last nine trials) differed between training groups (one-way ANOVA, d.f. = 54, F = 31.83, P = 9 x 10–12, N as before and N (UNC) = 12). Multiple comparisons between all groups (Tukey test, two-sided): ALA vs TRP, q = 12.66, P = 5 x 10–11; ALA vs HIS, q = 10.97, P = 2 x 10–9; ALA vs UNC, q =5.64, P = 0.001; TRP vs HIS, q = 1.63, P = 0.66; TRP vs UNC, q = 6.64, P = 0.0001; HIS vs UNC, q = 5.04, P = 0.004. Box plot: center line, median; box limits, interquartile range; and whiskers, s.d. d. Examples of behavioral responses during appetitive conditioning. Single trial examples of swimming trajectories during the 30 s after odor onset, but prior to food delivery (ALA fish). Trials of the same fish were chosen from the first (left panel) and last (right panel) training day. Top: trajectory plots of one CS+ trial (Ala, red) and one CS− trial (Trp, blue). Brightness encodes z-level in the water column. Center and bottom: histograms of fish position in each video frame extracted from the trajectories above. e. Same plots as in (d) for a TRP fish. f. Mean learning curves for individual components of appetitive behavior (cf. Fig. 2d and ref.33). Lines and shading show the mean (± s.e.m.) of ζ for the first three days of training (nine trials per day). Comparisons between CS+ and CS– (Wilcoxon signed rank test, two-sided, N = 43 animals, ALA, TRP, and HIS). z-level: day 1, T = 293, P = 0.03; day 2, T = 169, P = 0.0001; day 3, T = 111, P = 3 x 10–6. Speed: day 1, T = 291, P = 0.03; day 2, T = 85, P = 3 x 10–7; day 3, T = 121, P =6 x 10–6. Distance: day 1, T = 268, P = 0.01; day 2, T = 208, P = 0.001; day 3, T = 232, P = 0.003. Surface (peaks): day 1, T = 454, P = 0.82; day 2, T = 243, P = 0.005; day 3, T = 56, P = 1 x 10–8. Area: day 1, T = 271, P = 0.01; day 2, T = 297, P = 0.03; day 3, T = 199, P = 0.0007. Circling: day 1, T = 423, P = 0.55; day 2, T = 464, P = 0.92; day 3, T = 404, P = 0.41. ns: P ≥ 0.05; *P < 0.05; **P < 0.01; ***P < 0.001.
