Extended Data Fig. 3: Changes in preference over time in the Multiple Models experiment.

Points show mean and vertical bars show 95% confidence intervals based on the t-distribution. a Training phase. Lines show an asymptotic curve fitted to the data from session 16 onwards; the curve for A100 did not converge but is shown for illustration. Session 16 was the first session after a period of one week when no birds opened dishes. Curves fitted to the full time period all failed to converge. A100 session 9 had a small sample size (3) and as a result has very wide confidence intervals (−5.3, 8.3) that are not shown in full for clarity of the rest of the plot. N = 3470 dishes (M100), 769 (A100) and 2227 (V100). b Test phase, trends with time. Asymptotic curves fitted to these data, using the same method as the training phase, failed to converge. Instead, trend lines are a moving average across five sessions (centred on the third session). Sample sizes shown above each plot. c Test phase, comparing initial (yellow, session 1-3) and asymptotic (black, session 10 onwards, chosen to match Discrimination Experiment) preferences. N = 437 dishes (initial, 1 M), 384 (initial, 2 M), 2902 (asymptotic, 1 M), 4334 (asymptotic, 2 M).