Fig. 2
From: The interplay of uncertainty, relevance and learning influences auditory categorization
On average, participants weighted tones according to their relevance when making category decisions. (A) For both signal (grey) and distractor (purple) tones, their influence on category choice is computed using correlation of the respective tone frequencies with their associated average category choice probability (Methods). Error bars: standard error of the mean. Statistics: two-sided paired t-test between each subjects’ normalized correlation coefficient of signal tone frequencies with choice probability and correlation coefficient of distractor tone frequencies with choice probability. (B) Average accuracy in unbiased trials decreased with the number of distractor tones. Error bars: standard error of the mean. Statistics: one-way repeated measures ANOVA; ptone_position = 2.82e-58, Ftone_position(3, 165) = 226.61. (C) For trials with exactly one distractor tone, accuracy was higher when the distractor tone frequency was similar to the trial category (Blue: low-category trials, red: high-category trials). Binned frequencies of the distractor are shown on the x-axis. Error bars: standard error of the mean. Statistics: two-tailed Wilcoxon signed-rank test between accuracy in high- and low- category trials computed at each distractor frequency. N = 56 participants. †p < 0.05, *p < 0.01, **p < 0.001, ***p < 0.0001.
