Fig. 6: Normalized satisfaction and certainty ratings.
In each figure panel, the 25–75% quartiles of the data are displayed as a thick line around the median. Source data are provided as a Source Data file. a Distribution of normalized satisfaction ratings for each participant (P2: 152 trials; P3: 95 trials; C1: 104 trials). The median satisfaction score for each participant is displayed as an open circle. b Normalized satisfaction ratings per object across all participants (71 cat, 71 apple, 70 key, 67 towel and 72 toast sensations). On average, participants were more satisfied with the sensations for a cat than for a key (p = 0.005−1, two-sided Kolmogorov–Smirnov test Bonferroni corrected at α = 0.005, D = 0.33). No other differences were significant. c Self-rated certainty with which a participant could (correctly or incorrectly) identify the object corresponding to a set of stimulus parameters in the replay task. Participant P2 was significantly more certain of his answers when they were in fact correct, compared to when they were not (p = 0.003, Wilcoxon rank sum, α = 0.025, z = 2.73, 64 correct and 116 incorrect trials). No significant differences were found for P3 (25 correct and 81 incorrect trials) or C1 (71 correct and 10 incorrect trials).
