Fig. 5: Uncertainty increases aperiodic dynamics during sensation.

Aperiodic dynamics are estimated via neural sample entropy (a–c, SampEn) and 1/f power spectral density (PSD) slopes (d–f). a Temporal traces of sample entropy (mean ± within-subject SEM; n = 47 participants). The yellow background indicates the period of stimulus presentation. The black line indicates time points at which permutation tests indicated linear load effects. b Topography of mean linear load effect estimates (n = 47), with black dots representing the significant cluster. c Post hoc analysis of entropy estimates within significant cluster. Gray dots indicate individual outliers (defined as Cook’s distance > 2.5 × mean (Cook’s distance); n = 2 participants) and have been removed from the statistical post hoc assessment (n = 45 participants). Data are within-subject centered for visualization (see “Methods”; linear: b = 0.002, 95% CI = [0.001, 0.003], t(44) = 4.22). d Aperiodic slopes shallow with increasing target load (i.e., spectral rotation across low- and high-frequencies; mean ± within-subject SEM; n = 47 participants). Lower and upper insets highlight slope differences at low and high frequencies, respectively. e Topography of mean linear load effects on 1/f slopes (n = 47). Black dots indicate the significant occipital cluster used for post hoc assessments. f Same as c, but for occipital aperiodic slopes. Data are within-subject centered for visualization (see “Methods”; n = 3 outliers in gray; effective n = 44 participants in red; linear: b = 0.008, 95% CI = [0.003, 0.012], t(43) = 3.06). Panels b and e indicate p values from dependent samples regressions on target load, corrected for multiple comparisons via cluster-based permutations. Panels c and f indicate p values from two-tailed paired t-tests (linear effects vs. zero). Source data are provided as a Source Data file.