Visual temporal binding is dynamically shaped by sensory input, perceptual history and confidence.

(A) Schematic depiction of the sustained-stream temporal integration task. On each trial, two low–spatial-frequency Gabor patches of opposite orientation (±45°) alternated in counterphase for ~1 s with variable inter-stimulus intervals (ISIs, 0–146 ms, steps of 21 ms). Participants reported whether they perceived a fused plaid (integration) or separate gratings (segregation), followed by a confidence rating (1–4). (B) Psychometric functions showing the proportion of segregation responses as a function of ISI for all trials (black), trials following t–1 integration responses (red), and t–1 segregation responses (blue). Shaded areas denote SEM. The Point of Subjective Equality (PSE) was higher following t–1 integration trials, indicating a broader temporal binding window. (C) Individual PSE values as a function of previous perceptual outcome. Integration on the preceding trial biased perception toward integration on the current trial, indicating a stronger influence of recent perceptual choice-history on temporal binding processes (serial dependence effect). Black boxplots indicate the median and interquartile range (IQR), and individual dots represent participant-level data. (D) Mean z-scored confidence ratings as a function of ISI and t–1 judgment. Confidence decreased near the perceptual threshold (intermediate ISIs) and increased when temporal evidence clearly supported either integration or segregation. Confidence was higher after t–1 integration responses at short ISIs (21–63 ms) and higher after t–1 segregation trials at long ISIs (104–125 ms; q < 0.05, FDR-corrected), showing an interaction between prior experience and temporal evidence. (E) Across participants, mean confidence negatively correlated with temporal integration threshold (PSE; r = –0.28, p = 0.009; Parson Correlation Coefficient), indicating that individuals with narrower integration windows (higher temporal acuity) reported greater confidence. Together, these results show that visual temporal binding and metacognition are dynamically shaped by recent perceptual experience, reflecting an adaptive interplay between sensory evidence and perceptual history.

Oscillatory and aperiodic neural dynamics predict visual temporal integration performance.

(A) Topographical maps show the scalp distribution of correlations between individual alpha frequency (IAF) and temporal threshold of the psychometric function (PSE) during eyes-open (EO) and eyes-closed (EC) resting-state conditions. Faster IAFs over posterior–central sites were significantly associated with lower temporal thresholds (p < 0.001, cluster-corrected), indicating finer temporal resolution and narrower integration windows. Scatterplot shows the pooled correlation across participants over posterior-central electrodes. (B) Correlations between IAF and psychometric slope revealed no significant effects in either resting-state condition (p > 0.05), suggesting that alpha frequency primarily modulates temporal binding boundaries rather than perceptual precision. (C) Topographies and scatterplot depict positive correlations between aperiodic exponent and temporal threshold in the EO condition (p < 0.001, cluster-corrected), indicating that steeper aperiodic spectra (i.e., lower neural noise) were linked to finer temporal resolution. No reliable associations were found in the EC condition. (D) Higher aperiodic exponents were robustly correlated with steeper psychometric slopes in both EO and EC conditions (p < 0.001, cluster-corrected), showing that reduced neural excitation supports greater perceptual sensitivity and more stable temporal integration. Together, these results reveal that visual temporal acuity depends jointly on alpha oscillations and aperiodic activity, with faster alpha rhythms and steeper aperiodic slopes facilitating more precise temporal sampling of sensory input.

Resting-state neural dynamics predict individual differences in serial dependence and confidence.

(A) Topographies show correlations between individual alpha frequency (IAF) and serial dependence strength (ΔPSE = PSE t–1 integration – PSE t–1 segregation) for eyes-open (EO) and eyes-closed (EC) conditions. Faster IAFs over posterior–central electrodes were associated with lower ΔPSE in the EO condition (p < 0.001, cluster-corrected), indicating that individuals with faster alpha rhythms (i.e., higher temporal acuity) relied less on prior perceptual outcomes. No significant effects were observed in the EC condition (p > 0.05). (B) Correlations between the aperiodic exponent and ΔPSE revealed that flatter spectra (lower exponents, greater neural noise) were linked to higher ΔPSE values in the EO condition (p < 0.001, cluster-corrected), suggesting greater dependence on perceptual history to stabilize noisy sensory representations. This relationship was absent in the EC condition (p > 0.05). (C) Mean subjective confidence (z-scored) correlated positively with IAF in the EO condition (p < 0.001, cluster-corrected), showing that individuals with faster alpha cycles exhibited not only higher temporal acuity but also greater confidence in their perceptual judgments. No significant correlations were found for EC or for aperiodic activity (p > 0.05). Together, these findings demonstrate that oscillatory and aperiodic neural mechanisms jointly shape perceptual and metacognitive aspects of temporal integration: slower alpha rhythms and flatter spectra bias perception toward reliance on prior experience, whereas faster alpha oscillations enhance temporal precision and subjective certainty.

Resting-state EEG decomposition and comparison of oscillatory and aperiodic neural activity across eyes-closed and eyes-open conditions.

(A) Representative FOOOF spectral parameterization of resting-state EEG averaged over posterior–occipital electrodes for the eyes-closed (EC, left) and eyes-open (EO, right) conditions. The algorithm decomposes the original power spectral density (PSD; black line) into its aperiodic 1/f component (blue dashed line) and periodic oscillatory peaks (red line). Individual alpha frequency (IAF; vertical dashed line) and the aperiodic fit. (B) Topographical distributions of IAF during EC (left) and EO (middle) resting-state conditions, and their scalp-level difference map (EC–EO, right). No significant differences in IAF were observed between conditions (p > 0.05, cluster-corrected). (C) Topographical distributions of the aperiodic exponent during EC (left) and EO (middle), and the corresponding difference map (EC–EO, right). Aperiodic exponents were significantly higher during EC relative to EO (p < 0.001, two-tailed, cluster-based correction; black asterisks indicate significant electrodes), consistent with stronger aperiodic activity under eyes-closed rest.