Preprocessed data for: Clock-free optimal stopping in decision-making

Preprocessed data accompanying: Clock-free optimal stopping in decision-making: Stochastic sensitivity dynamics predict behavioral, neural, and perceptual transitions — Tom Froese (2026) This archive contains the preprocessed data needed to reproduce the figures reported in the paper, without downloading the full raw dataset from OSF. preprocessed.zip — unzip into the data/ directory of the code repository to produce data/preprocessed/ with five subdirectories: ClickTimes/ (click response times), Haptics/ (haptic feedback, 100 Hz), EDA/ (electrodermal activity, 25 Hz), EEG/ (global scalp potential and sensitivity-model fits, MAT files), and PAS/ (perceptual-awareness-scale analyses, MAT files). eeg_task_raw_fif.zip — the 62 cleaned continuous EEG recordings (MNE-Python *_task-raw.fif, one per participant) used by the aperiodic-exponent (1/f) analysis. Unzip into data/preprocessed/EEG/ to add the per-dyad pceXX/ folders; this lets you reproduce the neural exponent figure from preprocessed data alone. All files derive from the raw Perceptual Crossing Experiment dataset archived on OSF (https://osf.io/47n3p). Code repository: https://github.com/tom-froese/pce-sensitivity-dynamics. Changes in v3 (2026-07-13): EEG channel-order / montage correction. The per-trial EEG had been analysed under a manufacturer-default montage (with PO9/PO10) that did not match the shipped anatomical channel order (with AFz/Iz). Channel labels have been corrected to the true anatomical order throughout the pipeline, and eeg_task_raw_fif.zip has been rebuilt from raw on the corrected montage (all 31 dyads / 62 recordings). Channel-mean results—the global scalp potential and its sensitivity-model fit (grand R² ≈ 0.886) and the 1/f aperiodic exponent—are permutation-invariant and are unchanged; only the per-channel / topographic outputs (e.g. the R² scalp map) were affected, and these have been corrected. Removed the retired 7-ROI “by brain region” breakdown (regions of interest), which was not reported in the manuscript: the corresponding roi* fields have been dropped from globalScalpPotential_data.mat and globalScalpPotential_stats.mat in EEG/. The global-scalp-potential sensitivity-model fits themselves are retained (and corrected). Full provenance of the montage correction is documented in the project’s cap/montage provenance record (2026-07-13); the corrected analysis code is in the code repository above.