ReMind: A Retrospective Self-Report Paradigm for Studying Mind-Wandering Onset During Reading

This is the dataset for ReMind paper. Paper abstract: Mind-wandering (MW) is a frequent and pervasive phenomenon, yet it is commonly assessed using self-reports or probe-based methods that offer limited temporal precision regarding its onset. In this study, we introduce a novel paradigm, ReMind, that estimates the onset and duration of MW episodes during natural reading by combining retrospective self-reports with eye-tracking. Specifically, participants were asked to indicate the words where they believed their mind started and stopped wandering during each MW report. These selections were aligned with gaze timestamps to estimate MW onset. To test whether knowing MW onset benefits both self-report and probe-based paradigms, we simulated probes by randomly sampling time points during reading and trained logistic regression classifiers to distinguish MW episode from on-task reading using eye-tracking features identified in past studies. Classifiers trained on time windows anchored to MW onset outperformed those using arbitrary windows in both the self-report and simulated thought-probe paradigm (AUROC = 0.655 and 0.626, respectively, using leave-one-subject-out cross-validation). We also trained classifiers using eye-tracking features extracted from sliding windows around MW onset, where classification performance peaked near AUROC = 0.65 about 3 seconds after MW onset. Feature-level analyses revealed systematic changes in gaze behavior, including reduced fixation rate and increased pupil size following MW onset. These findings demonstrate the temporal progression of gaze behaviors from on-task reading to MW and suggest that our paradigm effectively captures the covert cognitive state of MW onset. Overall, this approach offers a valuable tool for studying MW onset and monitoring attention in naturalistic settings.