Temporal signatures of thought—neurodynamics distinguish on- and off-task thoughts

Our thoughts fluctuate dynamically, driven either by external stimuli and tasks (on-task thoughts) or drifting to task-unrelated contents (off-task thoughts or mind wandering). Although research has identified neural markers distinguishing different thought types, the temporal signature (dynamics) of on- and off-task thoughts remains poorly understood. This EEG study investigated different neurodynamical features-autocorrelation window (ACW), Lempel-Ziv complexity (LZC), power-law exponent (PLE), and median frequency (MF)-to differentiate these thoughts in their underlying dynamics during a signal-response task. Off-task thoughts exhibited prolonged ACW, reduced LZC, increased PLE, and smaller MF compared to on-task thoughts, establishing a distinct neurodynamic signature. Through statistical modeling, we identified a hierarchical background-foreground structure among these measures that unfolds along a temporal continuum, transitioning from longer block-level (17-second) to shorter trial-level (3-second) timescale. Notably, the longer background (block-level ACW) and shorter foreground (trial-level ACW and LZC) layers are tightly coupled during the "faster and shorter" on-task thoughts whereas they are more loosely related during "slower and longer" off-task thoughts. These findings, replicated in an independent dataset, demonstrate how the organization of our brain's dynamics, along a temporal continuum of longer background durations to shorter foreground durations, shapes on-task and off-task thoughts thereby yielding their distinct signatures.