Spontaneous Brain Dynamics Track the Episodic "When"

Across species, neurons track time over the course of seconds to minutes, which may feed the sense of time passing. Here, we asked whether neural signatures of time-tracking could be found in humans. Participants stayed quietly awake for a few minutes while being recorded with magnetoencephalography (MEG). They were unaware they would be asked how long the recording lasted (retrospective time) or instructed beforehand to estimate how long it will last (prospective timing). At rest, rhythmic brain activity is nonstationary and displays bursts of activity in the alpha range (alpha: 7-14 Hz). When participants were not instructed to attend to time, the relative duration of alpha bursts linearly predicted individuals' retrospective estimates of how long their quiet wakefulness lasted. The relative duration of alpha bursts was a better predictor than alpha power or burst amplitude. No other rhythmic or arrhythmic activity predicted retrospective duration. However, when participants timed prospectively, the relative duration of alpha bursts failed to predict their duration estimates. Consistent with this, the amount of alpha bursts was discriminant between prospective and retrospective timing. Last, with a control experiment, we demonstrate that the relation between alpha bursts and retrospective time is preserved even when participants are engaged in a visual counting task. Thus, at the time scale of minutes, we report that the relative time of spontaneous alpha burstiness predicts conscious retrospective time. We conclude that in the absence of overt attention to time, alpha bursts embody discrete states of awareness constitutive of episodic timing.