Low-frequency cortical oscillations are modulated by temporal prediction and temporal error coding

Abstract

Monitoring and updating temporal predictions are critical abilities for adaptive behavior. Here, we investigated whether neural oscillations are related to violation and updating of temporal predictions. Human participants performed an experiment in which they had to generate a target at an expected time point, by pressing a button while taking into account a variable delay between the act and the stimulus occurrence. Our behavioral results showed that participants quickly adapted their temporal predictions in face of an error. Concurrent electrophysiological (EEG) data showed that temporal errors elicited markers that are classically related to error coding. Furthermore, intertrial phase coherence of frontal theta oscillations was modulated by error magnitude, possibly indexing the degree of surprise. Finally, we found that delta phase at stimulus onset was correlated with future behavioral adjustments. Together, our findings suggest that low frequency oscillations play a key role in monitoring and in updating temporal predictions.

Separable EEG Features Induced by Timing Prediction for Active Brain-Computer Interfaces

“…It is reasonable for timing prediction to induce positive delta waveform and beta-gamma ERSP suppression, especially in the central brain area. To be specific, delta oscillation in the central area is believed to be instrumental for reflecting many voluntary mental activities, including the perceptual judgement [ 20 , 30 , 31 ], difficulty assessment [ 30 ], target detection [ 31 ], perceptual learning [ 28 , 32 ], etc. In the current study, the ERP profiles started to go upwards shortly after the predicted moment, which worked as a reflection of perceptual judgement, indicating subjects have realized that time was up.…”

“…Specifically, the ‘what’ template was observed with the assistance of magnetoencephalogram (MEG) [ 21 ] or high-resolution functional magnetic resonance image (fMRI) [ 19 , 22 ], but no EEG evidence was found. While the ‘when’ template is usually embodied as the estimation of specific time interval and could trigger specific patterns of neural oscillations in both the MEG [ 23 , 24 , 25 , 26 ] and EEG [ 27 , 28 ] studies. On that basis, this study investigated the endogenous signatures of timing prediction and tested whether distinct temporal templates can induce separable EEG features for active brain–computer interfaces.…”

Implicit temporal predictability biases slow oscillatory phase in auditory cortex and enhances pitch discrimination sensitivity

“…36 It is currently an open question to what extend entrained neural delta oscillations 37 are a signature of processing rhythmic input, or whether they pose a more parsimo- 38 nious mechanism of temporal prediction. Important evidence for an endogenous role 39 of delta oscillations in auditory temporal attention comes from two studies showing 40 that auditory processing fluctuates with the phase of spontaneously present delta 41 activity in auditory cortex, in absence of rhythmic stimulation [38,39]. 42 Furthermore, previous studies have shown that entrainment is subject to top- 43 down modulation, shown by enhanced phase coherence of slow oscillations in antici- 44 pation of temporally predictive input [11,16,35].…”

“…To our knowledge, only one study in the visual domain reported a role 49 of slow oscillations in single trial temporal predictions [5, theta band]. Furthermore, 50 a recent study [41] showed that delta phase in the target-onset time window reflects 51 adjustments to previously encountered violations of temporal predictions in an explicit 52 timing task. To date, to the best of our knowledge, no study has assessed whether 53 oscillations implement implicit temporal predictions for audition.…”

“…Due to computational constraints, we only specified a random 327 intercept, but no random slopes for the predictors. 328 To account for the circularity of the phase angles, we followed an approach 329 previously described by Wyart et al [55] (see also [16,41]) of using the sine and cosine 330 of the phase angles jointly as linear predictors in a regression. For both, the sin(ϕ) 331 and cos(ϕ), we specified an interaction with ∆pitch:…”