Disentangling sensory- and motor-related EEG activities elicited in the context of rhythmic sensorimotor interactions.

Chemin, Baptiste [UCL] Huang, Gan [UCL] Mouraux, André [UCL]

Disentangle rhythmic sensory- and motor-related brain activity in electroencephalographic signal is rendered difficult by the fact that both processes can be expected to elicit Steady-State Evoked- Potentials (SS-EPs) at the same frequencies. Previous behavioural studies have shown that when participants are asked to tap along a musical beat, their taps precede the actual occurrence of the beat by 30-50 ms. This phenomenon, referred to as mean negative asynchrony is smaller in musicians than non-musicians. Furthermore, in a pilot experiment (6 participants) we observed significant fluctuations in the tapping period, resulting in some amount of desynchronization between auditory input and motor output. Therefore, we investigated whether the fluctuations in motor output can be used as a “temporal signature” to disentangle sensory- and motor-related periodic brain activity. The EEG was recorded in 16 healthy participants while they tapped in synchrony with a periodic auditory beat. The onset of the taps was measured using a touch pad. In the original EEG signals, activity related to sensory processing can be expected to follow the strict periodicity of the auditory beat and, hence, to elicit an SS-EP concentrated at the beat frequency. In contrast, movement-related activity can be expected to exhibit temporal fluctuations following those of the actual tapping movement and, hence, to elicit a relatively attenuated SS-EP spread around the mean tapping frequency. We then applied a linear time warping function to the recorded EEG signals, so as to suppress the fluctuations in tapping onsets. In these time-warped signals, movement-related activity can be expected to become strictly periodic, whereas auditory-related activity will exhibit fluctuations. Therefore, movement- related activity is expected to elicit an SS-EP concentrated at the mean tapping frequency, whereas auditory-related activity is expected to elicit an attenuated SS-EP spread out in the frequency domain. Frequency-domain analysis revealed a clear SS-EP in both the original and the time-warped EEG signals. Notably, the scalp topographies of the two SS-EPs were markedly different, indicating that they predominantly reflect activity originating from different brain regions. This is explained by the fact that in the original EEG signals, the SS-EP peak is more strongly determined by auditory-related activity, while in the time-warped EEG signals, it is more strongly determined by movement-related activity. Supporting this interpretation, the scalp topography of the SS-EP obtained using the original EEG signal was maximal over fronto-central regions, compatible with bilateral activity originating from auditory cortices. In the time-warped EEG signals, the scalp topography extended over frontal and parietal regions contralateral to the tapping hand, compatible with activity originating from the contralateral sensorimotor cortex.