Lateralised dynamic modulations of corticomuscular coherence associated with bimanual learning of rhythmic patterns

Abstract

Human movements are spontaneously attracted to auditory rhythms, triggering an automatic activation of the motor system, a central phenomenon to music perception and production. Cortico-muscular coherence (CMC) in the theta, alpha, beta and gamma frequencies has been used as an index of the synchronisation between cortical motor regions and the muscles. Here we investigated how learning to produce a bimanual rhythmic pattern composed of low- and high-pitch sounds affects CMC in the beta frequency band. Electroencephalography (EEG) and electromyography (EMG) from the left and right First Dorsal Interosseus and Flexor Digitorum Superficialis muscles were concurrently recorded during constant pressure on a force sensor held between the thumb and index finger while listening to the rhythmic pattern before and after a bimanual training session. During the training, participants learnt to produce the rhythmic pattern guided by visual cues by pressing the force sensors with their left or right hand to produce the low- and high-pitch sounds, respectively. Results revealed no changes after training in overall beta CMC or beta oscillation amplitude, nor in the correlation between the left and right sides for EEG and EMG separately. However, correlation analyses indicated that left- and right-hand beta EEG-EMG coherence were positively correlated over time before training but became uncorrelated after training. This suggests that learning to bimanually produce a rhythmic musical pattern reinforces lateralised and segregated cortico-muscular communication.

Figure 1

Experimental Design. Sound was delivered binaurally. Force sensors were held between index finger and thumb. EMG was recorded from left and right FDI and FDS muscles. EEG was recorded with 64 channels and analyses were conducted using C3 and C4 data. Listening (L) trials before and after the 8 training (T) trials were considered as pre and post-training, respectively.

Figure 2

Absolute time difference between the inter-sound intervals that participants were instructed to perform and the intervals that participants actually produced before and after training. The black dot and bars represent the mean and confidence interval. Grey dots represent averaged data for individual participants. The significant reduction on the absolute time differences throughout training indicate successful training with higher production precision gained through practice.

Figure 3

Mean and confidence interval of the correlation coefficient between right and left hand beta CMC for FDI (pink) and FDS (blue) muscles before and after training (represented by larger dots and bars). Smaller dots represent data for individual participants. The left and right hand FDS muscle that were positively correlated pre-training became uncorrelated post-training, indicating segregation of left and right hand after training.

Figure 4

Cortico-muscular coherence in the beta band (16–36 Hz). Panel (A) shows demeaned beta CMC data throughout the rhythm pattern for a representative participant (ID 17) before (Pre) and after (Post) training for the right hand-C3 (blue) and left hand-C4 (pink) for FDI (first dorsal interosseus) and FDS (flexor digitorum superficialis) muscles. The shaded colours represent the low-pitch (pink) and high-pitch (blue) sounds that were learnt to be played using left and right hand, respectively. Panel (B) represents the topoplots of the same participant and the average of all participants for beta CMC for FDI and FDS muscles before (Pre) and after (Post) training.