P 80. Corticospinal synchrony might mediate feature extraction in implicit motor learning

Abstract

Dynamic adaptation and improvements of motor skills are achieved in terms of implicit motor learning. Neuronal and neuromuscular synchronization (coherence) contribute significantly to the accuracy of motor control. However, it remains to be defined whether this also applies for corticospinal network adaptations during the process of implicit motor learning. We aimed to gain insight into the cortical and corticospinal integration of implicit motor learning. We tested 15 healthy subjects using the Serial Reaction Time Task (SRTT): after a random baseline block [Bl] subjects performed four sequential blocks ([Sa] to [Sd]), followed by a random block [Rc], another sequential block [Se] and again one random block [Rd]. We recorded simultaneous 64-channel EEG and surface EMG of the Flexor Digitorum (FD) and Extensor Digitorum (ED) muscles. We analyzed the performance data, power spectral density (PSD) of cortical and muscular activation, as well as event-related cortical time–frequency power spectra and corticospinal time–frequency coherence spectra using the Morlet wavelet algorithm. Time-locked data was analyzed with respect to the button-press and pooled for performance blocks. As expected, subjects showed a significant decrease of reaction times in the sequence blocks [Sc] and [Sd] reflecting implicit motor learning (repeated measures ANOVA revealed significant differences between blocks ( F 7 = 4256, P = 0.001). [Sc] (one-tailed paired t -test, t 11 = 3.102, P = 0.04) and [Sd] (one-tailed paired t -test, t 11 = 2.888, P = 0.03) displayed significantly shorter reaction times compared to [Bl]). Corticomuscular coherence (CMC) time-locked to FD activation increased after the baseline, reached maximal peak amplitude when reaction time decreased in the sequence blocks and was strongest in low and high gamma (30–45 Hz and 60–100 Hz). The gamma CMC started about 200 ms before, ended about 300 ms after button-press and showed a biphasic time-course with the maximal peak amplitude from 60 to 80 ms after button-press ( Fig. 1 ). The main finding is a significant modulation of gamma band CMC in the sequence blocks compared to [Bl]. Interestingly, this increase occurred in a later phase of the motor response after button press ( Fig. 1 ) indicating stronger ‘afferent’ directionality and was already observed in the first sequence blocks even before reaction time decreased. Therefore, gamma CMC might contribute to the extraction of the implicit motor sequence and facilitate the cortical reorganization of motor program encoding.