Neural correlates of motor learning: Network communication versus local oscillations

Anaïs Mottaz; Branislav Savic; Leslie Allaman; Adrian G Guggisberg

doi:10.1162/netn_a_00374

Neural correlates of motor learning: Network communication versus local oscillations

Netw Neurosci. 2024 Oct 1;8(3):714-733. doi: 10.1162/netn_a_00374. eCollection 2024.

Authors

Anaïs Mottaz^{1

2

3}, Branislav Savic⁴, Leslie Allaman¹, Adrian G Guggisberg^{1

4}

Affiliations

¹ Division of Neurorehabilitation, Department of Clinical Neurosciences, University Hospital of Geneva, University of Geneva, Switzerland.
² SIB Text Mining Group, Swiss Institute of Bioinformatics, Carouge, Switzerland.
³ BiTeM Group, Information Sciences, HES-SO/HEG, Carouge, Switzerland.
⁴ Division of Neurorehabilitation, Department of Neurology, Inselspital, Bern University Hospital, University of Bern, Switzerland.

Abstract

Learning new motor skills through training, also termed motor learning, is central for everyday life. Current training strategies recommend intensive task-repetitions aimed at inducing local activation of motor areas, associated with changes in oscillation amplitudes ("event-related power") during training. More recently, another neural mechanism was suggested to influence motor learning: modulation of functional connectivity (FC), that is, how much spatially separated brain regions communicate with each other before and during training. The goal of the present study was to compare the impact of these two neural processing types on motor learning. We measured EEG before, during, and after a finger-tapping task (FTT) in 20 healthy subjects. The results showed that training gain, long-term expertise (i.e., average motor performance), and consolidation were all predicted by whole-brain alpha- and beta-band FC at motor areas, striatum, and mediotemporal lobe (MTL). Local power changes during training did not predict any dependent variable. Thus, network dynamics seem more crucial than local activity for motor sequence learning, and training techniques should attempt to facilitate network interactions rather than local cortical activation.

Keywords: Electroencephalography; Functional connectivity; Motor sequence learning; Neural plasticity.

Plain language summary

Both, local and network processing mechanisms support motor sequence learning. The aim of the present study was to compare the impact of these two processing types on motor learning. We measured EEG before, during, and after a finger-tapping task (FTT) in 20 healthy subjects. The results showed that only network dynamics, measured with functional connectivity, could predict learning, long-term expertise, and consolidation. Conversely, local activity, measured with event-related power decrease, did not predict any dependent measure. Specifically, network interactions of the primary motor area, the striatum, and the medial temporal lobe correlated with learning performance. Therefore, network dynamics seem more crucial than local activity for motor sequence learning and training techniques should facilitate network interactions rather than local cortical activation.