|Title||Looking for neurophysiological correlates of brain-computer interface learning|
|Publication Type||Conference Paper|
|Year of Publication||2019|
|Authors||Corsi, M-C, Chavez, M, Schwartz, D, George, N, Hugueville, L, Kahn, A, Dupont, S, Bassett, D, Fallani, FDe Vico|
Non-invasive Brain-Computer Interfaces (BCIs) are largely used to produce thought-provoked action, by exploiting the ability of subjects to voluntary modulate their brain activity through mental imagery. Despite its clinical applications [Jin, 2012; Prasad, 2010], controlling a BCI appears to be a learned skill. Several weeks or even months are needed to reach relatively high-performance in BCI control, without being sufficient for 15 to 30% of the users [Allison, 2010; Vidaurre, 2010]. This gap has motivated a deeper understanding of mechanisms associated with motor imagery (MI) tasks [Kaiser, 2014; Perdikis, 2014]. If similarities have been shown between MI-based BCI learning and motor sequence learning [McDougle, 2016; Wander, 2013], our understanding of the involved processes is still incomplete. Among the advanced reasons are the lacks of longitudinal studies long enough to observe consolidation effects associated with learning process, and of proper learning metrics based on the neurophysiology [Perdikis, 2018]. Here, we expected that MI-BCI learning is associated with the recruitment of areas distributed across the cortex beyond those targeted by the BCI. We also hypothesized that the associated properties, in terms of activations and functional connectivity, predict the learning success.