Decoding working memory task condition using magnetoencephalography source level long-range phase coupling patterns

摘要

Objective. The objective of the study is to identify phase coupling patterns that are shared acrosssubjects via a machine learning approach that utilises source space magnetoencephalography(MEG) phase coupling data from a working memory (WM) task. Indeed, phase coupling of neuraloscillations is putatively a key factor for communication between distant brain areas and istherefore crucial in performing cognitive tasks, including WM. Previous studies investigating phasecoupling during cognitive tasks have often focused on a few a priori selected brain areas or aspecific frequency band, and the need for data-driven approaches has been recognised. Machinelearning techniques have emerged as valuable tools for the analysis of neuroimaging data since theycatch fine-grained differences in the multivariate signal distribution. Here, we expect that thesetechniques applied to MEG phase couplings can reveal WM-related processes that are shared acrossindividuals. Approach. We analysed WM data collected as part of the Human Connectome Project.The MEG data were collected while subjects (n = 83) performed N-back WM tasks in two differentconditions, namely 2-back (WM condition) and 0-back (control condition). We estimated phasecoupling patterns (multivariate phase slope index) for both conditions and for theta, alpha, beta,and gamma bands. The obtained phase coupling data were then used to train a linear supportvector machine in order to classify which task condition the subject was performing with anacross-subject cross-validation approach. The classification was performed separately based on thedata from individual frequency bands and with all bands combined (multiband). Finally, weevaluated the relative importance of the different features (phase couplings) for classification by themeans of feature selection probability. Main results. The WM condition and control condition weresuccessfully classified based on the phase coupling patterns in the theta (62% accuracy) and alphabands (60% accuracy) separately. Importantly, the multiband classification showed that phasecoupling patterns not only in the theta and alpha but also in the gamma bands are related to WMprocessing, as testified by improvement in classification performance (71%). Significance. Ourstudy successfully decoded WM tasks using MEG source space functional connectivity. Ourapproach, combining across-subject classification and a multidimensional metric recentlydeveloped by our group, is able to detect patterns of connectivity that are shared across individuals.In other words, the results are generalisable to new individuals and allow meaningful interpretationof task-relevant phase coupling patterns.