The association between oscillatory burst features and human working memory accuracy

Abstract

Oscillatory power across multiple frequency bands has been associated with distinct working memory (WM) processes. Recent research has shown that previous observations based on averaged power are driven by the presence of transient, oscillatory burst-like events, particularly within the alpha, beta, and gamma bands. However, the interplay between different burst events in human WM is not well understood. The current EEG study aimed to investigate the dynamics between alpha (8-12 Hz)/beta (15-29 Hz) and high frequency activity (HFA; 55-80 Hz) bursts in human WM, particularly burst features and error-related deviations during the encoding and maintenance of WM in healthy adults. Oscillatory burst features within the alpha, beta and HFA bands were examined at frontal and parietal electrodes in healthy young adults during a Sternberg working memory task. Averaged power dynamics were driven by oscillatory burst features, most consistently the burst rate and burst power. Alpha/beta and HFA bursts displayed complementary roles in WM processes, in that alpha and beta bursting decreased during encoding and increased during delay, while HFA bursting had the opposite pattern, i.e., increased during encoding and decreased during the delay. Critically, weaker variation in burst dynamics across stages was associated with incorrect responses and impaired overall task performance. Together, these results indicate that successful human WM is dependent on the rise and fall interplay between alpha/beta and HFA bursts, with such burst dynamics reflecting a novel target for the development of treatment in clinical populations with WM deficits.