Working memory (WM) is the brain’s ability to retain information that is not directly available from the sensory systems. WM retention is accompanied by sustained firing rate modulation and changes of the large-scale oscillatory profile. Among other changes, beta-band activity elevates in task-related regions, presumably stabilizing WM retention. Alpha-band activity, in turn, is stronger in task-irrelevant regions, serving to protect WM trace from distracting information. Although a large body of experimental evidence links neural oscillations to WM functions, theoretical understanding of their interrelations is still incomplete. In this study, we used a computational approach to explore a potential role of beta and alpha oscillations in control of WM stability. First, we examined a single bistable module that served as a discrete object representation and was resonant in the beta-band in the active state. We demonstrated that beta-band input produced differentially stronger excitatory effect on the module in the active state compared to the background state, while this difference decreased with the input frequency. We then considered a system of two competing modules, selective for a stimulus and for a distractor, respectively. We simulated a task, in which a stimulus was loaded into the first module, then an identical oscillatory input to both modules was turned on, after which a distractor was presented to the second module. We showed that beta-band input prevented loading of high-amplitude distractors and erasure of the stimulus from WM. On the contrary, alpha-band input promoted loading of low-amplitude distractors and the stimulus erasure. In summary, we demonstrated that stability of WM trace could be controlled by global oscillatory input in a frequency-dependent manner via controlling the level of competition between stimulus-encoding and distractor-encoding circuits. Such control is possible due to difference in the resonant and non-linear properties between the background and the active states.
bioRxiv Subject Collection: Neuroscience