Attention-related modulation of frontal midline theta oscillations in cingulate cortex during a spatial cueing Go/NoGo task.

Recently frontal midline theta (FMθ, 4-8 Hz) oscillations have been consistently reported and proposed as the potential neural mechanism for response inhibition, a core component of human executive functions. However, it remains unclear to what extent the increase of FMθ power during response inhibition tasks is influenced by other non-inhibitory cognitive processes that are usually required for the tasks. In this study, we examined attention-related effects on FMθ during response inhibition by revisiting the EEG data from healthy young adults (N = 30) while performing a spatial cueing Go/NoGo task (Hong et al., 2017). Such an experimental design enabled us to manipulate selective attention that was voluntarily deployed to the stimuli triggering response inhibition. Although commission error rates were low for the NoGo trials, response preparation and prepotent motor activities were qualified by observing significant contingent negative variation (CNV) and lateralized readiness potential (LRP) in both the Go and the NoGo trials. We observed an increase of FMθ power in both attended and ignored conditions of the NoGo trials compared with the Go trials, while FMθ power was significantly smaller in the ignored condition than in the attended condition. Furthermore, source localization analysis suggested cingulate cortex as the main origin for FMθ in the NoGo trials, and FMθ power differences between the attended and ignored conditions were localized in cingulate gyrus. We further provided evidence for attention-dependent FMθ generation by correlating theta source density with ERP signatures of attention (N1), response preparation (CNV), and motor activity (LRP). Overall, our findings suggest that during the Go/NoGo task, the increase of FMθ power was not purely driven by response inhibition, and part of FMθ generators in cingulate gyrus was significantly modulated by attention-related processing (possibly conflict monitoring).