Neural dynamics of stimulus-response representations during inhibitory control

Abstract

The investigation of action control processes is one major field in cognitive neuroscience, and several theoretical frameworks have been proposed. One established framework is the "Theory of Event Coding" (TEC). However, only rarely, this framework has been used in the context of response inhibition and how stimulus-response association or binding processes modulate response inhibition performance. Particularly the neural dynamics of stimulus-response representations during inhibitory control are elusive. To address this, we examined N=40 healthy controls and combined temporal EEG signal decomposition with source localization and temporal generalization multivariate pattern analysis (MVPA). We show that overlaps in features of stimuli used to trigger either response execution or inhibition compromised task performance. According to TEC, this indicates that binding processes in event file representations impact response inhibition through partial repetition costs. In the EEG data, reconfiguration of event files modulated processes in time windows well-known to reflect distinct response inhibition mechanisms. Crucially, event file coding processes were only evident in a specific fraction of neurophysiological activity associated with the inferior parietal cortex (BA40). Within that specific fraction of neurophysiological activity, the decoding of the dynamics of event file representations using temporal generalization MVPA suggested that event file representations are stable across several hundred milliseconds, and that event file coding during inhibitory control is reflected by a sustained activation pattern of neural dynamics.

Protocol to decode representations from EEG data with intermixed signals using temporal signal decomposition and multivariate pattern-analysis

“…Thus, it is also necessary to consider the representational content of such signals more closely. The dynamics of the representational content can be examined using multivariate pattern analysis (MVPA) ( Carlson et al., 2011 ; King and Dehaene, 2014 ; Prochnow et al., 2021 ; Takacs et al., 2020a , 2021 ).…”

“…Importantly, the application of the current protocol does not require more detailed information on the task or theoretical knowledge on cognitive flexibility. Optional: For the interested reader, we recommend the method sections of previous papers that successfully combined RIDE and MVPA to study stimulus-response binding ( Takacs et al., 2020a ), response-response binding ( Takacs et al., 2021 ) and inhibitory control ( Prochnow et al., 2021 ). Note: We use EEG data that was recorded in a dimly lit room with 60 Ag/AgCl electrodes, placed in equidistant positions in an elastic cap (EasyCap, Germany).…”

“…Optional: For the interested reader, we recommend the method sections of previous papers that successfully combined RIDE and MVPA to study stimulus-response binding ( Takacs et al., 2020a ), response-response binding ( Takacs et al., 2021 ) and inhibitory control ( Prochnow et al., 2021 ).…”

“…The combination of temporal signal decomposition and decoding methods seem not only feasible but also desirable in areas of human cognition in which multiple aspects of coding are intermixed ( Petruo et al., 2021 ; Prochnow et al., 2021 ; Takacs et al., 2020a , 2021 ). This method combination has the potential to increase the number of MVPA applications in fields where it is important to disentangle different aspects of the EEG signal.…”

“…This method combination has the potential to increase the number of MVPA applications in fields where it is important to disentangle different aspects of the EEG signal. It was suggested, that such distinction can be important cognitive control ( Petruo et al., 2021 ; Prochnow et al., 2021 ), that has the overlapping motor and non-motor components. However, classes should be defined even more cautiously to decode higher-level, abstract representations.…”