A statistical comparison of EEG time- and time-frequency domain representations of error processing

Area of Science:

• Cognitive Neuroscience
• Electrophysiology
• Human Behavior

Background:

• Successful behavior requires error detection and adjustment.
• Electrophysiological signatures of error processing include the error-related negativity (ERN) and delta/theta power.
• The relationship between phase-locked (ERN) and non-phase-locked (delta/theta power) activity remains unclear.

Purpose of the Study:

• To investigate the relationship between the ERN and delta/theta power in error processing.
• To determine the predictive power of ERN and delta/theta power on single-trial accuracy.
• To explore the contribution of both overlapping and independent information from these signals.

Main Methods:

• Utilized single-trial analysis of electrophysiological data.
• Employed logistic regression to model trial accuracy prediction.
• Compared prediction models using ERN, delta/theta power, or both.

Main Results:

• Both ERN and delta/theta power independently predicted single-trial accuracy with approximately 70% accuracy.
• A combined model incorporating both ERN and delta/theta power yielded superior prediction accuracy.
• Phase-locked EEG activity (ERN) was found to be as predictive as non-phase-locked activity (delta/theta power).

Conclusions:

• The phase-locked and non-phase-locked components of the EEG signal offer comparable predictive value for response accuracy.
• Single-trial ERPs and delta/theta power contain both shared and unique information regarding error processing.
• Integrating both ERN and delta/theta power enhances the understanding of error detection mechanisms.