Abstract:

Following errors, participants usually recruit more cognitive resources to change error-related behaviors; this phenomenon is termed post-error adjustment. Generally, behavioral adjustments in post-error trials behave as slower subsequent responses and improved accuracy. It is worth noting that we cannot successfully perceive every error we commit in daily life. Several studies found that post-error slowing occurred only after aware errors, suggesting that only aware errors contribute to the phenomenon of post-error adjustment. Moreover, these studies emphasized the role of top-down control in the processing of error awareness. However, a few studies came to the opposite conclusion, finding that post-error adjustment could be modulated by unaware errors in an implicit manner. These studies emphasized the role of bottom-up control in the processing of error awareness. Notably, previous studies have demonstrated that post-error adjustment involves both proactive and reactive cognitive control. Proactive control refers to a goal-driven manner that is actively maintained with sustained attention before the occurrence of cognitively demanding events. Reactive control refers to a bottom-up manner, in which the attentional control is mobilized when the goal-related event is reactivated. Thus, whether different control strategies are adopted by aware and unaware errors remains unclear.
To investigate the above issue, we recruited 36 participants to execute an error awareness task based on the go/no-go task. However, data from five participants were removed due to poor EEG records or poor behavioral performance. In the go/no-go error awareness task, participants were instructed to withhold their responses in certain circumstances. The first was when a word was presented on two consecutive trials, and the second was when the font color of the word and its meaning were inconsistent. Additionally, the usage of an error signal button might lead to a response bias toward signaling or not signaling an error. If participants tended to signal errors, they might signal their correct responses as errors, increasing the false alarm rates. If participants did not tend to signal errors, aware errors might be classed as unaware errors. In this case, the measurement of unaware errors might be contaminated by potential conscious error trials. Thus, participants were instructed to respond to indicate their perceived response accuracy in both error and correct cases during the rating screen in the current experiment.
Since previous studies have found that neural oscillations reveal the processing of proactive and reactive control, the time-frequency analysis is conducted in this experiment. It has been suggested that the alpha band (8-14 Hz) reflects the trial-by-trial behavioral adjustment. Thus, alpha power is chosen as the neural indicator. As a result, the post-error reaction time indicated two dissociated behavior patterns with speeding up following aware errors and slowing down following unaware errors. However, accuracy in trials following aware and unaware errors was significantly higher than for trials following correct go. At the neural level, alpha (-500 to 500 ms) power was stronger for aware errors than for unaware errors. Moreover, the alpha was activated before the subjective report of error awareness for aware errors, but the alpha was activated after the subjective report of error awareness for unaware errors.
Current behavioral results showed that aware and unaware errors both successfully optimized post-error performance, but the two error types adopted different methods to adjust post-error behaviors. The time-frequency analysis revealed that aware errors led to sustained attention control after responses, but unaware errors led to temporary attention control induced by the subjective report of error awareness. Therefore, these findings might suggest that the adjustments following aware errors were based on a strategy such as proactive control, whereas the adjustments following unaware errors were based on a strategy such as reactive control.

Key words: error awareness, post-error adjustment, alpha, proactive control, reactive control