ИСТИНА

Maintenance of cognitive control depends upon performance monitoring, including error detection, which can be either internal or driven by an external feedback. Only unexpected error information is known to evoke respective brain processing. Considering fluctuations of cognitive control within a subject over trials, we expect three basic scenarios: (1) Optimal levels of both task-specific processing and motor threshold provide for fast correct responses. (2) Lowered motor threshold results in increased probability of committing fast erroneous responses, which can be detected internally since specific processing is not necessarily compromised and both correct and incorrect motor programs are active. (3) Lowered level of task-specific processing results in uncertainty and a delay in responses, thus decreasing the likelihood of internal error detection and increasing unexpectedness of the external feedback. In the current study, we focused on the distinction between fast and slow behavioral responses. EEG data were collected from 50 healthy right-handed adults. We used an auditory condensation tasks involving a two-alternative choice. The task did not push the participants towards fast forced choices and employed four target stimuli that were procedurally equable and did not require inhibition of prepotent responses. For each participant, responses were classified into fast and slow relative to individual median response times. Visual feedback was given 500 ms after the response, thus allowing measuring both error-related and feedback-related brain events. Previous reports suggest that after errors, either post-response or post-feedback fronto-central theta power is enhanced, reflecting error detection, while post-feedback frontal beta power is increased after correct responses, reflecting positive learning signals. We found a much greater error-related theta power increase at fronto-central sites for fast responses compared with slow responses. Conversely, feedback-related theta activity at fronto-central sites was more prominent for slow responses compared with fast responses. Only for slow correct responses, positive feedback induced a post-feedback increase in frontal beta power. Our findings are compatible with the view that fast responses involve little internal uncertainty, thus allowing internal error detection, while making the external feedback predictable and non-informative. On the other hand, slow responses imply higher internal uncertainty, making the external feedback unexpected and thus informative in terms of reinforcement learning.