Resumen
We tested whether changes in prestimulus neural activity predict behavioral performance (decision time and errors) during a prolonged visual task. The task was to classify ambiguous stimuli?Necker cubes; manipulating the degree of ambiguity from low ambiguity (LA) to high ambiguity (HA) changed the task difficulty. First, we assumed that the observer?s state changes over time, which leads to a change in the prestimulus brain activity. Second, we supposed that the prestimulus state produces a different effect on behavioral performance depending on the task demands. Monitoring behavioral responses, we revealed that the observer?s decision time decreased for both LA and HA stimuli during the task performance. The number of perceptual errors lowered for HA, but not for LA stimuli. EEG analysis revealed an increase in the prestimulus 9?11 Hz EEG power with task time. Finally, we found associations between the behavioral and neural estimates. The prestimulus EEG power negatively correlated with the decision time for LA stimuli and the erroneous responses rate for HA stimuli. The obtained results confirm that monitoring prestimulus EEG power enables predicting perceptual performance on the behavioral level. The observed different time-on-task effects on the LA and HA stimuli processing may shed light on the features of ambiguous perception.