The Electrocortical Correlates of Fluctuating States of Attention During Vigilance Tasks

This study investigated the electrocortical correlates of attention. Sixteen subjects (seven females, nine males) engaged in a forty-minute target-detection vigilance task. In addition to the demands of the vigilance task, the subjects were asked to press a button during the vigilance task if they had experienced task-unrelated thoughts (TUTs) or daydreams. Continuous electroencephalograms (EEG) were recorded from the subjects during the entire task. During the vigilance task, task-irrelevant tones were presented every 2-4 seconds. Event-related potentials (ERPs) elicited by these tones were generated. The task was divided into four ten-minute periods. The analysis of the performance data indicated performance decrements across periods. The performance decrement was reflected by increased reaction times (RT) to correct detections, and decreased proportion of correct detections, or A'. As the task progressed, the subjects became more conservative with their responses which was indicated by a significant decrease in the subjects' response criterion, or B''. In addition, the analysis revealed a significant increase in the number of reports of TUTs or daydreams over the course of the task. The EEG data were analyzed for differences in absolute power across periods as well as before and after TUT responses. ERPs elicited by the task-irrelevant tones were analyzed in the same manner. The quantitative analysis of the EEG data (QEEG) indicated no significant difference in the absolute power of the four frequency bands (delta, theta, alpha, and beta) across periods. However, when the QEEG were converted into band ratios (beta/alpha and beta/alpha+theta), the pre-TUT and the post-TUT QEEG were found to be significantly different in the posterior parietal sites. The ERP amplitude components were not significantly different before and after a TUT response or across periods, however excessive artifact contamination of the ERP data resulted in the elimination of ERP data from 2/3 of the subjects. This may have resulted in decreased power and therefore reduced the chances of finding a significant differences in the ERP data.