Computerized assessment of sustained attention: interactive effects of task demand, noise, and anxiety

In a sample of 163 college undergraduates, the effects of task demand, noise, and anxiety on Continuous Performance Test (CPT) errors were evaluated with multiple regression and multivariate analysis of variance. Results indicated significantly more omission errors on the difficult task. Complex interaction effects of noise and self-reported anxiety yielded more omissions in quiet intermittent white noise, particularly for high-anxious subjects performing the difficult task. Anxiety levels tended to increase from pretest to posttest, particularly for low-anxious subjects in the quiet, difficult-task condition, while a decrease was seen for high-anxious subjects in the loud, easy-task condition. Commission errors were unrelated to any predictor variables, suggesting that "attention" cannot be considered a unitary phenomenon. The variety of direct and interactive effects on vigilance performance underscore the need for clinicians to use a variety of measures to assess attentional skills, to avoid diagnosis of attention deficits on the basis of a single computerized task performance, and to rule out anxiety and other contributors to poor vigilance task performance.