Recent research suggests that sustained attention is punctuated by periodic lapses that produce cyclic variations in sustained human performance approximately every 90 s. It has been suggested that a norepinephrine based arousal system may be responsible for these variations. However, both cholinergic and noradrenergic pathways are known to mediate attention. Furthermore, the current understanding of the 1.5 min cycle is derived from post hoc analysis of observed data and suffers from a lack of deductive, theoretically derived and tested models. The current study used reaction time and electroencephalographic data collected during a 15 min visual continuous performance task to determine if cholinergic pathways played a role in cyclic arousal and provide deductive support for the 1.5 min cycle. A spectral analysis procedure was used to detect the prevalence of the 1.5 min cycle in performance and cortical activity. The 1.5 min cycle was found in performance but not cortical arousal. Sinusoidal and pulse-wave models were then tested for goodness of fit. Models with a periodicity fixed near 90 s predicted performance better than other periodicities. The results support the role of norepinephrine, but not acetylcholine, in mediating cyclic arousal, and the role of the 90 s cycle in predicting performance.