Earlier work by our group and others has pointed to a role for movement during sleep in sleep-panic attacks. Specifically, our group has reported that panic disorder patients, as a group, appear to move more during sleep than agematched controls, whereas the subgroup of panic disorder who ''panic'' during sleep move less on the nights they experience sleep-panic attacks than they do on nights without sleep-panic attacks. We studied the movement of sleep-panic patients in a more detailed fashion with more than one sleep movement index. Fourteen patients with sleep-panic attacks were compared with 14 waking panic patients, 13 social phobic patients, and 14 normal controls. Subjects from the other groups were age matched to the sleep-panic group. Their comparison study night corresponded to the night number of the sleep-panic attack. Sleep-panic patients did move less on panic nights than did the normal controls on the corresponding sleep-panic night according to two separate sleep movement indices. Although not statistically significant, sleep-panic patients also moved less on their panic night than did either of the other anxious groups on corresponding nights. Rechtshaffen and Kales' Movement Time (MT) measure appears to overestimate actual min of movement during sleep in all subjects. The movement noted in sleep-panic patients may have some role in the pathophysiology of sleep-panic attacks. A possible mechanism is explained.