Within each subject, microstates were identified as continuous epochs within which all topographies were assigned to the same class. No minimal microstate duration was required. Individual microstate profiles were computed for each class, consisting of mean microstate duration (“duration”), mean number of microstates per second (“occurrence”) and percentage of total analysis time occupied in that state (“percent total time”). For each microstate measure, a two way ANOVA was performed with subject group as between, and microstate class as within subject factors. Significant results including group were post hoc tested using unpaired, double-ended t-tests. Relationships between each microstate measure and the current severity (in partial remission = 1, mild = 2, moderate = 3 and severe = 4) of panic attacks according to the DSM-III-R criteria was assessed using Spearman's rank correlation coefficient. Statistical significance was defined as p<0.05.
