Although sleep and circadian rhythm disturbances are significant features of bipolar disorder and are associated with illness severity and recurrence, less is known about their significance prior to illness onset. Therefore, this study investigated these variables using objective assessment methods within a sample at high-risk of bipolar spectrum disorders. Thirty-one high-risk and 24 age- and gender-matched control participants wore an actigraph for 7 days to obtain sleep and circadian activity data. Self-report measures of mood, bed and get-up times, and cognitive style were also obtained. High-risk participants exhibited greater variability in duration, fragmentation, and efficiency of sleep, shorter sleep duration, and later and more variable bedtimes than controls. They also had lower relative amplitude of activity patterns and made more positive self-appraisals for hypomanic experiences. Logistic regression showed that positive self-appraisal and variability in bedtime significantly discriminated between the two groups. Results suggest that sleep and circadian activity differences are apparent in high-risk participants and, therefore, may not simply represent an artifact of the illness. Such individuals also report cognitive styles consistent with those observed in bipolar patients. Relevance of these results to current evolutions of the instability theory of bipolar disorder is discussed. © 2009 Wiley Periodicals, Inc. J Clin Psychol 65: 1–16, 2009.