ABSTRACTBackground:Sleep disturbance is a common consequence of providing care to a loved one with Alzheimer's disease (AD). We explored the usefulness of the Pleasant Events and Activity Restriction (PEAR) model for predicting multiple domains of sleep disturbance.Methods:Our sample consisted of 125 spousal AD caregivers. Participants completed the Pittsburg Sleep Quality Index (PSQI) and were questioned regarding the frequency with which they engaged in pleasant events and the extent to which they felt restricted in engaging in social and recreational activities in the past month. Participants were classified into one of three groups: HPLR = High Pleasant Events + Low Activity Restriction (= reference group; N = 38); HPHR/LPLR = either High Pleasant Events + High Activity Restriction or Low Pleasant Events + Low Activity Restriction (N = 52); and LPHR: Low Pleasant Events + High Activity Restriction (N = 35). These three groups were compared on the seven subscales of the PSQI.Results:Significant differences were found between the HPLR and LPHR groups on measures of subjective sleep quality, sleep latency, habitual sleep efficiency, sleep disturbance, and daytime dysfunction. Additionally, significant differences were found between the HPLR and HPHR/LPLR groups on subjective sleep quality, sleep latency, and habitual sleep efficiency, and between the HPHR/LPLR and LPHR groups on sleep disturbance and daytime dysfunction.Conclusions:This study provides broad support for the PEAR model and suggests that interventions focusing on behavioral activation may potentially provide benefits to non-affective domains including sleep.