Traditional behavioral tests, such as the open field test, measure an animal's responsiveness to a novel environment. However, it is generally difficult to assess whether the behavioral response obtained from these tests relates to the expression level of motor activity and/or to avoidance of anxiogenic areas. Here, an automated home cage environment for mice was designed to obtain independent measures of motor activity levels and of sheltered feeding preference during three consecutive days. Chronic treatment with the anxiolytic drug chlordiazepoxide (5 and 10 mg/kg/day) in C57BL/6J mice reduced sheltered feeding preference without altering motor activity levels. Furthermore, two distinct chromosome substitution strains, derived from C57BL/6J (host strain) and A/J (donor strain) inbred strains, expressed either increased sheltering preference in females (chromosome 15) or reduced motor activity levels in females and males (chromosome 1) when compared to C57BL/6J. Longitudinal behavioral monitoring revealed that these phenotypic differences maintained after adaptation to the home cage. Thus, by using new automated behavioral phenotyping approaches, behavior can be dissociated into distinct behavioral domains (e.g., anxiety-related and motor activity domains) with different underlying genetic origin and pharmacological responsiveness.