Repeated acquisition and performance chamber for mice: A paradigm for assessment of spatial learning and memory

Molecular genetic manipulation of the mouse offers the possibility of eluci dating the function of individual gene products in neural systems underlyin g learning and memory. Many extant learning paradigms for mice rely on nega tive reinforcement, involve simple problems that are relatively rapidly acq uired and thus preclude timecourse assessment, and may impose the need to u ndertake additional experiments to determine the extent to which noncogniti ve behaviors influence the measures of learning. To overcome such limitatio ns, a multiple schedule of repeated acquisition and performance was behavio rally engineered to assess learning vs rote performance within-behavioral t est session and within-subject utilizing an apparatus modified from the rat (the repeated acquisition and performance chamber; RAPC). The multiple sch edule required mice to learn a new sequence of door openings leading to sac charin availability in the learning component during each session, while th e sequence of door openings for the performance component remained constant across sessions. The learning and performance components alternated over t he course of each test session, with different auditory stimuli signaling w hich component was currently in effect. To validate this paradigm, learning vs performance was evaluated in two inbred strains of mice: C57BL/6J and 1 29/SvJ. The hippocampal dependence of this measure was examined in lesioned C57BL/6J mice. Both strains exhibited longer latencies and higher errors i n the learning compared to the performance component and evidenced declines in both measures across the trials of each session, consistent with an acq uisition phenomenon. These same measures showed little or no evidence of ch ange in the performance component. Whereas three trials per session were ut ilized with C57BL/65 mice in each component, behavior of 129/SvJ mice could only be sustained for two trials per component per session, demonstrating differences in testing capabilities between these two strains under these e xperimental conditions and thus precluding the ability to make systematic s train comparisons of learning capabilities. Hippocampal lesions in C57BL/6J mice resulted in substantially longer latencies and increased errors in th e learning but not the performance component, demonstrating the importance of this region to spatial learning as measured in the RAPC. In aggregate, t his positive reinforcement-based operant paradigm to evaluate murine spatia l learning detects strain differences and hippocampal dependence and permit s explicit differentiation of the impact of noncognitive contributions to l earning measures on a within-subject, within-session basis. (C) 2000 Academ ic Press.