Abstract
Working memory (WM) is required to bridge the time between the moment of sensory perception and the usage of the acquired information for subsequent actions. Its frequent and pharmacoresistent impairment in mental health disorders urges the development of rodent paradigms through back-translation of human WM tests, ideally avoiding the confounds of alternation-based assays. Here we show, that mice can acquire a delayed-matching-to-position (DMTP) operant spatial WM (SWM) paradigm that is akin to the combined attention and memory (CAM) task previously developed for rats, and that relies on a 5-choice wall [5-CSWM, 5-choice based operant testing of SWM (5-CSWM)]. Requiring ca. 3 months of daily training with a non-illuminated operant box in the default state, mice could attain a performance level of ≥70% choice accuracy with short (2 s) delays in the DMTP 5-CSWM task. Performance decreased with extended delays, as expected for WM processes. Modafinil (15 and 30 mg/kg) and guanfacine (0.3 and 1 mg/kg) showed no consistent efficacy in enhancing task performance. We also found, that mice did not improve beyond chance level, when trained in the DNMTP-version of the 5-CSWM. Our results outline the methodical possibility and constraints of assessing spatial WM in mice with an operant paradigm that provides high control over potentially confounding variables, such as cue-directed attention, motivation or mediating strategies like body-positioning.