The rodent tasks with food rewards are useful methods to evaluate memory functions, including hole-board and corridor tests. The AMBITUS system (a square corridor with several food rewards), as a combination of these tests, was developed for the investigation of a variety of parameters associated with exploration and cognitive performance in rodents. Experiments were performed to characterize these behaviors in healthy rats and a new "schizophrenia-like" rat substrain with impaired learning ability to reveal the reliability in tests related to these functions.A square corridor was constructed with equally spaced sites along each wall (4 inside and 4 outside) resulting in 16 side-boxes for food rewards. Photocells at each box recorded the visits into the side-boxes (as exploratory activity), while the eating parameters were obtained from video records. The animals were exposed to two types of tasks repeatedly in two series: all (16) or only the inside (8) boxes (Task 1 or Task 2, respectively) were baited.Most of the rats acquired Task 1, and their performance improved by repetition, but the new substrain showed decreased exploration and learning capacity. The introduction of Task 2 caused prompt preference of the baited inner side-boxes, and gradually improved working and reference memory during the trials. The manual and automated scoring of the visits into the side-boxes showed significant (r=0.97) correlation.The results proved that healthy animals could perform the simple tasks in the square corridor after a few repetitions. The semi-automated AMBITUS system might be appropriate to detect cognitive flexibility after different manipulations, and it provides immediate, online assessment of exploratory behavior of a large number of animals within a short period of time, and it reduces the possibility of experimenter bias.