T-Maze Delayed Non-Matching to Place

The delayed non-matching to place (DNMTP) task in the T-maze is widely considered the gold-standard hippocampal-dependent working memory test for rodents. The protocol follows a sample-delay-choice structure in which the animal is forced to visit one arm, held for a defined delay, and then must choose the opposite arm to receive a food reward. Unlike simple alternation, DNMTP explicitly rewards the non-matching response and imposes parametric delays that can be titrated to map the temporal limits of spatial working memory. This protocol is directly analogous to the delayed non-matching to sample tasks used in primate neuropsychology.

The signature feature of DNMTP is its exquisite sensitivity to hippocampal damage: bilateral hippocampal lesions produce delay-dependent deficits that spare short-delay performance while devastating longer delays. Prefrontal cortex lesions produce a different pattern, with steeper forgetting slopes but relatively preserved short-delay accuracy. These dissociable lesion profiles make DNMTP a powerful tool for localizing circuit-level dysfunction. Performance is measured as percent correct at each delay, choice latency, and the parameters of the fitted delay-performance function.

ConductMaze implements DNMTP with fully automated trial management, including programmable door sequences, delay timing, reward dispensing, and pseudorandom arm assignment. The software ensures that the non-matching rule is consistently enforced across trials and sessions, generates balanced trial sequences, and computes delay-dependent forgetting functions with integrated curve fitting. Optional correction trials can be enabled for training sessions, wherein incorrect choices result in a repeated trial with the same sample arm.