T-Maze Delayed Alternation

The T-maze delayed alternation task introduces a variable delay between the sample and choice runs of a forced alternation procedure, enabling researchers to probe the temporal limits of spatial working memory. On each trial, the animal is forced to enter one goal arm, then returned to the start position and held for a defined delay (typically 10-120 seconds) before being given a free choice. By testing multiple delays within or across sessions, the protocol generates a delay-dependent forgetting curve that reveals the time course of spatial information decay in working memory.

The primary analysis relates correct choice percentage to delay duration, yielding a forgetting function whose slope and intercept are independently informative. Steep forgetting curves indicate rapid working memory decay, while a downward shift of the entire curve suggests a global encoding deficit. The delay at which performance drops to chance (50%) provides a single-parameter summary of working memory capacity. This protocol is exquisitely sensitive to hippocampal and prefrontal cortical damage and is widely used to characterize mnemonic impairments in rodent models of schizophrenia, Alzheimer disease, and aging.

ConductMaze manages trial-by-trial delay intervals through automated guillotine door scheduling, allowing the software to interleave different delays within a session in a pseudorandom order. The system precisely controls door open/close timing, logs the actual delay experienced by each animal, and computes delay-dependent performance curves with built-in curve-fitting routines. Real-time monitoring ensures that the animal remains in the start compartment during the delay and detects premature door approaches.