Touchscreen Paired Associates Learning

The touchscreen paired associates learning (PAL) task requires rodents to learn specific object-location associations and select the correct spatial location for each visual stimulus presented on a touchscreen display. On each trial, two or three stimuli are shown simultaneously, each in a different screen location, but only one stimulus is in its correct paired location. The task is a direct rodent analogue of the CANTAB Paired Associates Learning test used clinically to detect early hippocampal dysfunction in Alzheimer's disease. Performance depends on hippocampal pattern separation and completion, entorhinal-hippocampal spatial processing, and perirhinal object-identity circuits.

The primary dependent variable is the percentage of correct responses, which must exceed chance level (33% for three-choice or 50% for two-choice variants) to demonstrate learning. Errors are classified as stimulus-driven (selecting the wrong location for a given stimulus) or location-driven (responding to a preferred location regardless of stimulus), providing insight into whether deficits reflect associative binding or spatial bias. Correction trials, in which incorrect trials are repeated until a correct response occurs, yield perseverative error counts. Additional metrics include trials to criterion, response latency, reward collection latency, and session completion rate.

ConductMaze manages PAL protocols by rendering stimulus-location configurations on calibrated touchscreen panels and detecting touch responses with spatial precision. The system counterbalances stimulus-location pairings across sessions, applies configurable correction trial logic, and classifies errors by type in real time. Learning curves, error distributions, and latency profiles are exported alongside raw trial-by-trial data. The platform supports flexible stimulus sets and location arrays for parametric difficulty manipulation.