Visual X Maze

The Visual X Maze is a four-arm cross-shaped maze developed in collaboration with Cedars-Sinai Medical Center and described by Vit et al. in Scientific Reports. Each arm of the X maze terminates in a visual stimulus display — either a backlit panel or small monitor — that presents distinct color, contrast, or pattern cues. The task requires the animal to navigate from a central start zone to the arm displaying a designated target stimulus, testing visual-spatial discrimination, color vision, and contrast sensitivity. Unlike traditional maze tasks that rely on distal spatial cues, the Visual X Maze directly challenges the animal's ability to discriminate between proximal visual stimuli, making it particularly sensitive to retinal pathology, optic nerve damage, and early-stage visual cortex dysfunction.

In the standard protocol, the animal is placed in the center of the X maze facing a neutral wall. Four arms extend outward, each displaying a different visual stimulus — for example, vertical green gratings (target S+), horizontal red gratings, blue checkerboard, and gray uniform background. The target stimulus arm contains a food reward or escape platform, while entry into non-target arms results in a brief confinement penalty. Across trials, the position of the target stimulus is pseudorandomly rotated among the four arms to prevent spatial bias, forcing the animal to rely on visual cue identification rather than allocentric spatial memory. The task has been extensively validated in Alzheimer's disease transgenic mouse models (5xFAD, APP/PS1), where it detects subtle visual discrimination deficits months before conventional memory impairments emerge.

ConductMaze provides complete control over the Visual X Maze through per-arm stimulus display management, automated arm entry detection via infrared beam-break sensors at each arm entrance, motorized guillotine doors for controlled access, and food well delivery in the target arm. The software pseudorandomizes target arm position across trials, enforces inter-trial intervals, and computes trial-by-trial accuracy, error patterns, and learning curves. Integration with overhead video tracking enables simultaneous analysis of path efficiency and decision-point hesitation behavior.