The Austin (or Milner) Maze consists of a 10 × 10 array of electrical contacts through which the subject must discover a hidden pathway by touching successive points in the matrix. Trials needed to reach 3 consecutive errorless performances, errors committed, and time taken to reach this criterion are commonly regarded as indicators of frontal lobe function. This study evaluated the equivalence of a computer-based version of this task. Thirty-two male computer science students completed both forms of the maze in counterbalanced order, separated by an interval of 4 weeks. In a 2-year follow-up, 25 remaining subjects completed the tasks in reverse order, again separated by a 4-week interval. Correlational data and analysis of variance supported the proposition that the computer-based version could act as a substitute for the more tediously administered traditional form. Although the sample characteristics may limit the generalizability of this conclusion, the application and development of the software by independent investigators may facilitate its clinical application and allow a more efficient resolution of relevant theoretical issues.