An important aspect of cognitive functioning that is often impaired following traumatic brain injury (TBI) is visuospatial learning and memory. The Austin Maze task is a measure of visuospatial learning that has a long history in both clinical neuropsychological practice and research, particularly in individuals with TBI. The aim of this study was to evaluate visuospatial learning deficits following TBI using a new computerised version of the Austin Maze task. Twenty-eight individuals with moderate-to-severe TBI were compared to 28 healthy controls on this task, together with alternative neuropsychological measures, including the WAIS-III Digit Symbol and Digit Span subtests, the Trail Making Test, WMS-III Logical Memory, and Rey Osterrieth Complex Figure Test. The results demonstrated that TBI individuals performed significantly more poorly on the Austin Maze task than control participants. The Austin Maze task also demonstrated good convergent and divergent validity with the alternative neuropsychological measures. Thus, the computerised version of the Austin Maze appears to be a sensitive measure that can detect visuospatial learning impairments in individuals with moderate-to-severe TBI. The new computerised version of the task offers much promise in that it is more accessible and easier to administer than the conventional form of the test.

The Austin Maze is a neuropsychological assessment tool used to measure cognitive function. A computerised version of the tool has recently been developed and shown to be equivalent to the conventional version in terms of performance. However, controversy remains regarding which specific cognitive constructs the conventional and computer versions of the Austin Maze purport to measure. The aim of this study was to investigate which cognitive constructs are associated with Austin Maze performance and whether these constructs remain equivalent across conventional and computer versions. Sixty-three healthy people completed both conventional and computerised versions of the Austin Maze in addition to a number of established measures of planning, error utilisation, working memory, visuospatial ability and visuospatial memory. Results from a series of regression analyses demonstrated that both versions of the Austin Maze were predominantly associated with visuospatial ability and visuospatial memory. No executive measures, including those of planning, error utilisation or working memory, significantly contributed to any Austin Maze performances. This study complements previous research and supports equivalency of the conventional and computer versions of the Austin Maze.

The Austin Maze has declined in use in both clinical and research contexts due to difficulties in accessing the conventional button-box form of the test. Computer-based versions of the Austin Maze offer a potential means of making the test more accessible, but as yet there is limited evidence regarding the equivalence of computer and conventional versions of the Austin Maze. The present study compared performance on a computer version of the Austin Maze by Bray and McDonald with performance on the traditional button maze in 63 participants aged 18–27 years. The results showed no differences between the computer and conventional versions in terms of mean scores and distributions, and performances on the two versions were significantly correlated. Examination of correlates found no relationship between Austin Maze performance and years of education or age for either version of the Austin Maze performance. Intellectual function was modestly associated with performance on the conventional version but not the computer version. Overall, these findings suggest that scores on the Bray and McDonald computer version of the Austin Maze produces comparable scores to the conventional form of the test and can be interpreted using existing normative data.

This study aimed to investigate which abilities are measured by the Austin Maze. One hundred and eight university students were administered a battery of eight neuropsychological tests including, the Austin Maze, the Tower of London, the Wisconsin Card Sort Test, Block Design, the Visual Spatial Learning Test, Digit Span Backwards, the Brown-Peterson Task and the Wide Range Achievement Test of Reading. Results indicated that visuospatial ability and memory both significantly contributed to performance on the Austin Maze, but differed in the degree to which they explained the performance depending on which measure of maze performance was employed. It appears that visuospatial ability is measured in early trials of the Austin Maze when individuals are orienting themselves to the path. In later trials individuals must call upon visuospatial memory to consolidate the details of the path. Executive function and working memory were not found to be significantly implicated in performance on the Austin Maze. (JINS, 1999, 5, 1–9.)