The Cambridge Neuropsychological Test Automated Battery (CANTAB) is a system of neuropsychological tests frequently used to track the progression of cognitive deficits in mild cognitive impairment (MCI) and Alzheimer's disease (AD). We investigated test-retest reliability in seven CANTAB tests. Twenty-five MCI patients, with either AD-like or conflicting/normal cerebrospinal fluid profiles underwent three testing sessions at 6-month intervals, including the following tests: Reaction Time and Rapid Visual Information Processing (assessing attention and reaction times); Delayed Matching-to-Sample, Paired Associates Learning, Spatial Recognition Memory and Pattern Recognition Memory (assessing memory); Spatial Working Memory (assessing executive functions). No significant difference was found when comparing the two groups. Many CANTAB measures obtained low or marginal test-retest coefficients. We observed a marked improvement in Spatial Working Memory (SWM) in both groups when comparing the baseline performance with the 6-month follow-up, but no difference in performance between 6- and 12-month follow-ups. A similar trend was documented in Paired Associates Learning (PAL), but the effect size was small. Such improvement may result from a practice effect, likely due to the learning of an effective strategy. Our evidence raised an important issue concerning the need for methodological caution when interpreting the results of longitudinal studies using SWM and PAL.
Keywords: AD biomarkers; CANTAB; mild cognitive impairment; practice effect; spatial working memory.