RESUMO
OBJECTIVE: Standard evaluation of the Trail Making Test (TMT) only incorporates completion times. However, the analysis of different error types may provide more insight into underlying cognitive processes and could also increase diagnostic accuracy. This cross-sectional observational study compared three different TMT error types and assessed their diagnostic utility in patients with subjective cognitive decline (SCD), mild cognitive impairment (MCI), and Alzheimer's dementia (AD) with or without depression. METHOD: We evaluated 618 outpatients of a memory clinic with SCD (N = 190), MCI (N = 210), or AD (N = 218). Of these, 157 had comorbid depression. TMT completion times, total error rates, and the three error types "sequencing error," "perseverative error," and "proximity error" were examined. RESULTS: Results indicated that patients with MCI or AD committed more errors on TMT B, and specifically more perseverative errors than patients with SCD (p < 0.001). Depression was not associated with any TMT error type. Including TMT errors in models predicting diagnosis group by TMT completion times did not increase predictive accuracy, measured by areas under the curve. CONCLUSIONS: The findings do not indicate any impact of comorbid depression on TMT errors. Moreover, TMT error analysis does not seem to provide additional diagnostic utility for SCD, MCI, and AD diagnoses.