Automated Computerized Neuropsychological Diagnostics in Multiple Sclerosis

Abstract

Abstract: Cognitive deficits are common in multiple sclerosis (MS), necessitating diagnosis and constant monitoring. However, frequent neuropsychological assessments are costly and not always feasible. A computerized assessment of basic cognitive functions could contribute to better patient care in this regard. The primary cognitive inventory of the Memory and Attention Test (MAT-COBI) constitutes such a tool. In this feasibility study, we tested N = 44 patients with MS with a standardized neuropsychological assessment consisting of established screening procedures (Brief International Cognitive Assessment in MS [BICAMS] and Brief Repeatable Battery [BRB] and three subtests of the Test Battery of Attentional Performance (TAP). The results of these procedures were considered a gold standard for assessing which patients were cognitively impaired or cognitively unaffected. We also administered the MAT-COBI and compared the results to the gold standard. While the results of an exploratory correlation analysis pointed toward the interpretation that the the MAT-COBI adequately operationalized the basic cognitive domains of attention, memory, and executive, sensitivity was insufficient (32.3 %), while specificity reached 92.3 %. A closer look at the single cognitive domains showed marginal differences across domains (sensitivity ranged from 0–24 %, specificity: 89.5 %-100 %). Based on these results, the MAT-COBI does not appear to be a suitable tool for differentiating between cognitively preserved and cognitively impaired patients with MS. We discuss possible explanations and implications for automated cognitive diagnostics.