Computer detection of cognitive impairment and associated neuropsychiatric dimensions from the content analysis of verbal samples.

This is a report of a study testing the capacity of a computerized measure of the content analysis of five minute verbal samples to detect and measure cognitive impairment and comorbid neuropsychiatric dimensions in 117 drug-abusing inpatients. The cognitive impairment scores obtained from the computerized procedure correlated significantly with independent scores from the Trails B and Stroop Color and Word test as well as with ANAM (Automated Neuropsychiatric Assessment Metric) neuropsychological tests, including the Matching to Sample Efficiency and Accuracy, the Code Substitution Efficiency, the Continuous Performance Task Efficiency and Accuracy, the Code Substitution Delayed Recall Accuracy, and the Simple Reaction Time Efficiency. When the computerized verbal-content-analysis-derived cognitive impairment scores were combined with scores of selected other ANAM measures, more and higher intercorrelations occurred with Trails A, Trails B, the Stroop Color and Word test, and the Wisconsin Card Sort test. In addition, validated measures of a broad range of associated neuropsychiatric dimensions can be obtained simultaneously from the same five minute verbal samples providing the cognitive impairment scores. No significant effects were found on the cognitive impairment scores of age, education, gender, race, and duration of drug-abuse abstinence.