Comparing Clinical Diagnosis of Domains of Impairment in individuals with Mild Cognitive Impairment with the NIH Toolbox‐Cognitive Battery Performance

Abstract

AbstractBackgroundMild Cognitive Impairment (MCI) represents the phase between age‐associated cognitive decline and Alzheimer’s disease and related dementias. Examination of these individuals may be especially important through computer‐based batteries such as the NIH Toolbox‐Cognition Battery (NIHTB‐CB). To inform NIHTB‐CB validity, we compared MCI patients’ performance against categories of cognitive impairment defined by the consensus‐based clinical diagnosis form of the National Alzheimer’s Disease Coordinating Center (NACC) Unified Data Set UDS3. We expected a correspondence between MCI consensus‐based clinical deficit areas and specific NIHTB‐CB task performance.MethodRetrospective data analysis was conducted on 65 MCI participants from the Michigan Alzheimer’s Disease Research Center, with a mean age of 74±8.3, 65% female, who completed UDS3 and NIHTB‐CB measures. The UDS3 forms consist of interviews, medical/neurological examinations, and neuropsychological testing, with a clinical consensus‐based diagnosis form listing domains of cognitive impairment representing memory, language, attention, executive, and visuospatial functioning. Four domain impairment groupings with a minimum of 10 participants each were established (Executive Functioning, n = 10; Memory, n = 27; Memory+Executive Functioning, n = 17; and Memory+Language n = 11) and compared to NIHTH‐CB performance. ANOVAs adjusted for demographics and post hoc tests were conducted to differentiate across clinical consensus categories for the MCI participants.ResultTwo NIHTB‐CB measures yielded significant differences across the clinical grouping: Picture Sequence Memory (PSM, F = 3.3, p = 0.03) and Pattern Comparison (PC, F = 4.1, p = 0.01). For PSM, post hoc tests demonstrated higher scores for the Executive compared to Memory+Executive (PSM: p<0.02) groups with a trend for Executive versus Memory groups (p<.07). For PC, the Memory group performed better than both the Memory+Executive (p<.02) and Memory+Language (p<.04).ConclusionThe better memory performance on PSM of the MCI Executive group compared to the two groups with memory impairment is consistent with expectation. PC, as a measure of executive ability and processing speed was sensitive to clinically diagnosed executive impairment. Lower scores for the Memory+Language group on PC were unexpected. Efficient language functioning may be necessary for complex pattern comparison as dimensions shift. Findings support the validity of NIHTB‐CB. They also suggest the need for additional research on language demands for executive measures.