Interindividual and intraindividual variability in amnestic mild cognitive impairment (aMCI) measured with an online cognitive assessment.

INTRODUCTION: Mean cognitive performance is worse in amnestic mild cognitive impairment (aMCI) compared to control groups. However, studies on variability of cognitive performance in aMCI have yielded inconclusive results, with many differences in variability measures and samples from one study to another. METHODS: We examined variability in aMCI using an existing older adult sample (n = 91; 51 with aMCI, 40 with normal cognition for age), measured with an online self-administered computerized cognitive assessment (Cogniciti's Brain Health Assessment). Our methodology extended past findings by using pure measures of variability (controlling for confounding effects of group performance or practice), and a clinically representative aMCI sample (reflecting the continuum of cognitive performance between normal cognition and aMCI). RESULTS: Between-group t-tests showed significantly greater between-person variability (interindividual variability or diversity) in overall cognitive performance in aMCI than controls, although the effect size was with a small to moderate effect size, d = 0.44. No significant group differences were found in within-person variability (intraindividual variability) across cognitive tasks (dispersion) or across trials of a response time task (inconsistency), which may be because we used a sample measuring the continuum of cognitive performance. Exploratory correlation analyses showed that a worse overall score was associated with greater inter- and intraindividual variability, and that variability measures were correlated with each other, indicating people with worse cognitive performance were more variable. DISCUSSION: The current study demonstrates that self-administered online tests can be used to remotely assess different types of variability in people at risk of Alzheimer`s. Our findings show small but significantly more interindividual differences in people with aMCI. This diversity is considered as "noise" in standard assessments of mean performance, but offers an interesting and cognitively informative "signal" in itself.

Further development of the Multiple Errands Test: standardized scoring, reliability, and ecological validity for the Baycrest version.

OBJECTIVES: (1) To determine the summary scores on the Baycrest Multiple Errands Test (BMET) that best discriminate between community dwelling people with traumatic brain injury or stroke and matched controls; (2) to determine interrater reliability; (3) to evaluate further the ecological validity. DESIGN: Case-control. SETTING: Large, university-affiliated health care center and participants' homes. PARTICIPANTS: People with acquired brain injury (n=27) and healthy matched controls (n=25). INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: (1) BMET; (2) performance-based measure of instrumental activities of daily living: the Assessment of Motor and Process Skills; (3) self-report and significant other report of daily life function, the Dysexecutive Questionnaire, the Sickness Impact Profile, and the Mayo-Portland Adaptability Inventory. RESULTS: Performance on the BMET was significantly different between people with acquired brain injury and controls (P<.05); good to strong correlations (>.50) were found in more than one third of the correlations between the BMET and measures of IADL and everyday function. Interclass correlation coefficients (ICCs) on BMET summary scores were high (ICC=.71-.88), illustrating very good interrater reliability. CONCLUSIONS: This study extends the psychometric findings of the Multiple Errands Test, thus further confirming its value for clinical and research purposes. It is a reliable and ecologically valid assessment that provides a standard way of categorizing executive performance errors in a naturalistic environment.