Monthly At-Home Computerized Cognitive Testing to Detect Diminished Practice Effects in Preclinical Alzheimer's Disease

Abstract

Introduction: We investigated whether monthly assessments of a computerized cognitive composite (C3) could aid in the detection of differences in practice effects (PE) in clinically unimpaired (CU) older adults, and whether diminished PE were associated with Alzheimer's disease (AD) biomarkers and annual cognitive decline.Materials and Methods:N = 114 CU participants (age 77.6 ± 5.0, 61% female, MMSE 29 ± 1.2) from the Harvard Aging Brain Study completed the self-administered C3 monthly, at-home, on an iPad for one year. At baseline, participants underwent in-clinic Preclinical Alzheimer's Cognitive Composite-5 (PACC5) testing, and a subsample (n = 72, age = 77.8 ± 4.9, 59% female, MMSE 29 ± 1.3) had 1-year follow-up in-clinic PACC5 testing available. Participants had undergone PIB-PET imaging (0.99 ± 1.6 years before at-home baseline) and Flortaucipir PET imaging (n = 105, 0.62 ± 1.1 years before at-home baseline). Linear mixed models were used to investigate change over months on the C3 adjusting for age, sex, and years of education, and to extract individual covariate-adjusted slopes over the first 3 months. We investigated the association of 3-month C3 slopes with global amyloid burden and tau deposition in eight predefined regions of interest, and conducted Receiver Operating Characteristic analyses to examine how accurately 3-month C3 slopes could identify individuals that showed >0.10 SD annual decline on the PACC-5.Results: Overall, individuals improved on all C3 measures over 12 months (β = 0.23, 95% CI [0.21–0.25], p < 0.001), but improvement over the first 3 months was greatest (β = 0.68, 95% CI [0.59–0.77], p < 0.001), suggesting stronger PE over initial repeated exposures. However, lower PE over 3 months were associated with more global amyloid burden (r = −0.20, 95% CI [−0.38 – −0.01], p = 0.049) and tau deposition in the entorhinal cortex (r = −0.38, 95% CI [−0.54 – −0.19], p < 0.001) and inferior-temporal lobe (r = −0.23, 95% CI [−0.41 – −0.02], p = 0.03). 3-month C3 slopes exhibited good discriminative ability to identify PACC-5 decliners (AUC 0.91, 95% CI [0.84–0.98]), which was better than baseline C3 (p < 0.001) and baseline PACC-5 scores (p = 0.02).Conclusion: While PE are commonly observed among CU adults, diminished PE over monthly cognitive testing are associated with greater AD biomarker burden and cognitive decline. Our findings imply that unsupervised computerized testing using monthly retest paradigms can provide rapid detection of diminished PE indicative of future cognitive decline in preclinical AD.

Practice Effect of Repeated Cognitive Tests Among Older Adults: Associations With Brain Amyloid Pathology and Other Influencing Factors

“…In our study, the attenuated PE in memory indices was associated with the presence of high amyloid burden but not with brain MRI features, including hippocampal volume, implying that PE in memory tests could be more indicative of β-amyloidosis [which is specific for Alzheimer’s disease (AD)] instead of biomarkers of neurodegeneration or neuronal injury ( Jack et al, 2016 ). Consistent with our results, a recent report from the Harvard Aging Brain Study, of 114 cognitively unimpaired older adults, showed that lower PE in a self-administered computerised cognitive composite battery over the first 3 months was associated with more global amyloid burden (based on PiB-PET imaging) and tau deposition in the entorhinal cortex and inferior-temporal lobe (based on Flortaucipir PET imaging) ( Jutten et al, 2021 ). These findings imply the usefulness of PE as an early detection tool for signs of disease burden prior to the emergence of cognitive impairment, which might inform participant stratification and biomarker testing strategies for clinical trials.…”

“…This observation underscores the potential value of diminished practice effects as an adjunct metric to traditional assessments for the sensitive detection of preclinical AD. Several previous studies have consistently shown that diminished PE over repeated cognitive testing (mainly episodic memory measures) was associated with subsequent cognitive decline and increased risk of mild cognitive impairment (MCI) or dementia ( Duff et al, 2007 ; Sanchez-Benavides et al, 2016 ; Jutten et al, 2020 , 2021 ). In contrast, previous evidence on the association between PE and AD biomarkers and neuropathology remained inconsistent ( Duff et al, 2018 ; Ihara et al, 2018 ; Jutten et al, 2020 ).…”

“…Exploring factors that influence practice effects can be informative of potential heterogeneity of measurement bias and in developing mitigation strategies to minimise such bias ( Calamia et al, 2012 ). On the other hand, the magnitude of practice effect per se may also have indicative value for cognitive impairment or existing brain pathologies ( Duff et al, 2007 ; Jutten et al, 2021 ). From this perspective, PE may represent not merely a source of measurement error but potentially valuable information from a clinical and scientific perspective ( Duff et al, 2007 ).…”

Pepinemab antibody blockade of SEMA4D in early Huntington’s disease: a randomized, placebo-controlled, phase 2 trial

“…While predefined primary endpoints of cognition, OTS and PTAP, failed to achieve statistical significance (one-sided P = 0.028 and 0.06, respectively) other indications of cognitive benefit of pepinemab treatment include a nominally significant effect on the HD-CAB index (one-sided P = 0.007), a prespecified exploratory endpoint, as well as post hoc analysis demonstrating a nominally significant treatment effect on apathy severity (P = 0.017), a component of the prespecified exploratory PBA-s that has been reported to correlate with cognitive decline. It was also of particular interest that learning effects appear to be compromised early in HD progression, as previously reported for early Alzheimer's disease 44,45 , and that pepinemab treatment appears to improve or rescue this cognitive feature. It would be desirable to add a direct assessment of ability to learn in future studies.…”

“…An important consideration for evaluation of cognitive performance is the role of ‘learning effects’ during early evaluation sessions. It has previously been reported for Alzheimer’s disease that reduced learning effects are associated with increased levels of biomarker burden and greater cognitive decline 44 , 45 . A comparison of change in HD–CAB index for EM cohort B1 and LP cohort B2 subjects treated with placebo only (Fig.…”