Can Digital Memory Tests Detect Tau-Related Cognitive Change Before Alzheimer Disease?

Digital cognitive assessments that tax hippocampal memory have shown promise as low-burden markers of early Alzheimer disease, in some analyses outperforming amyloid and tau measures for predicting cognitive decline.¹ Casey Vanderlip shared his recent analysis of screening data from the AHEAD and A4 trials, which paired digital memory assessments with amyloid and tau positron emission tomography, and found the largest deficits in amyloid-positive participants with neocortical tau, with little cognitive difference observed with amyloid positivity alone. The deficits were driven by hippocampal-dependent memory tasks, whereas attention and psychomotor speed were unaffected across biological stages. Vanderlip noted that the cross-sectional design precludes determining when these changes first emerge, that the differences were observed at the group level in highly selected trial populations, and that longitudinal studies are needed to establish whether these measures can support trial enrichment or individual-level clinical decisions. Related work in A4 participants found that combining a brief digital memory battery with plasma phosphorylated tau 217 identified cognitively unimpaired individuals at elevated risk of future cognitive decline.²

Psychiatric Times: Your research used screening data from the AHEAD and A4 trials; why were these secondary prevention cohorts the right place to test digital memory measures?

Casey Vanderlip: AHEAD and A4 are 2 of the largest preclinical Alzheimer disease trials, making them particularly valuable for understanding how cognition relates to underlying disease biology before clinical impairment emerges. Both studies collected digital cognitive assessments alongside amyloid and tau PET imaging, allowing us to examine subtle cognitive differences across biological stages of disease. I think an important strength is that we're studying these relationships in actual clinical trial populations, where understanding which cognitive measures are most sensitive to early disease changes has direct implications for trial design and outcome selection. Having 2 independent cohorts also allowed us to examine whether these relationships were consistent across different populations and study designs.

PT: Amyloid-positive participants with neocortical tau showed the largest deficits. What does that tell us about how tau spread relates to cognition at the preclinical stage?

Vanderlip: I think this suggests that as tau pathology spreads beyond the medial temporal lobe into neocortical regions, it may begin to disrupt the broader neural circuits supporting cognition, particularly memory. We saw relatively little cognitive difference with amyloid positivity alone, whereas individuals with more extensive tau pathology showed substantially greater deficits. This suggests that the regional extent of tau pathology may be particularly important in determining when the underlying biology of Alzheimer disease begins to translate into measurable cognitive impairment. What's especially interesting is that we're detecting these differences in individuals who are still considered cognitively unimpaired by conventional clinical criteria.

PT: Why were the deficits driven by hippocampal-dependent memory tasks?

Vanderlip: The hippocampus and surrounding medial temporal lobe structures are particularly vulnerable to Alzheimer disease pathology, even in its earliest stages. The memory tasks that showed the strongest effects in our study rely on hippocampal-dependent computations, such as distinguishing between similar representations, forming associations, and learning in the presence of interference. These processes may be disrupted both directly by pathology affecting the medial temporal lobe and indirectly as pathology spreads and alters the broader neural networks supporting memory. I think our findings highlight the importance of designing cognitive assessments that specifically tax the neural systems and computations most vulnerable to early Alzheimer disease, rather than relying solely on more general measures of cognitive function.

PT: In your analysis, attention and psychomotor speed were unaffected across stages. How might that shape which tasks clinicians choose to monitor?

Vanderlip: It's important not to view these findings as a suggestion that attention or processing speed should be ignored. We studied a highly selective population of cognitively unimpaired individuals from preclinical Alzheimer disease trials. In more heterogeneous clinical populations, changes in attention and processing speed may be particularly informative, especially given that other conditions and even different presentations of Alzheimer disease can affect these domains. Rather, our findings suggest that in the early stages of typical Alzheimer disease, memory may be particularly sensitive to underlying disease progression. I think this reinforces the importance of including sensitive memory assessments when monitoring early Alzheimer disease, while recognizing that a broader evaluation of cognitive function remains essential.

PT: You saw the cognitive signature grew with biological disease stage. How early in that progression does it become detectable?

Vanderlip: I think this is a really important question, but one we can't directly answer with the current study given its cross-sectional design. We observed relatively little cognitive difference with amyloid positivity alone, while individuals with medial temporal tau showed some intermediate differences and those with neocortical tau showed the clearest deficits. However, this doesn't tell us exactly when these cognitive changes first emerge or how they evolve as pathology progresses. Longitudinal studies will be crucial for understanding how cognitive changes unfold alongside accumulating pathology, and ultimately determining how early sensitive digital assessments can detect the cognitive consequences of Alzheimer disease.

PT: How might digital assessments help identify and enrich participants for early-stage trials?

Vanderlip: One of the challenges in preclinical Alzheimer disease trials is identifying individuals who are likely to experience measurable cognitive change over the course of a study. Our findings suggest that digital cognitive assessments may provide useful information alongside biological markers by capturing subtle differences in cognitive function. Because these tools are relatively brief, low-cost, and scalable, they could potentially be incorporated into trial screening to help characterize cognitive vulnerability and inform participant selection. However, our study wasn't designed to evaluate trial enrichment directly, so longitudinal studies will be important to determine whether these measures can reliably predict future decline and ultimately improve trial efficiency.

PT: What would need to happen for these tools to move into routine clinical care?

Vanderlip: I think there are several important steps before digital cognitive assessments can be widely adopted in clinical care. First, we need to establish their feasibility and reliability across more diverse, real-world populations, including whether they can be administered remotely and without supervision. We also need to understand how robust these measures are across different demographic groups and testing environments, and whether they are sufficiently sensitive to detect meaningful cognitive changes over time. Perhaps most importantly, we need to demonstrate that the information these assessments provide actually improves clinical decision-making and patient care. Ultimately, the goal isn't simply to make cognitive testing more accessible, but to ensure that these tools provide meaningful information that clinicians can use to better support their patients.

PT: What would you say are the main limitations of these findings, and what should clinicians keep in mind when interpreting them?

Vanderlip: I think there are a few important limitations to keep in mind. First, we studied highly selected populations from preclinical Alzheimer disease trials, so it's not yet clear how well these findings generalize to more diverse clinical populations. Second, the cognitive differences we observed were at the group level. While we detected differences between biological stages, we know much less about how accurately these assessments can characterize disease or predict future decline in individual patients. Finally, because this was a cross-sectional study, we can't determine how these cognitive changes unfold over time. I think it's important for clinicians to recognize that these are promising research tools, but more work is needed to establish their clinical utility, particularly for interpreting results and making decisions at the individual patient level.

PT: What do you hope the future of this research looks like?

Vanderlip: Digital cognitive assessments are receiving a lot of attention right now, and I think there's good reason to be excited about their potential. But with that excitement comes a responsibility to rigorously understand both their capabilities and their limitations. We need to establish what these tools can reliably measure, how sensitive they are to early disease-related changes, and where they provide meaningful advantages over existing approaches. I'm particularly interested in understanding how digital cognitive measures relate to underlying Alzheimer disease biology and how those relationships evolve over time. Ultimately, the goal isn't simply to develop newer or more convenient ways to assess cognition, but to use these tools to improve how we detect, monitor, and treat Alzheimer disease, and most importantly, to make a meaningful difference in patients' lives.

Mr Vanderlip is an MD/PhD student at the University of California, Irvine. He received a CTAD Rising Star Award for his research in Alzheimer disease.

References

1. Vanderlip CR, Stark CEL; Alzheimer’s Disease Neuroimaging Initiative. Digital cognitive assessments as low-burden markers for predicting future cognitive decline and tau accumulation across the Alzheimer’s spectrum. Alzheimers Dement. 2024;20(10):6881-6895.

2. Vanderlip CR, Stark CEL. Integrating plasma p-tau217 and digital cognitive assessments for early detection in Alzheimer’s disease. Alzheimers Dement. 2025;21:e70355.