Sensitivity comparison of longitudinal cognitive function indicators of Alzheimer’s disease after mild cognitive impairment: a prospective cohort study

Why this matters for families and doctors

Many people diagnosed with mild memory problems want to know: will this stay stable, or is it the first step toward Alzheimer’s disease? This study followed hundreds of adults with mild cognitive impairment over several years to see which common memory and daily-function tests best forecast who will later develop Alzheimer’s. By tracking how scores change over time instead of relying on a single snapshot, the researchers built a more realistic picture of brain aging and found a simple combination of tests and a genetic marker that can help doctors sort patients into higher- and lower-risk groups.

Following memory changes over the years

The team drew on data from the Alzheimer’s Disease Neuroimaging Initiative, a large North American project that has been monitoring volunteers since 2006. They focused on 596 people who all started with mild cognitive impairment and had at least three repeat visits. Over a median of about two and a half years, 184 of these individuals went on to develop Alzheimer’s dementia, while 412 stayed in the mild stage. At each visit, participants completed a panel of well-known tests, including general thinking tests, memory tasks, and questionnaires about everyday abilities such as managing bills or preparing meals.

Which tests send the clearest warning signals?

Not all tests turned out to be equally useful for predicting who would decline. Worsening scores on several measures of overall thinking and daily function strongly signaled higher risk of Alzheimer’s. In particular, two tools stood out: the Clinical Dementia Rating Sum of Boxes (CDR-SB), which rates how well someone manages basic and complex daily activities, and the Functional Activities Questionnaire (FAQ), which focuses on more detailed day-to-day tasks. When these scores rose over time, the chance of progressing to Alzheimer’s increased sharply. In contrast, better performance on memory tests (such as recalling word lists) and on a brief mental status exam was linked to a lower likelihood of conversion.

Looking beyond straight-line decline

Most past studies have treated changes on cognitive tests as if they followed a simple straight-line path. The authors of this paper suspected that real-life decline is more uneven, with certain stages where abilities drop faster. They used a flexible statistical approach that allows test scores to bend and curve over time and ties these changing paths directly to the chances of developing Alzheimer’s. This method captured clear non-linear patterns, especially for CDR-SB and FAQ, and yielded models that fit the data well and produced reasonably accurate risk predictions over two, five, and even eight years. In other words, how a person’s scores move over time—rather than a single score at one visit—carries important clues about their future.

The role of genes in speeding up decline

The researchers also examined a well-known genetic risk factor: the APOE ε4 variant. People can carry zero, one, or two copies of this variant. Across all the models, having APOE ε4 consistently raised the chance of progressing from mild impairment to Alzheimer’s. When they separated participants by how many copies they carried, a dose–response pattern emerged. Those with two copies tended to decline faster and convert earlier, especially on the CDR-SB and FAQ measures, while those without the variant declined more slowly. This suggests that the same test scores may mean different things for different people depending on their genetic background.

What this means for patients and clinics

For patients and families, the study’s message is that repeated check-ups focusing on everyday function can offer more than reassurance or worry in the moment—they can provide a window into the future. For clinicians, combining APOE ε4 genotyping with careful, long-term tracking of CDR-SB and FAQ scores appears to be a practical way to flag individuals at highest risk of developing Alzheimer’s dementia. While the models are not perfect and still need to be tested in broader, more diverse groups, they point toward a future in which doctors rely on patterns over time, rather than one-off scores, to guide monitoring, counseling, and early intervention in people living with mild cognitive impairment.

Citation: Guo, G., Song, W., Wang, A. et al. Sensitivity comparison of longitudinal cognitive function indicators of Alzheimer’s disease after mild cognitive impairment: a prospective cohort study. Sci Rep 16, 14503 (2026). https://doi.org/10.1038/s41598-026-44192-2

Keywords: Alzheimer’s disease, mild cognitive impairment, cognitive decline, APOE genetic risk, functional assessment