Behavioral Changes Predict Alzheimer’s Disease Risk

In addition to its well-known effects on cognition, especially memory, Alzheimer’s disease has also been associated with psychiatric and behavioral problems, and emerging evidence suggests that these changes could help identify patients at risk of developing dementia. A recent study published in Brain used NACC dataset to validate that approach, showing that people meeting specific neuropsychiatric criteria were significantly more likely to develop autopsy-confirmed Alzheimer’s disease within five years. The results point toward a scalable diagnostic strategy that could help identify patients who may qualify for clinical trials and targeted treatments. 

Linking Alzheimer’s disease to psychiatric symptoms isn’t a new idea. “The index case, the very first person described as having the disease that was named after [Alois] Alzheimer was a woman who came to the hospital not with cognitive symptoms, but with emotional dysregulation,” says Zahinoor Ismail, MD, FRCPC, chair of the Canadian Conference on Dementia and a professor of psychiatry at the University of Calgary. Ismail, the senior author on the new paper, has long studied the link between dementia and mental health. 

It’s a tricky problem. “If an older person has behavioral changes, is it a [pre-existing] psychiatric condition, neuropsychiatric symptoms that are presenting in response to life stressors, underlying infections … other medical conditions or medications, or are these symptoms reflections of the underlying amyloid and tauopathies that we link with Alzheimer’s disease?” says Ismail. To address those nuances, he and his colleagues have developed a framework called mild behavioral impairment (MBI), which describes the new onset of persistent neuropsychiatric symptoms that are not better explained by other mechanisms. Like mild cognitive impairment (MCI), MBI often precedes the onset of more typical symptoms of dementia. 

After more than a decade of refining and testing those criteria, the researchers were ready to apply them to a large population of patients. Previous results had shown strong links between MBI and later cognitive decline, as well as changes in biomarkers associated with Alzheimer’s disease, but those studies were relatively small and relied on indirect measures of neuropathology. “We wanted to see, in brain tissue, was there greater Alzheimer’s disease pathology in those who had MBI, [but] of course this is very hard because you have to have autopsies,” says Ismail. 

Turning to NACC, he and an international team of collaborators conducted a case-control study of more than a thousand NACC participants whose brains had been assessed at autopsy, yielding definitive diagnoses. They used results from neuropsychiatric questionnaires to identify people who met the criteria for MBI. “We had to use multiple visits, because the [neuropsychiatric assessment] measures only a four-week reference range, and the persistence criterion of MBI is symptoms persisting for at least six months,” says Ismail. The investigators then looked for participants who’d died and received autopsies within five years, and grouped them into cases who’d previously developed MBI, and controls who hadn’t. 

Thanks to the immense size of NACC dataset, over a thousand people had completed multiple neuropsychiatric assessments and been autopsied within five years. About half had developed autopsy-confirmed Alzheimer’s disease, providing large samples of both cases and controls for analysis. The results were clear. “Those with MBI were much more likely to have [Alzheimer’s disease] pathology in their brain than those without MBI,” says Ismail. 

In addition to Alzheimer’s disease, the researchers also looked at two other forms of dementia: Lewy body disease and a distinctive neuropathology associated with a protein biomarker called TDP-43. “We had hypothesized that we would also see more Lewy bodies in the MBI group,” says Ismail, adding that Lewy body dementia, in particular, is associated with hallucinations and other behavioral pathologies. However, the data showed no link between MBI and either Lewy body dementia or TDP-43. “It wasn’t the home run for Lewy body disease that some might have hoped for, but it does tell us [that] behavior is core and fundamental to the Alzheimer’s disease process,” says Ismail. 

Ismail’s team is already working on some of those studies, including efforts to identify the mechanisms underlying MBI development, and potential strategies for treating Alzheimer’s-associated neuropsychiatric disease. “This is really paradigm-shifting, because there’s been this kind of false dichotomy that you treat cognitive symptoms with Alzheimer’s medication, and you treat neuropsychiatric or behavioral symptoms with psychiatric medication, but I would contest that,” says Ismail. For example, data from trials of Aducanumab, an anti-beta-amyloid antibody therapy, show that the drug can yield significant improvements in a patient’s behavioral symptoms and mood as well as cognition. Ismail hopes that the adoption of MBI will help identify high-risk patients earlier and improve the treatment of their various symptoms. 

Sharif SF, Guan DX, Bodnar TS, Joseph JT, Zetterberg H, Smith EE, Ismail Z. Neuropsychiatric symptoms and progression to pathologically confirmed Alzheimer's disease. Brain. 2025 Oct 3;148(10):3694-3704. doi: 10.1093/brain/awaf156. PMID: 40279515; PMCID: PMC12493056.