Apathy and depression dominate Alzheimer’s symptom networks in early and late disease alike, study finds

Apathy and depression are the most common neuropsychiatric symptoms in Alzheimer’s disease, whether the illness begins before or after the age of 65, according to a network-based analysis of more than 1,300 patients from the Amsterdam Dementia Cohort.

Early-onset Alzheimer’s disease, defined in the study as onset before 65 years, included 655 participants with a mean age of 58.4 years and 53 percent women. The late-onset group comprised 705 participants aged 65 or older, with a mean age of 70.8 years and 52 percent women. For each group, the researchers estimated a symptom network using Ising models for binary data, a statistical framework in which each dichotomized questionnaire item, scored as present or absent, becomes a node, and the edge weight between two nodes reflects the strength of the conditional association between those symptoms after accounting for all other symptoms in the network.

This approach differs from simply counting how often symptoms co-occur. The researchers then compared the early-onset and late-onset networks by visual inspection, by the strength of individual edges, by node strength centrality, defined as the sum of edge weights connected to a node, and by a formal Network Comparison Test of global network strength, a summary measure of the overall connectivity of each network.

The headline result is one of similarity. Apathy was the most prevalent symptom in both groups, affecting 58 percent of the early-onset patients and 55 percent of the late-onset patients, with depression close behind at 48 percent and 43 percent respectively. Global network strength did not differ significantly between the two groups, with the comparison yielding a p-value greater than .05. In other words, the overall density of symptom interconnections is statistically indistinguishable between patients whose disease starts early and those whose disease starts late.

Beneath that global similarity, however, the analysis revealed subtle differences in which specific connections carried the most weight. In the early-onset network, the strongest edges linked aggression with disinhibition and aggression with irritability, each with an edge weight of 1.22. In the late-onset network, the strongest edge connected depression with anxiety, at an edge weight of 1.49. These edge weights describe the conditional association between symptom pairs within each group’s network and are not directly comparable as causal claims; they indicate which symptom pairs are most tightly bound together once all other symptoms are taken into account.

Node strength centrality told a complementary story. In the early-onset group, aggression and anxiety showed the highest node strength, at 3.61 and 3.60 respectively, meaning these symptoms had the largest summed connections to the rest of the network. In the late-onset group, aggression again ranked highest at 3.83, but the second most central node was disinhibition, at 3.54.

The authors conclude that early-onset and late-onset Alzheimer’s disease show largely overlapping neuropsychiatric symptom networks, with subtle differences in how symptoms manifest between the groups. Because the study is cross-sectional, it captures symptom co-occurrence at a single point in time and cannot establish whether one symptom causes another or how networks evolve as the disease progresses. The participants all came from a single memory clinic cohort, and the NPI-Q items were dichotomized as present or absent, which simplifies the graded severity that clinicians typically observe.

Rather than treating each symptom as an isolated item on a checklist, the Ising framework treats the symptom profile as an interconnected system, allowing researchers to quantify which connections are strongest and which symptoms act as hubs. Applied here to a large, well-characterized cohort, the method shows that the broad structure of neuropsychiatric comorbidity in Alzheimer’s disease is robust across age of onset, even as the fine-grained wiring of the symptom network varies in measurable ways.

Subject of Research: Differences in neuropsychiatric symptom co-occurrence networks between early-onset and late-onset Alzheimer's disease

Article Title: Neuropsychiatric symptom networks in early- and late-onset Alzheimer's disease

Article References: Weltings, E., van Dam, M., Butterbrod, E., Tijms, B. M., van Harten, A. C., de Boer, S. C. M., van der Flier, W. M., Pijnenburg,, Y. A. L., & Sikkes, S. A. M. (2026). Neuropsychiatric symptom networks in early- and late-onset Alzheimer's disease. Alzheimer's Research & Therapy. https://doi.org/10.1186/s13195-026-02225-y

Image Credits: AI Generated

DOI: 10.1186/s13195-026-02225-y

Keywords: Alzheimer's disease, neuropsychiatric symptoms, early-onset, late-onset, network analysis, apathy, depression, NPI-Q, Ising model, Amsterdam Dementia Cohort, Neuropsychiatric, symptom