Skip to main content

Simple brain scan reveals early signs of Alzheimer’s disease

PREDICTOM study reveals slowdown in communication between different brain regions from the earliest stages of Alzheimer’s disease.

16 September 2025
A man with memory problems due to early stage Alzheimer’s disease makes a note.
© LightField Studios, Shutterstock

IHI project PREDICTOM aims to boost our ability to detect people at risk of developing dementia (including Alzheimer’s disease) before the first symptoms appear. Now, the project has demonstrated that a type of brain scan called an EEG (electroencephalogram) can distinguish between healthy people and people with very subtle symptoms of the disease. The findings are published in the Journal of Dementia and Alzheimer’s Disease.

The challenge: detecting the earliest signs of Alzheimer’s disease

In an EEG, electrodes capable of measuring electrical activity in the brain are attached to the person’s scalp. For this study, the PREDICTOM team turned to a public EEG database at Chung-Ang University Hospital comprising EEG recordings plus detailed clinical information from over 1 100 people, including people with Alzheimer’s disease, mild cognitive impairment (MCI), and subjective cognitive decline (SCD) as well as healthy controls. MCI is often viewed as ‘normal aging’; however, studies show that people with MCI develop Alzheimer’s disease earlier than healthy people of the same age. Meanwhile SCD lies between normal aging and MCI. Today, SCD is very hard to diagnose as those who have it may sense that their cognitive abilities are declining, but this is rarely picked up by standard tests.

However, identifying people with SCD and MCI is critical as treatments and lifestyle changes designed to slow the progress of Alzheimer’s disease are most likely to be effective during the earlier stages of the disease.

Using graph theory to track brain waves

Most previous studies of Alzheimer’s disease using EEG data focused on things like the speed of the brain rhythms. In contrast, the PREDICTOM team turned to graph theory to analyse the data.

“Graph theory is a branch of mathematics used to model complex networks – like the internet, social networks or, in our case, the brain’s communication patterns and functionality among different brain regions,” explains lead author Ioulietta Lazarou of Greece’s Centre for Research and Technology Hellas. “In our analysis, brain regions are represented as ‘nodes’, and the functional connections between them (based on EEG signal synchronisation) as ‘edges’. This approach allowed us to quantify how efficiently and robustly the brain communicates.”

The analysis revealed clear differences in how electrical signals moved between brain regions in people with different levels of cognitive symptoms. People with Alzheimer’s disease showed the slowest brain activity and seriously reduced communication between the different regions of the brain. Crucially, changes in brain network activity were also clearly visible in people with SCD and MCI.

“We found that even people with very subtle memory concerns, who don’t yet show signs of cognitive impairment, already have measurable changes in how different parts of the brain communicate,” says Dr Lazarou.

PREDICTOM: targeting non-invasive, low-cost ways of detecting dementia

The discovery suggests that EEGs could help doctors spot signs of Alzheimer’s earlier than traditional tests and examinations; this is important because EEGs are low-cost and widely accessible, and could even be used in the home.

“The study directly supports PREDICTOM’s goal of identifying non-invasive, scalable biomarkers for early detection and patient stratification in Alzheimer’s disease,” says Dr Lazarou. “By demonstrating that network-level alterations in brain activity can be detected in resting-stage EEG, we move closer to the project’s vision of developing practical, explainable tools to support early diagnosis and monitoring.”

The researchers now plan to validate their results in other EEG datasets, including those from PREDICTOM cohorts. They also aim to look for links between brain network disruptions as detected by EEGs, and clinical and cognitive data. If links are found, EEG data could eventually be used as a marker to track disease progression or treatment response.

Meanwhile the interdisciplinary, public-private nature of the PREDICTOM project was key to getting the results, says Dr Lazarou: “Academic partners provided methodological expertise and data analysis, while industry partners contributed perspectives on translational relevance, clinical utility, and data integration across modalities. This collaboration ensured the study stayed both scientifically rigorous and aligned with real-world application needs.”