Researchers at Örebro University have developed two new AI models that can analyze the brain’s electrical activity and accurately distinguish healthy individuals from patients with dementia, including Alzheimer’s disease.
Early diagnosis is crucial so that proactive measures can be taken that slow the progression of the disease and improve the patient’s quality of life.
Muhammad Hanif, computer science researcher, Örebro University
In the study An explainable and efficient deep learning framework for EEG-based diagnosis of Alzheimer’s disease and frontotemporal dementia, the researchers combined two advanced AI methods: temporal convolutional networks and LSTM networks. The program analyzes EEG signals and can almost perfectly determine whether a person is sick or healthy.
Can distinguish healthy from sick people with 80% certainty
Comparing three groups – Alzheimer’s disease, frontotemporal dementia and healthy people – the method achieved an accuracy of more than 80 percent. The researchers also use an explanatory AI technique that shows which parts of the EEG signal affect the diagnosis. This helps doctors interpret how the system reaches its conclusions.
In the second study, Privacy-Preserving Dementia Classification from EEG via EEGNetv4 Hybrid Fusion and Federated LearningThe researchers developed a small, resource-efficient AI model – less than a megabyte – that also protects patient privacy. With federated learning, multiple healthcare providers can collaborate to train the AI system without sharing patient data. Despite the privacy protection, the model achieves an accuracy of over 97%.
“Traditional machine learning models often lack transparency and face privacy issues. Our study aims to address both of these issues,” says Muhammad Hanif, associate lecturer in computer science at Örebro University.
AI detects patterns in the brain’s electrical signals
Researchers have succeeded in combining different methods of interpreting electrical signals from the brain. By dividing EEG signals into different frequency bands – alpha, beta and gamma waves – AI can identify patterns linked to dementia. Algorithms can detect long-term changes in signals and recognize subtle differences between diagnoses. Additionally, explainable AI technology ensures that the system is no longer a “black box”: it clearly shows the basis for its decisions.
In their studies, the researchers demonstrate how AI can become a fast, inexpensive and privacy-friendly tool for the early diagnosis of dementia. EEG is already a simple and inexpensive method that can be used in primary care. Combined with AI models that can run on wearable devices, this opens up the possibility of wider use in healthcare – from specialist clinics to future at-home tests.
AI test could be used at home in future
“Early diagnosis is essential to implement proactive measures that slow disease progression and improve quality of life. If solutions like this are fully implemented, they could ease the burden on everyone involved – patients, caregivers, relatives and healthcare professionals,” says Muhammad Hanif.
The studies were carried out in collaboration between researchers at Örebro University and several international institutions, including universities in the United Kingdom, Australia, Pakistan and Saudi Arabia.
“We plan to continue the research by expanding to larger and more diverse datasets, exploring more EEG features, and including other types of dementia such as vascular dementia and dementia with Lewy bodies. At the same time, we will use explainable AI and ensure strict protection of patient data,” explains Muhammad Hanif.
