A team of engineers from Florida Atlantic University has developed a groundbreaking AI model that utilizes electroencephalography (EEG) to distinguish between two major types of dementia: Alzheimer’s Disease (AD) and Frontotemporal Dementia (FTD). This advancement could significantly improve diagnostic accuracy and patient care for millions affected by these debilitating disorders.
Dementia encompasses a range of disorders that gradually impair memory, cognitive abilities, and daily functioning. According to estimates, approximately 7.2 million Americans aged 65 and older will be affected by AD by 2025. FTD, although less common, ranks as the second leading cause of early-onset dementia, typically impacting individuals between their 40s and 60s. Both conditions present overlapping symptoms, which often complicates accurate diagnosis.
New Insights into Brain Activity
Current diagnostic methods, such as MRI and PET scans, are effective but expensive and require specialized equipment. EEG provides a non-invasive, portable, and cost-effective alternative by measuring brain activity through sensors that capture various frequency bands. However, the analysis of EEG signals has been challenging due to noise and individual variability, leading to inconsistent results in distinguishing AD from FTD.
The research team utilized deep learning techniques to create a model that enhances the accuracy and interpretability of EEG data. Their findings, published in the journal Biomedical Signal Processing and Control, indicated that slow delta brain waves serve as a critical biomarker for both AD and FTD, particularly in the frontal and central regions of the brain. The model demonstrated over 90% accuracy in differentiating between dementia patients and cognitively normal individuals, while also predicting disease severity with relative errors of less than 35% for AD and 15.5% for FTD.
The research revealed that brain activity in AD is more widespread and affects multiple brain regions and frequency bands, such as beta, signifying more extensive damage. In contrast, FTD primarily impacts the frontal and central areas. This distinction contributes to the understanding that AD is generally easier to detect than FTD.
Advancements in Clinical Diagnosis
The model features a two-stage design that first identifies healthy individuals before differentiating between AD and FTD. This approach improved the model’s specificity in identifying healthy participants from 26% to 65%, achieving an overall accuracy of 84%. By merging convolutional neural networks and attention-based long short-term memory (LSTM) networks, the model not only identifies the type of dementia but also assesses its severity.
According to Tuan Vo, a doctoral student and first author of the study, “What makes our study novel is how we used deep learning to extract both spatial and temporal information from EEG signals. By doing this, we can detect subtle brainwave patterns linked to Alzheimer’s and frontotemporal dementia that would otherwise go unnoticed.” The model also employs Grad-CAM technology to visualize which brain signals influenced its decisions, providing clinicians with valuable insights into the diagnostic process.
The research aligns with previous studies indicating that AD tends to affect a broader range of brain areas, resulting in lower cognitive scores compared to FTD, which has more localized impacts. Hanqi Zhuang, co-author and associate dean at FAU, stated, “This difference explains why Alzheimer’s is often easier to detect. However, our work also shows that careful feature selection can significantly improve how well we distinguish FTD from Alzheimer’s.”
The implications of this study are profound. The integration of engineering, artificial intelligence, and neuroscience could revolutionize dementia diagnosis, offering clinicians efficient tools for real-time tracking of disease progression. Stella Batalama, dean of the College of Engineering and Computer Science, emphasized the potential of this research: “With millions affected by Alzheimer’s and frontotemporal dementia, breakthroughs like this open the door to earlier detection, more personalized care, and interventions that can truly improve lives.”
The findings from this research pave the way for a more nuanced understanding of dementia and highlight the critical role that advanced technology can play in enhancing healthcare outcomes.
