Machine learning approach for early onset dementia neurobiomarker using EEG network topology features

Area of Science:

• Neuroscience
• Artificial Intelligence
• Gerontology

Background:

• Digital health technologies and AI offer novel approaches for managing cognitive decline in older adults.
• Early detection of dementia is crucial for effective intervention and maintaining independence.
• Neurotechnology research, particularly AI for social good, aims to improve well-being for individuals with disabilities.

Purpose of the Study:

• To investigate the use of machine learning algorithms and electroencephalography (EEG) for early-onset dementia prognosis.
• To assess the potential of AI in analyzing neuro-biomarkers for cognitive decline.
• To evaluate digital non-pharmacological therapies and cognitive-behavioral interventions for dementia management.

Main Methods:

• An empirical task using an EEG-based passive brain-computer interface was developed to assess working memory decline.
• Network neuroscience techniques were applied to EEG time series data for mild cognitive impairment prediction.
• Three experimental tasks, including emotional working memory tasks and an image oddball task, were employed in a pilot study.

Main Results:

• Findings from a pilot study in Poland demonstrated the feasibility of predicting cognitive decline.
• EEG responses to emotional stimuli and visual oddball tasks were analyzed to validate the methodology.
• The study confirmed the potential of machine learning in modeling mild cognitive impairment prediction.

Conclusions:

• The proposed three experimental tasks utilizing AI show critical potential for early-onset dementia prognosis in older adults.
• This research highlights the role of AI in advancing neurotechnology for social good.
• The findings support the development of digital health solutions for cognitive health management in aging populations.