Neurology and Neurosurgery

Sept. 26, 2024

Mayo Clinic researchers and colleagues have developed a new approach for analyzing routine electroencephalography (EEG) recordings to identify brain activity associated with neurodegenerative diseases. The artificial intelligence method doesn't require prior selection of EEG channels or frequency bands — avoiding bias and potentially making EEG a robust tool for clinical diagnosis of neurodegenerative disease.

The researchers analyzed more than 12,000 EEG recordings in the Mayo Clinic EEG database. Using machine learning techniques, the researchers identified distinct patterns of brain activity corresponding to various stages of cognitive decline. The patterns differentiate the brain waves of individuals with Alzheimer's disease and Lewy body dementia from those of cognitively unaffected individuals.

Traditional methods of quantitative EEG analysis often require prior selection of clinically meaningful EEG features. As a result, those methods are susceptible to bias, limiting the clinical utility of routine EEGs in the diagnosis and management of neurodegenerative disorders.

"Compared to other clinically available tests, scalp EEGs are noninvasive and relatively inexpensive tests that are available at most neurological practices worldwide," says David T. Jones, M.D., a neurologist at Mayo Clinic in Rochester, Minnesota. "With continued development, data-driven methods might improve the clinical utility of EEG in memory care. EEG potentially could help with early identification of mild cognitive impairment and differentiate between various neurodegenerative causes."

For more information

Li W, et al. Data-driven retrieval of population-level EEG features and their role in neurodegenerative diseases. Brain Communications. 2024;6:fcae227.

Refer a patient to Mayo Clinic.