Neurophysiological hallmarks of Huntington’s disease progression: an EEG and fMRI connectivity study

Abstract

Electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) can provide corroborative data on neurophysiological alterations in Huntington’s disease (HD). However, the alterations in EEG and fMRI resting-state functional connectivity (rsFC), as well as their interrelations, at different stages of HD remain insufficiently investigated. This study aimed to identify neurophysiological alterations in individuals with preclinical HD (preHD) and early manifest HD (EMHD) by analyzing EEG and fMRI rsFC and examining their interrelationships. We found significant differences in EEG power between preHD individuals and healthy controls (HC), with a decrease in power in a specific frequency range at the theta-alpha border and slow alpha activity. In EMHD patients, in addition to the decrease in power in the 7–9 Hz range, a reduction in power within the classic alpha band compared to HC was observed. The fMRI analysis revealed disrupted functional connectivity in various brain networks, particularly within frontal lobe, putamen-cortical, and cortico-cerebellar networks, in individuals with the HD mutation compared to HC. The analysis of the relationship between EEG and fMRI rsFC revealed an association between decreased alpha power, observed in individuals with EMHD, and increased connectivity in large-scale brain networks. These networks include putamen-cortical, DMN-related and cortico-hippocampal circuits. Overall, the findings suggest that EEG and fMRI provide valuable information for monitoring pathological processes during the development of HD. A decrease in inhibitory control within the putamen-cortical, DMN-related and cortico-hippocampal circuits, accompanied by a reduction in alpha and theta-alpha border oscillatory activity, could potentially contribute to cognitive decline in HD.

The performance of domain-based feature extraction on EEG, ECG, and fNIRS for Huntington’s disease diagnosis via shallow machine learning

“…Odish et al (2018) collected 3minute EEG data samples from 26 HD-gene carriers and 25 controls, and tested only SVMs to classify patients with 83% accuracy based on the single feature of power spectral density. Ponomareva et al (2023) used computational analysis to reveal notable differences in both EEG power and fMRI functional connectivity between preclinical HD patients and healthy controls, indicating such anomalies could be detected even before symptomaticity.…”