Increased electromagnetic variability in Lewy body dementia: A MILOS study

Abstract

Background Lewy body dementia (LBD) includes dementia with Lewy bodies and Parkinson’s disease dementia, which shared a common pathological core feature of cognitive fluctuations. However, the clinical identification of fluctuations is still controversial. With Electroencephalogram (EEG) and Magnetoencephalography (MEG), it becomes possible to observe the neural correlates of fluctuations with a high temporal resolution in LBD. In this multimodel imaging and computational modelling of dementia with lewy bodies (MILOS) study, we combined EEG/MEG to explore the electro‐magnetic biomarkers of fluctuations in LBD. We hypothesized that LBD would show an increased variation in EEG/MEG compared with controls, indicating a potential underlying mechanism for fluctuations in LBD. Method Nine patients with LBD (mean age=72) and twelve healthy controls (mean age=67) were recruited and an animal/non‐animal object recognition task was conducted. 64‐channel EEG and 306‐channel MEG (including gradiometer (GRM) and magnetometer (MAG)) were simultaneously recorded during this task. Subsequently participants had MRI with T1‐weighted images for EEG/MEG source coregistration. Forward models for EEG/MEG were 3‐Shell Sphere and Single Shell respectively and minimum norm Least‐Squares as the inversion method. The variations of source post‐stimulus in object recognition was assessed. We used standardized clinical batteries to test cognitive functions and overall scalp sensors for EEG/MEG sensor‐level analyses. Result HC had shorter reaction time and higher accuracy rate than LBD in object recognition. The variations of reaction time across trials in object recognition was significantly higher in LBD than HC. LBD had weaker overall EEG amplitudes and stronger MEG variations across trials for post stimulus than HC. Phase lock value (PLV) indicated different EEG synchronization patterns between groups with higher theta and lower beta synchronization in LBD than HC. The MEG variations across trials in LBD were involved with theta, alpha and beta bands. The MEG variations were negatively correlated with MMSE, ACER, ACE‐III but positively correlated with UPDRS scores. Source variations in LBD involved a network of parietal areas, including postcentral gyrus and superial parietal areas. Conclusion Our study revealed the possible biomarkers of fluctuations in LBD by using EEG/MEG. It highlighted the importance of electrophysiological correlates of fluctuations in understanding the etiology of LBD.