Abstract
Electrophysiological methods, that is M/EEG, provide unique views into brain health. Yet, when building predictive models from brain data, it is often unclear how electrophysiology should be combined with other neuroimaging methods. Information can be redundant, useful common representations of multimodal data may not be obvious and multimodal data collection can be medically contraindicated, which reduces applicability. Here, we propose a multimodal model to robustly combine MEG, MRI and fMRI for prediction. We focus on age prediction as a surrogate biomarker in 674 subjects from the Cam-CAN dataset. Strikingly, MEG, fMRI and MRI showed additive effects supporting distinct brain-behavior associations. Moreover, the contribution of MEG was best explained by cortical power spectra between 8 and 30 Hz. Finally, we demonstrate that the model preserves benefits of stacking when some data is missing. The proposed framework, hence, enables multimodal learning for a wide range of biomarkers from diverse types of brain signals.
Highlights
Non-invasive electrophysiology assumes a unique role in clinical neuroscience
Our study focuses on the following questions: 1) Can magnetic resonance imaging (MRI)-based prediction of age be enhanced with MEG-based electrophysiology? 2) Do functional MRI (fMRI) and MEG carry non-redundant clinically relevant information? 3) What are the most informative electrophysiological markers of aging? 4) Can potential advantages of multimodal learning be maintained in the presence of missing values?
Linear predictions were based on distinct features from anatomical MRI, fMRI and MEG that have been commonly associated with aging
Summary
Non-invasive electrophysiology assumes a unique role in clinical neuroscience. Magneto- and electophencephalography (M/EEG) have an unparalleled capacity for capturing brain rhythms without penetrating the skull. First results suggest that the prediction error of models trained to learn age from brain data of healthy populations provides clinically relevant information (Cole et al, 2018; Ronan et al, 2016; Cole et al, 2015) related to neurodegenerative anomalies, physical and cognitive decline (Kaufmann et al, 2019). Compared to EEG, MEG benefits from the magnetic transparency of the skull, which facilitates source localization by reducing the risk of errors due to an incorrect head conductivity model, and by limiting the large-scale mixing of neural sources This significantly increases the signal-to-noise ratio for MEG in higher frequencies, rendering it a formidable technique for studying cortical oscillatory activity (Lehtelaet al., 1997; Gobbeleet al., 1998). Our study focuses on the following questions: 1) Can MRI-based prediction of age be enhanced with MEG-based electrophysiology? 2) Do fMRI and MEG carry non-redundant clinically relevant information? 3) What are the most informative electrophysiological markers of aging? 4) Can potential advantages of multimodal learning be maintained in the presence of missing values?
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
