A Latent State Space Model for Estimating Brain Sources and Connectivity from Electroencephalogram (EEG) Data

• Activity Number: 443 - Latent Variables, Causal Inference, Machine Learning and Other Topics in Mental Health Statistics
• Type: Contributed
• Date/Time: Thursday, August 12, 2021 : 4:00 PM to 5:50 PM
• Sponsor: Mental Health Statistics Section
• Author(s): Qinxia Wang* and Ji Meng Loh and Yuanjia Wang
• Companies: Columbia University and New Jersey Institute of Technology and Columbia University
• Keywords: State space models; Kalman filter; Multi-channel EEG signals; Patient heterogeneity ; Latent sources; Alcoholism

Abstract

Modern neuroimaging technologies have substantially advanced the measurement of brain activities. With superb temporal resolution capturing fast-changing brain activities, Electroencephalogram (EEG) has emerged as an increasingly useful tool to study cortical connectivity. Challenges of modeling EEG signals include complex brain activities involving interactions of unknown sources, low signal-to-noise ratio and between-subject heterogeneity. In this work, we propose a state space model that jointly analyzes multi-channel EEG signals and learns dynamics of sources corresponding to brain activities. Our model borrows strength from spatially correlated signals and uses low-dimensional latent sources to explain observed channels. It can account for patient heterogeneity and quantify the effect of subject covariates on the latent space. We use an EM algorithm, Kalman filtering and bootstrap resampling to fit the state space model and provide comparisons between patient diagnostic groups. We apply the approach to a case-control study of alcoholism and reveal significant attenuation of brain activities in response to visual stimuli in alcoholic patients compared to healthy controls.