Neurophysiological gradient of cortical traveling waves

1673 

Abstract Submission 

Xiaobo Liu¹, Sylvain Baillet²

¹McGill University, Montreal, Quebec, ²Montreal Neurological Institute, Montreal, Quebec

Xiaobo Liu  
McGill University
Montreal, Quebec

Sylvain Baillet  
Montreal Neurological Institute
Montreal, Quebec

Identifying the spatiotemporal structure of macroscopic brain dynamics remains a challenge to understand the mechanisms of human brain functions and their dysfunctions. Traveling waves are recurrent and consistent observations at multiple spatial scales and in diverse preparations, including animal models and noninvasive human brain data collected with time-resolved techniques. However, practical tools for characterizing these phenomena of cortical flow in a quantitative manner are only currently fledging. We proposed to use 3-D optical flow decompositions of time-resolved cortical activity, which enable the rigorous delineation of various spatiotemporal elements of propagation patterns such as sources and sinks and translational components (Lefèvre & Baillet, 2008 & 2009 ; Khan et al., 2011). Here we expand this toolkit to test whether intrinsic activity of the human brain showcases a predominant spatiotemporal structure of propagation across the cortex.

Our data analysis pipeline is summarized in Fig.1. We used resting-state MEG and structural MRI (T1) data from the Open MEG Archive (OMEGA). The MEG data were analyzed using Brainstorm with default parameters, unless specified. Preprocessed-MEG data were resampled at 368 Hz, based on the observation that 95% of the sensor signal power spectrum density (PSD) was contained below 92 Hz across participants. We used Brainstorm's default overlapping-spheres and minimum-norm imaging method for mapping sensor data on the cortical surface of the participants.

To characterize the dynamics of cortical activity, we derived the optical flow of cortical activity at each time point and ran the Helmholtz-Hodge decomposition (HHD ; Khan et al., 2011) of this vector field, which yielded propagation patterns into elemental diverging and curling components. We further assigned local minima of diverging components as sources and local maxima as sinks of cortical activity, respectively. All inferential statistics were run using two-sample t-tests with false discovery rate (FDR) correction.

We observed a consistent gradient of cortical flow dynamics along the cortical anatomy (Fig. 2): frontal cortical sources were significantly stronger than cortical sinks strength in occipital regions ( Pfdr < 0.05). We also observed that the sinks in the limbic network have greater strength than the sources of the sensorimotor network. These observations were systematic along the antero-posterior sagittal (R = -0.60, P < 0.0001) and the dorso-ventral axial direction (R = 0.38, P < 0.0001). These results suggest that spontaneous, task-free brain activity propagates according to a marked spatiotemporal structure consistent with the grand traits of neuroanatomy.

Our study implements a novel framework for quantifying cortical propagation patterns, to delineate the spatio-temporal organization of cortical activity. The results of our investigation indicate that spontaneous neurophysiological activity displays a high degree of organization within a kinetic cortical framework, showcasing a discernible trajectory from sensation to cognition.

EEG/MEG Modeling and Analysis ¹

MEG ²

Neurophysiology of Imaging Signals

Computational Neuroscience

Data analysis

MEG

Modeling