Streams of Thought: An ICA Methodology for Lagged Resting State Analysis
Erik Jahner1,3, Xiao-Fei Yang2,3, Mary Helen Immordino-Yang2,3; 1University of California Riverside, 2University of Southern California, 3Brain and Creativity Institute
The human brain is an ongoing dynamic system not activated by experience but nudged from intrinsic activity into new network configurations during perception and learning. Ongoing neural activity during rest is assumed to reflect these intrinsic dynamics in a relatively closed system state revealing the neural ensembles of thought. Traditionally, inter-regional connectivity in this system is measured by obtaining time-locked correlations in BOLD activity using fMRI. However, we know that neural activity unfolds across time and is not isolatent at some behavioral or perceptual reference point. Indications are that lagged network dynamics are a more fundamental property of network dynamics than the traditional resting-state network approaches. This exploratory study is a theoretical and methodological examination of how a lagged analysis of resting state dynamics in fMRI could expose persistent representations of knowledge in the neocortex. A novel approach using independent component analysis (ICA) on surface maps is applied to resting-state data from 54 adolescents. ICA methodologies allow for the reconstruction of individual representations of the lagged threads for use as regressors with some notable limitations.These methods reveal lagged structures with interpretable but different topographical and temporal information than traditional resting-state analyses. The group level results are symmetrical between hemispheres and early extracted components may represent high level perceptual processes. These results also do not correlate with known traditional resting-state networks further supporting the theory that traditional networks are not internally isolatent.The methods and interpretation of ICA as applied to a lagged matrix of resting state dynamics are presented.
Topic Area: METHODS: Neuroimaging