24 August 2017 Adaptive windowing and windowless approaches to estimate dynamic functional brain connectivity
Author Affiliations +
Abstract
In this work, we discuss estimation of dynamic dependence of a multi-variate signal. Commonly used approaches are often based on a locality assumption (e.g. sliding-window) which can miss spontaneous changes due to blurring with local but unrelated changes. We discuss recent approaches to overcome this limitation including 1) a wavelet-space approach, essentially adapting the window to the underlying frequency content and 2) a sparse signal-representation which removes any locality assumption. The latter is especially useful when there is no prior knowledge of the validity of such assumption as in brain-analysis. Results on several large resting-fMRI data sets highlight the potential of these approaches.
Conference Presentation
© (2017) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Maziar Yaesoubi, Vince D. Calhoun, "Adaptive windowing and windowless approaches to estimate dynamic functional brain connectivity", Proc. SPIE 10394, Wavelets and Sparsity XVII, 1039411 (24 August 2017); doi: 10.1117/12.2274425; https://doi.org/10.1117/12.2274425
PROCEEDINGS
11 PAGES + PRESENTATION

SHARE
Back to Top