Abstract #1125

Efficacy of different dynamic functional connectivity methods to capture cognitively relevant information

Hua Xie^1,2, Javier Gonzalez-Castillo², Daniel A. Handwerker², Peter Molfese², Peter A. Bandettini^2,3, and Sunanda Mitra¹

¹Department of Electrical & Computer Engineering, Texas Tech University, Lubbock, TX, United States, ²Section on Functional Imaging Methods, Laboratory of Brain and Cognition, National Institute of Mental Health, National Institutes of Health, Bethesda, MD, United States, ³FMRI Core, National Institute of Mental Health, National Institutes of Health, Bethesda, MD, United States

With a multitask dataset (rest, memory, video, and math) serving as ground truth, we evaluated the efficacy of four different methods of estimating dynamic functional connectivity (dFC)—namely sliding window correlation (SWC), sliding window correlation with L1-regularization (SWC_L1), dynamic conditional correlation (DCC), and multiplication of temporal derivatives (MTD)—to capture cognitively relevant information. We used dFC estimates of each method as inputs for k-means, and evaluated how well they segregate scan periods for different tasks. We found that moving average DCC produces best results, especially for short window length (WL ≤ 9sec), suggesting DCC may more reliably reveal dFC linked to mental states.

This abstract and the presentation materials are available to members only; a login is required.

Join Here