A New Cross-Frequency Granger Causality Method for Investigating Cortical Dynamics
Abstract
A New Cross-Frequency Granger Causality Method for Investigating Cortical Dynamics Objectives: The goal of this short-term innovative research (STIR) project is to develop and assess a novel method for measuring multi-scale interactions across brain regions based on cross-frequency Granger causality. Approach: During the proposed five-month project, we will first develop and calibrate a cross-frequency Directed transfer Function (xDTF) method (Aim 1) and then test the method using experimental data (Aim 2). Aim 1 will be accomplished by applying complex demodulation to shift the frequencies of time-series signals (Aim 1A) and multivariate vector autoregressive modeling (MVAR) to estimate Granger causality between different frequency-shifted time series (Aim 1B). The xDTF method will be calibrated using simulated data (Aim 1B). We will test the xDTF method using human electrophysiology recordings (scalp, intracranial) and compare results with phase-amplitude coupling estimates based on the same experimental data (Aim 2). Preliminary studies have confirmed the feasibility of the work proposed. Significance: This project will establish a novel method for investigating non-linear interactions across brain regions based on frequency coupling. The proposed xDTF method differs from other MVAR-based Granger causality methods that identify interactions only in the same frequency. We anticipate that future applications of this method will include investigation of causal interactions in multimodal recordings on different time-scales as well as external stimuli, including time- varying sound streams in the environment.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Oct 11, 2018
- Source ID
- W911NF1710532
Entities
People
- Dana Boatman
Organizations
- Army Contracting Command
- Johns Hopkins University
- United States Army