A GLM based approach to assess CFC.
Code to measure cross-frequency coupling between two signals as described in this manuscript:
A statistical framework to assess cross-frequency coupling while accounting for confounding analysis effects, J. Nadalin, L-E Martinet, E. Blackwood, M-S Lo, A. S. Widge, S. S. Cash, U. T. Eden, M. A. Kramer, 2019.
ExampleCode.m
: Run the cells in this file to produce example voltage traces and surfaces in (Phi_low, A_low, A_high)-space (as in Figure 4 of the manuscript). Four simulations are present: (i) no CFC, (ii) PAC only, (iii) AAC only, and (iv) both PAC and AAC.
simfun.m
: Code to simulate signals V_low, V_high, with induced cross-frequency coupling and measure output statistics R_PAC and R_AAC along with confidence intervals and p-values.
glmfun.m
: Code to evaluate the coupling statistics R_PAC and R_AAC, along with confidence intervals and p-values, between two signals.
glmfun_with_indicator.m
: Cn example of how to update glmfun to test for effect of condition (e.g. pre and post stimuli) on coupling
The voltage traces for the human data can be found in Patient_Data.mat
, and the rodent data can be found at https://github.com/tne-lab/cl-example-data
The Chaotic System Toolbox is required to generate surrogate data.
Any questions/comments please direct to Jessica Nadalin (jnadalin@bu.edu) or Mark Kramer (mak@bu.edu)