Unpacking the transient event dynamics that underlie spontaneous changes and induced responses in electrophysiology

This repository contains the scripts and software to run the simulations and real data analysis published in:

Andrew J. Quinn, Freek van Ede, Matthew J. Brookes, Simone G. Heideman, Magdalena Nowak, Zelekha A. Seedat, Diego Vidaurre, Catharina Zich, Anna C. Nobre, Mark W. Woolrich (2019) Unpacking Transient Event Dynamics in Electrophysiological Power Spectra. Brain Topography. https://doi.org/10.1007/s10548-019-00745-5

Requirements

The analysis requires the following software running on a Unix-Type operating system.

• MatLab 2018a or greater (may run on earlier versions but not tested)
• MatLab Signal Processing Toolbox
• MatLab Wavelet Toolbox
• HMM-MAR Toolbox
• distributionPlot package

Getting started

1. Download or clone this repository to your computer
2. Ensure that you have a recent MatLab with access to the Signal Processing Toolbox and Wavelet Toolbox.
3. Download HMM-MAR and distributionPlot.
4. Edit the file paths in the top of hmm_0_initialise to point to the location of these toolboxes on your computer
5. run hmm_0_initialise in MatLab, if this returns without error then you are good to go. If you see warnings then some dependencies may be missing, follow the instructions in the warning message.
6. Work through the hmm tutorial scripts

Contents

hmm_0_initialise runs the initial setup and configuration for these analyses

hmm_1_dynamics_illustration creates and plots the power simulations from figure 1

hmm_2_envelope runs an amplitude-envelope HMM on simulated data and creates figure 2

hmm_3_embedded runs an time-delay embedded HMM on simulated data and creates figure 3

hmm_4_realdata_trialwise runs an time-delay embedded HMM and task-evoked analysis on source-space MEG data and creates figures 4, 5, 6 and 7