Trial-by-trial analysis of combined EEG and fMRI shows dynamics of cognitive function

Abstract

Event Abstract Back to Event Trial-by-trial analysis of combined EEG and fMRI shows dynamics of cognitive function Ana Diukova1*, P. Mallikarjun1, D.P Auer1 and P.F Liddle1 1 University of Nottingham, United Kingdom Introduction: While most of the studies analysed responses to visual stimuli [1] and the error-related negativity [3] & [2], we report reliable recording and results of the trial-by-trial analysis of the auditory ERP in spite of the noisy MR environment. Further application in auditory ERP research includes comparison between healthy controls and patients with psychosis. Methods: Simultaneous recording of EEG/fMRI was applied to identify BOLD changes associated with target detection task. Twenty healthy volunteers participated in the study. Paradigm and study design: Continuous EEG and whole-brain fMRI recordings were made while the participants were required to press a button for infrequent target stimuli and ignore frequent non-target stimuli. Stimuli were presented randomly (Target/Standard Tone=15%/85%; ISI=3 sec). EEG Recording: EEG data were recorded using the MRI compatible BrainAmpMR EEG amplifier and the BrainCap electrode cap with 30 sintered Ag/AgCl ring electrodes, sampling rate was 5000 Hz. fMRI Acquisition. fMRI was performed using a Philips 3T MRI scanner (TE/TR=35/2100ms, matrix 64x64, voxel size 3.25x3.25x3 mm, 35 slices). Data Analysis: EEG data was analysed using Brain Vision Analyzer software and EEGLAB 5.03. Extended infomax ICA was performed on the MR-denoised single subject continuous EEG data. Then data was epoched across the target stimuli. The 31 ICs for each subject were screened for maps resembling the typical ERP topography, the time-frequency distribution was plotted and the appropriate IC was chosen. The resulting single-trial IC ERP amplitude vector was used as a parametric modulation for MRI design matrix. The fMRI data was realigned, normalised and spatially smoothed with 8 mm Gaussian kernel, using SPM2. The design matrix was generated using synthetic HRF and its first derivative. The design matrix consisted of an onset vector of target stimuli, IC ERP amplitude vector as a parametric regressor and six rigid-body motion correction parameters entered as confounding covariates. Results: At the group level (n=20), the amplitude of single trial IC ERP co-varied with BOLD response within the frontal and parietal cortex, cingulate gyrus, thalamus, cerebellum, posterior temporal lobe and insula (Figure). These findings confirm previous fMRI evidence of widespread cerebral activity during performance of auditory oddball task, and furthermore demonstrate that the BOLD activation is correlated with EEG activity. tn_pic1 pic1