Source imaging of high-density visual evoked potentials with multi-scale brain parcellations and connectomes

Abstract

We describe the multimodal neuroimaging dataset VEPCON (OpenNeuro Dataset ds003505). It includes raw data and derivatives of high-density EEG, structural MRI, diffusion weighted images (DWI) and single-trial behavior (accuracy, reaction time). Visual evoked potentials (VEPs) were recorded while participants (n=20) discriminated briefly presented faces from scrambled faces, or coherently moving stimuli from incoherent ones. EEG and MRI were recorded separately from the same participants. The dataset contains pre-processed EEG of single trials in each condition, behavioral measures, structural MRIs, individual brain parcellations at 5 spatial resolutions (83 to 1015 regions), and the corresponding structural connectomes computed from fiber count, fiber density, average fractional anisotropy and mean diffusivity maps. For source imaging, VEPCON provides EEG inverse solutions based on individual anatomy, with Python and Matlab scripts to derive activity time-series in each brain region, for each parcellation level. The BIDS-compatible dataset can contribute to multimodal methods development, studying structure-function relations, and to unimodal optimization of source imaging and graph analyses, among many other possibilities.

Structure supports function: Informing directed and dynamic functional connectivity with anatomical priors

“…One participant was excluded because of too many motion artifacts, leaving 19 datasets for analysis. Further details on the recordings and preprocessing pipeline can be found in the original manuscript of the VEPCON dataset (OpenNeuro Dataset ds003505; Pascucci et al, 2021 ).…”

“…A detailed description of the MR acquisition and preprocessing can be found in the original manuscript of the VEPCON dataset (OpenNeuro Dataset ds003505; Pascucci et al, 2021 ).…”

Connectome spectrum electromagnetic tomography: a method to reconstruct electrical brain source-networks at high-spatial resolution

“…Recently, two multimodal neuroimaging datasets have been released that have great potential to help develop source imaging tools as well as validate their performance and clinical relevance (VEPCON datasets from now on) 21 . In these datasets, visual evoked potentials were recorded for 20 participants while they discriminated between either briefly presented faces and scrambled faces, or coherently moving stimuli and incoherent ones.…”

“…Visual evoked potentials of well-known neurophysiological paradigms (such as face or motion perception) provide data with high signal-to-noise ratio (in comparison to other types of EEG experimental paradigms) and their activation response is well documented in other imaging modalities or animal models. We reconstructed brain activity maps from the VEPCON's EEG data 21 (Fig. 3a).…”

“…Others have already attempted to use structural connectivity information to constrain the M/EEG inverse solution by enforcing smoothness among connected sources 21,22 , sparsity among connectivity wavelets 19 , and other methods 20 . These articles suggest a potential improvement in the localization accuracy due to the anatomical constraints.…”