Simultaneous EEG-fMRI during a neurofeedback task, a brain imaging dataset for multimodal data integration

Abstract

Combining EEG and fMRI allows for integration of fine spatial and accurate temporal resolution yet presents numerous challenges, noticeably if performed in real-time to implement a Neurofeedback (NF) loop. Here we describe a multimodal dataset of EEG and fMRI acquired simultaneously during a motor imagery NF task, supplemented with MRI structural data. The study involved 30 healthy volunteers undergoing five training sessions. We showed the potential and merit of simultaneous EEG-fMRI NF in previous work. Here we illustrate the type of information that can be extracted from this dataset and show its potential use. This represents one of the first simultaneous recording of EEG and fMRI for NF and here we present the first open access bi-modal NF dataset integrating EEG and fMRI. We believe that it will be a valuable tool to (1) advance and test methodologies for multi-modal data integration, (2) improve the quality of NF provided, (3) improve methodologies for de-noising EEG acquired under MRI and (4) investigate the neuromarkers of motor-imagery using multi-modal information.

Extracting electrophysiological correlates of functional magnetic resonance imaging data using the canonical polyadic decomposition

“…Results in this section are focused on the second motor‐imagery dataset from Lioi et al ( 2020 ) and the resting‐state dataset from Deligianni et al ( 2014 , 2016 ). Hence, mention of the motor‐imagery dataset in the following section refers by default to the second motor‐imagery dataset.…”

“…Data released by Lioi and colleagues described in (Lioi et al, 2020 ) contain two sets of EEG‐fMRI data acquired during a motor‐imagery neurofeedback task. Subjects were asked to perform motor‐imagery of clenching their right hand according to a block design, alternating between motor‐imagery and resting blocks.…”

“…For both motor‐imagery datasets, pre‐processed EEG data are located in subjects' eeg_pp subdirectories within the derivatives directory. Readers may refer to (Lioi et al, 2020 ) for additional information regarding the motor‐imagery datasets. For all datasets, raw BOLD‐fMRI data were taken from subjects' main directory (i.e.…”

“…| EMPIRICAL DATA 3.1 | Open-source datasets We further apply our methodology to three open-source EEG-fMRI datasets, covering resting-state and task-based experimental conditions.Data released by Lioi and colleagues described in(Lioi et al, 2020) contain two sets of EEG-fMRI data acquired during a motor-imagery neurofeedback task. Subjects were asked to perform motor-imagery of clenching their right hand according to a block design, alternating between motor-imagery and resting blocks.…”

Bimodal electroencephalography-functional magnetic resonance imaging dataset for inner-speech recognition

“…The tool is developed in matlab and is an open source. Lioi et al [8] proposed a bimodal motor imagery neurofeedback based dataset aquired with simultaneous EEG-fMRI recordings. The data were recorded from 30 subjects performed kinaesthetic motor imagery of the right hand to bring the ball to the target.…”