# A multi-session simultaneous EEG-fMRI dataset with online experience sampling

Here we introduce a multi-session dataset that includes simultaneous EEG-fMRI with online measures of continuous behavior and spontaneous mental experience. Data components, organized in Brain Imaging Dataset Structure (BIDS) format, include simultaneous EEG-fMRI recordings with carbon wire loop sensors embedded in EEG caps for artifact removal, electrocardiogram (ECG), behavioral task responses, experience sampling ratings, and mental health surveys, from 24 healthy individuals aged 18-35. Tasks performed during EEG-fMRI were the Gradual Onset Continuous Performance Task (GradCPT) and a resting state condition with intermittent experience sampling assessing 13 unique dimensions of thought contents and processes (36 trials including 468 total ratings per participant). The same task protocol was completed on two different days, resulting in approximately 1 hour 20 minutes of EEG-fMRI data per individual. These data enable the study of the neural bases of various spontaneous cognitive processes, including attentional fluctuations and mind-wandering, thereby promoting insights into the behavioral relevance of resting state brain activity. The dataset also provides a means to study the reliability of temporal relationships between fMRI and EEG data features across different sessions within the same individuals. 

** Inclusion Criteria **
Participants in the study met the following inclusion criteria: 
- Aged 18 to 35 years old
- Spoke English
- Right handed
- Normal/corrected-to-normal vision

** Exclusion Criteria ** 
Participants meeting any of the following criteria were excluded from the study: 
- Unable to provide consent
- Reported having a current or history of psychiatric/neurological disorders
- Reported having a chronic medical condition
- Were pregnant
- Were prisoners
- Were unable to understand English 
- Were contraindicated for MRI
- Had allergies to saline gel
- Had cold, flu or COVID-19 symptoms within the two weeks preceding participation 
- Were unable to fit into one of our three EEG cap sizes (54cm, 56cm, and 58cm circumference)
- Wore glasses and did not have access to contacts. 

# Consent
Institutional review board approval and consent were obtained. Participants were recruited from the Philadelphia, Pennsylvania area, primarily drawing from communities in and around Drexel University and Temple University

# Clinical Measures
- State Trait Anxiety Inventory (STAI-S)
- DSM-5 Self-Rated Level 1 Cross-Cutting Symptom Measure-Adult (DSM XC)
- WHO Disability Assessment Schedule-12 (WHODAS-12)
- General Anxiety Disorder-7 (GAD-7)
- Patient Health Questionnaire-9 (PHQ-9)
- Ruminative Response Scale (RRS)
- Mind-Wandering Deliberate-Spontaneous scale (MWD-S)

# MRI Acquisition Parameters 
MRI data acquisition. MRI data were acquired using a 64-channel head coil on 3.0 Tesla Siemens MAGNETOM Prisma. To accommodate the EEG-fMRI set-up, the 64-channel head coil included an aperture for routing cables from the top of the EEG cap into the scanner bore to connect to EEG amplifiers. The first 22 subjects and the first session of the 23rd subject were run on Siemens software version Syngo MR E11. The second session of the 23rd subject and both sessions of the 24th subject were run on software version Syngo MR XA30. All acquisition parameters remained consistent across both software versions with the exception of “nonlinear gradient correction,” set to “false” for the first 22 subjects and set to “true” after the software upgrade. This change reflected differences in the default settings of the scanner rather than a deliberate change made by the researchers. After localizer scans, a 3D high resolution MPRAGE structural T1-weighted image was acquired (TR = 2400 ms; TI = 1000 ms; TE = 2.22 ms; slices = 208; FoV = 256 x 240mm, voxel size = 0.8 mm3 isotropic; flip angle = 8°; partial Fourier off; pixel bandwidth = 220Hz/Px) during the first scanning session. During a participant’s second scanning session, a fast T1-weighted, 2D MRI sequence was collected (TR = 190 ms; TE = 2.46 ms; in-plane voxel size = 0.8 x 0.8mm, slices = 25; slice thickness = 4.00 mm; flip angle = 70°; pixel bandwidth = 280 Hz/Px). Following the T1-weighted scan, we ran a B0 field map for the purpose of correcting echo-planar imagining (EPI) fMRI images (TR = 789 ms; TE 1 = 4.92 ms; TE 2 = 7.38 ms; slices = 81; FoV = 192mm x 192mm; voxel size = 2.0 mm3 isotropic; flip angle = 45°; partial Fourier off; pixel bandwidth = 668 Hz/Px). All BOLD fMRI sequences were acquired with a CMRR multiband gradient-echo EPI single shot sequence aligned with the AC-PC plane (TR = 2000 ms; TE = 25.00 ms; flip angle = 70°; slices = 81; FoV = 192mm x 192mm; voxel size = 2.0 mm3 isotropic; multi-band acceleration factor = 3; phase encoding direction = anterior-to posterior). A total of four fMRI runs across two tasks (one GradCPT run, three rs-ES runs) were run for 46 out of 47 scanning sessions, where one session (sub-003_ses-001) was ended early due to time constraints and only 3 runs were collected.  

# EEG Acquisition 
EEG data were collected using an MR-compatible system (Brain Products Gmbh, Gilching, Germany), including the BrainAmp MRPlus, the BrainAmp ExG, and the BrainCap MR with four Carbon Wire Loop (CWL) sensors embedded. Additional EEG-fMRI specific equipment included the Brain Products PowerPack (MR-compatible battery pack for the amplifiers), Triggerbox (sends and receives MRI and stimulus triggers to EEG recording software), SyncBox interface and main unit (syncs EEG phase with MR clock), the BrainVision USB 2 Adapter (connection hub for all EEG components), and the MR sled (mobile surface for mounting the EEG components into the scanner bore). Recorded channels included 4 CWLs, 31 cortical channels and one electrocardiography (ECG) channel (channel 32) placed on the back. In addition, the cap also contained a reference and ground electrode. The cap’s built-in serial current-limiting resistors included 15kOhms of resistance for the ground and reference electrodes, 10kOhms of resistance for all 31 EEG electrodes, and 20kOhms of resistance for the ECG electrode. All electrodes were connected to the BrainAmp MRPlus. The four CWLs sit over the frontal left, frontal right, parietal left, and parietal right locations on the cap and were connected to the bipolar BrainAmp ExG. The CWL channels record minute head movements in the scanner environment, used to perform advanced cardioballistic, helium pump, and movement artifact correction. Cortical electrodes were arranged according to the international 10-20 system. Electrodes were filled using high-chloride abrasive electrolyte-gel (Abralyt HiCl). A layer of Surgilast, an elastic dressing retainer, was placed over the cap after set-up to ensure electrode contact with the scalp and minimize cap movement. Electrode impedance was kept below 20kOhms and measured twice, once during set up, then inside the scanner just before the closing of the head coil. The EEG data were recorded using BrainVision Recorder V.1.25.0201 software at a sampling rate of 5kHz on a Windows operating system using 

** TR markers in EEG data **
As MRI TR markers come in, they are recorded in the EEG data to sync timing between EEG and fMRI data during EEG artifact correction and analyses. The length of the resting state task varied from run to run depending on the participant’s pace and the randomized duration of the fixation cross across trials. This necessitated manual termination of the scan upon conclusion of each rs-MDES run. In the majority of cases, the forced stop would end the functional scan in the middle of the collection of a volume, meaning that the EEG recording would receive an extra fMRI volume marker appended to the end of the recording; however, the scanner would not complete the measurement. This appended an extra fMRI marker (“T  1”) to the end of the EEG data which did not correspond to an fMRI volume. Exceptions to this rule are two runs which had two additional fMRI markers on the end of the EEG recording (sub-013_ses-002_run-003; sub-019_ses-001_run-003), four runs where there was one less fMRI marker in the EEG data than fMRI volumes recorded (sub-001_ses-002_run-003; sub-023_ses-002 all runs), and eight runs where the recorded fMRI markers and volumes were correctly aligned (sub-002_ses-001_run-002, run-003; sub-003_ses-002_run-001; sub-008_ses-001_run-001; sub-008_ses-002_run-002; sub-009_ses-002_run-002; sub-010_ses-001_run-001; sub-012_ses-002_run-003). This issue affects only the last markers in the run, where up to two volume markers in the EEG data may need to be omitted from analyses. All fMRI markers present in the EEG data are aligned with the first fMRI volume. 

# References for BIDS conversion
----------
Appelhoff, S., Sanderson, M., Brooks, T., Vliet, M., Quentin, R., Holdgraf, C., Chaumon, M., Mikulan, E., Tavabi, K., Höchenberger, R., Welke, D., Brunner, C., Rockhill, A., Larson, E., Gramfort, A. and Jas, M. (2019). MNE-BIDS: Organizing electrophysiological data into the BIDS format and facilitating their analysis. Journal of Open Source Software 4: (1896).https://doi.org/10.21105/joss.01896

Pernet, C. R., Appelhoff, S., Gorgolewski, K. J., Flandin, G., Phillips, C., Delorme, A., Oostenveld, R. (2019). EEG-BIDS, an extension to the brain imaging data structure for electroencephalography. Scientific Data, 6, 103.https://doi.org/10.1038/s41597-019-0104-8