Visual-Evoked Potential (VEP) Event-Related Files from the General Anesthesia and Brain Activity (GABA) Study and Infant Sibling Project (ISP)

HAPPE+ER software was optimized for developmental data using a subset of EEG files from 4-month and 10-month old infants in the General Anesthesia and Brain Activity (GABA) Study. While medically necessary, 1-2 million infants each year undergo general anesthesia – a process that sedates brain activity and impacts early sensory experiences during a time typically characterized by rapid neurocognitive development. The GABA study examines sensory and socioemotional neurodevelopment longitudinally from infancy through childhood in individuals who have and who have never undergone general anesthesia during different windows in the first year of life. The GABA study was carried out in accordance with the recommendations of the Institutional Review Board at Boston Children’s Hospital. All caregivers provided assent for their child’s participation in the GABA study and for the release of the deidentified data. 

To facilitate the use and understanding of HAPPE+ER software, we have provided a subset of the validation files from the GABA study to serve as a tutorial dataset for how to run event-related potential (ERP) data through this automated processing pipeline. Five files (a.raw - e.raw) are from four 4-month and one 10-month old infants during a pattern reversal visual-evoked potential (VEP) paradigm. Pattern reversal occurred every 500 milliseconds. The pattern stimulus onset is indicated in each file by the code: vep+. Data was collected using a 128-channel EGI HydroCel Geodesic Sensor Net and EGI Net Amps 400, sampled at 1000Hz with an online reference to channel CZ. 

We have also included a subset of files from the Infant Sibling Project (ISP), an investigation examining infants at high versus low familial risk for autism spectrum disorder over the first 3 years of life. Baseline EEG data was collected while a young child sat in a parent’s lap watching a research assistant blow bubbles or show toys for several minutes. The Infant Sibling Project was carried out in accordance with the recommendations of the Institutional Review Board at Boston University and Boston Children’s Hospital (#X06-08-0374), with written informed consent from all caregivers prior to their child’s participation in the study.  All files here have been deidentified, including alteration of exact acquisition dates.  Acquisition times have not been altered. For additional information about data collection paradigms, and sample studies published on the larger ISP data set, please see the following references:

1. Levin, A. R., Varcin, K. J., O’Leary, H. M., Tager-Flusberg, H., and Nelson, C. A. (2017). EEG power at 3 months in infants at high familial risk for autism. J. Neurodev. Disord. 9, 1–13.
2. Gabard-Durnam, L.J., Wilkinson, C., Kapur, K. et al. Longitudinal EEG power in the first postnatal year differentiates autism outcomes. Nat Commun 10, 4188 (2019). https://doi.org/10.1038/s41467-019-12202-9

Here we provide a subset of the full dataset with a simulated VEP signal added into the data, as example files for HAPPE+ER. To create these files, we selected a subset of 39 spatially-distributed channels in the baseline EEG files and created sixteen 30-second files using continuous segments of relatively artifact-free (clean) baseline data from the full-length files. Next, from 30-second sections of the same individuals’ EEG that were artifact-laden, we ran ICA and extracted artifact independent components (identified by an expert and labeled artifact by both ICLabel and MARA automated algorithms). We inserted the artifact ICs into that individual’s clean 30-second data segment to create an additional 16 artifact-added files. We then selected a channel from a simulated VEP dataset (included here as simulated_full.set) with a stereotyped and prominent simulated VEP waveform, in this case Oz, and added its timeseries (included here as simulated_singleChan.set) to each channel of the clean and artifact-added files to create two VEP datasets with a known ERP morphology (sim-artifact_a-p and sim-clean_a-p). For additional information about the creation of this simulated data and VEP data with a known ERP morphology, please refer to Monachino et al., in revision; DOI: https://doi.org/10.1101/2021.07.02.450946.

Additional files included below are the HAPPE+ER data and pipeline quality metric output spreadsheets for the five GABA study data files for an example run, the output spreadsheet containing the ERP timeseries from the generateERPs script, the .mat file containing the parameter settings for HAPPE+ER for that run, an Excel file with the bad channels for each file, and a tutorial document illustrating the results of this example run.