PROJECT SUMMARY (See instructions):
A key barrier in using neurophysiological data for research is the crushing computational burden involved in
processing large volumes of complex signal data, typically in European Data Format (EDF). The goal of this
project is to create an AI-ready resource of semantically annotated collection of electroencephalogram (EEG)
recordings and sleep polysomnography data, called the Neurophysiological AI-Ready Data Resource
(NAIRD), to transform the access and sharing of such resources for brain health research. NAIRD will
leverage data sources from prior repositories that cover two major brain health domains: epilepsy, as a part
of the Center for Sudden Unexpected Death in Epilepsy (SUDEP) Research (CSR), and sleep, as a part of
the National Sleep Research Resource (NSRR). NAIRD will make this rich collection of electrophysiological
signals with associated individual-level health outcomes not only FAIR (Findability, Accessibility,
Interoperability, and Reusable), but also interactable and AI-ready. Pre-computed innovative features will be
readily available with raw data source in a way that is free from redundant-processing needs using our future-proof data format in a scalable database environment. This overall goal will be achieved through four Aims:
Aim 1: Develop and operationalize a future-proof, universal self-descriptive sequential data format (U2S) to
represent neurophysiological data, with a web-based interactive visualization interface for on-the-fly creation
of user-specified data subsets; Aim 2: Convert the prospectively acquired Center for SUDEP Research data
collection to U2S format and import the converted data to WaveSphere, our NAIRD data access platform
designed and developed specifically to take advantage of the U2S data format; Aim 3: Convert the
retrospectively integrated National Sleep Research Resource polysomnography data collection to U2S
format and import the converted data to WaveSphere for interactive access and secondary analysis; Aim 4:
Evaluate and disseminate WaveSphere for continued enhancement and outreach. If successful, NAIRD will
achieve data interactability and usability at a level not seen before, enabling and supporting brain health
research by creating one of the largest and most comprehensive resource of FAIR, interactable, and AI-ready neurophysiological data. Areas impacted by NAIRD-enabled research include neurological diseases
and related conditions, through the application of machine learning algorithms for earlier predication,
prevention and improved care management.