DEVELOPING A FAIR PHYSIOLOGICAL SIGNAL DATA RESOURCE

DEVELOPING A FAIR PHYSIOLOGICAL SIGNAL DATA RESOURCE

Bibliographic Details
Main Author: Li, Xiaojin
Language:English
Published: Case Western Reserve University School of Graduate Studies / OhioLINK 2019
Subjects:
Computer Science
Online Access:http://rave.ohiolink.edu/etdc/view?acc_num=case1561916871230216
id ndltd-OhioLink-oai-etd.ohiolink.edu-case1561916871230216
record_format oai_dc
spelling ndltd-OhioLink-oai-etd.ohiolink.edu-case15619168712302162021-08-03T07:11:33Z DEVELOPING A FAIR PHYSIOLOGICAL SIGNAL DATA RESOURCE Li, Xiaojin Computer Science Physiological signal data, such as electroencephalography (EEG) and electrocardiography (ECG), provide critical diagnostic and prognostic information for a wide range of patient care and clinical research in many neurological diseases. With the recent exponential growth of physiological signal data, it becomes a challenge for researchers and clinicians to manage and process these data. Data resources following FAIR data principles (Findable, Accessible, Interoperable, and Reusable) along with effective data management and visualization will support the expanded use of existing annotated repositories. Such data resources will support, for example, the development of algorithms to automatically interpret signal data, and leading to the development of best practices for disease management.In this thesis, we focus on developing a FAIR physiological signal data resource and proposes a general architecture with a data preprocessing, management, visualization, sharing and analysis framework for large signal datasets. The overall innovation is a paradigm-changing approach for curating and reusing physiological signal data by overcoming the bottleneck due to the existing file-based, desktop-constrained, and inefficient data access paradigm. The contributions of the thesis consist of: a key-value based universal self-descriptive sequential data format; a scalable pipeline for clinical-event related signal data preprocessing; and a web-based system, named WaveSphere, for integrated signal data management, query, interactive visualization, annotation, and on-the-fly user-specified dataset generation and sharing. Based on our framework, we develop two customized systems: SpindleSphere for sleep spindle research, and SeizureBank with an analysis-ready dataset for epileptic seizure research. With our systems and the analysis-ready signal dataset, we propose several new signal analysis approaches applied to different clinical research studies. We propose a hybrid classifier for automatic sleep stage scoring and a parallel spindle detection algorithm for large-scale polysomnography signal data. In addition, we propose a feature-based seizure identification technique along with a cross-dataset evaluation benchmark for epileptic seizure research. Comparative evaluations of our framework with traditional file-based approaches demonstrate the improvement in scalability and performance with 1.86 terabytes (TB) of patient data. Therefore, our approach streamlines the ability of researchers to preprocess, manage, visualize, share and analyze large-scale physiological signal data in the neurological domain. 2019-08-28 English text Case Western Reserve University School of Graduate Studies / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=case1561916871230216 http://rave.ohiolink.edu/etdc/view?acc_num=case1561916871230216 unrestricted This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws.
collection NDLTD
language English
sources NDLTD
topic Computer Science
spellingShingle Computer Science
Li, Xiaojin
DEVELOPING A FAIR PHYSIOLOGICAL SIGNAL DATA RESOURCE
author Li, Xiaojin
author_facet Li, Xiaojin
author_sort Li, Xiaojin
title DEVELOPING A FAIR PHYSIOLOGICAL SIGNAL DATA RESOURCE
title_short DEVELOPING A FAIR PHYSIOLOGICAL SIGNAL DATA RESOURCE
title_full DEVELOPING A FAIR PHYSIOLOGICAL SIGNAL DATA RESOURCE
title_fullStr DEVELOPING A FAIR PHYSIOLOGICAL SIGNAL DATA RESOURCE
title_full_unstemmed DEVELOPING A FAIR PHYSIOLOGICAL SIGNAL DATA RESOURCE
title_sort developing a fair physiological signal data resource
publisher Case Western Reserve University School of Graduate Studies / OhioLINK
publishDate 2019
url http://rave.ohiolink.edu/etdc/view?acc_num=case1561916871230216
work_keys_str_mv AT lixiaojin developingafairphysiologicalsignaldataresource
_version_ 1719455598829895680

Similar Items