상세정보
Export to Refworks부가기능
Data Driven Algorithms for Analyzing Imaging and Electroencephalography (EEG) Data in Neuroscience
상세 프로파일
| 자료유형 | 학위논문 |
|---|---|
| 서명/저자사항 | Data Driven Algorithms for Analyzing Imaging and Electroencephalography (EEG) Data in Neuroscience. |
| 개인저자 | Zheng, Jingyi. |
| 단체저자명 | University of California, Davis. Statistics. |
| 발행사항 | [S.l.]: University of California, Davis., 2019. |
| 발행사항 | Ann Arbor: ProQuest Dissertations & Theses, 2019. |
| 형태사항 | 98 p. |
| 기본자료 저록 | Dissertations Abstracts International 81-04A. Dissertation Abstract International |
| ISBN | 9781085796163 |
| 학위논문주기 | Thesis (Ph.D.)--University of California, Davis, 2019. |
| 일반주기 |
Source: Dissertations Abstracts International, Volume: 81-04, Section: A.
Advisor: Hsieh, Fushing. |
| 이용제한사항 | This item must not be sold to any third party vendors. |
| 요약 | In an era of data explosion, how to effectively and efficiently process and extract underlying knowledge from the data becomes the main issue of data science. With the powerful computation resources, we could use machine learning models which are faster and have excellence performance to handle large dataset. Though ML models are handy, they are not suitable for all problems, especially for some combinatorial problems like in the field of neuroscience. To solve this problem, we mainly develop data-driven algorithms to systematically analyze different types of neuroscience data, imaging data and Electroencephalography (EEG) data.In neuroscience, calcium imaging is a two-photon indicator that shows the calcium status to monitor large neuron populations, and further reliably detect action potentials in vivo imaging of biological systems. With the new calcium imaging technique, we can monitor the whole brain-wide activity along the temporal axis. In our work, we focus on the prediction and classification of Epileptic Seizures from Brain-wide Calcium Imaging Video Data. To analyze the calcium imaging video, data-driven computation methodologies are developed to understand the underlying nervous system mechanics that leads to epilepsy. We first build classifiers to diagnose epilepsy, and then predict the happening time of next epileptic seizures. Furthermore, we explore the spatiotemporal structure of the video, and encode it into a global system-state trajectory, then extract a patterned signature that mechanistically defines recurrent epileptic events exhibited by a large complex system. Such a spatiotemporal structural dependency also points out which communities are the main driving forces underlying the recurrent dynamics.We also implement our algorithms to the Electroencephalography (EEG) data. EEG is an electrophysiological monitoring method to record electrical activity of the brain using different electrodes, which are considered as the EEG channels that are placed on scalp. We propose an effective information processing approach to explore the association among EEG channels under different circumstances. Particularly, we design four different experimental scenarios and record the EEG signal under motions of eye-opening and body-movement. With sequences of data collected in time order, we first compute the mutual conditional entropy to measure the association between two electrodes. Using the hierarchical clustering tree and data mechanics algorithm, we could effectively identify the association between particular EEG channels under certain motion scenarios. We also implement the weighted random forest to further classify the classes (experimental scenarios) of the EEG time series. |
| 일반주제명 | Statistics. Neurosciences. Information technology. Information science. |
| 언어 | 영어 |
| 바로가기 | 
: 이 자료의 원문은 한국교육학술정보원에서 제공합니다. |
