Sybeldon, Matthew
(2017)
Addressing Nonstationarity in EEG Applications.
Master's Thesis, University of Pittsburgh.
(Unpublished)
Abstract
Electroencephalography (EEG) is a measure of electrical activity from the brain that is used for numerous applications. EEG can be used for detecting the onset of epilepsy in a patient. It can also allow a disabled person to operate a computer using a brain computer interface
(BCI). However, EEG measurements are usually interpreted as being generated from a random process. This random process is not stationary, meaning that the distribution governing
the measurements can change over time. In seizure detection, this nonstationarity may indicate a seizure. In BCI applications, nonstationarity is problematic due to the invalidation of previously trained classifiers. This thesis provides techniques to both leverage and mitigate nonstationarity depending on the application. For seizure detection, a statistical detector requiring no previously labelled data or expert knowledge is outlined to detect seizures in the EEG. This detector achieved an average seizure prediction time of 70 seconds. In BCI applications, a combination of mutual information and ensemble learning is used to identify previously learned data most similar to incoming data and reduce calibration requirements. It was shown that this learning scheme provided adequate results for typical participants and outperformed state of the art techniques for steady state visual evoked potential BCIs
for participants whose EEG violated typical assumptions.
Share
| Citation/Export: |
|
| Social Networking: |
|
Details
| Item Type: |
University of Pittsburgh ETD
|
| Status: |
Unpublished |
| Creators/Authors: |
|
| ETD Committee: |
|
| Date: |
1 February 2017 |
| Date Type: |
Publication |
| Defense Date: |
8 November 2016 |
| Approval Date: |
1 February 2017 |
| Submission Date: |
11 November 2016 |
| Access Restriction: |
5 year -- Restrict access to University of Pittsburgh for a period of 5 years. |
| Number of Pages: |
43 |
| Institution: |
University of Pittsburgh |
| Schools and Programs: |
Swanson School of Engineering > Electrical Engineering |
| Degree: |
MSEE - Master of Science in Electrical Engineering |
| Thesis Type: |
Master's Thesis |
| Refereed: |
Yes |
| Uncontrolled Keywords: |
Electoencephalography, Nonstationarity, Seizure Detection, Brain Computer Interface, Transfer Learning |
| Date Deposited: |
01 Feb 2017 17:06 |
| Last Modified: |
01 Feb 2022 06:15 |
| URI: |
http://d-scholarship.pitt.edu/id/eprint/30292 |
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
 |
View Item |
![[feed]](/images/feed-icon-32x32.png){kind=icon}