Loading...
Please wait, while we are loading the content...
Similar Documents
Detection of neonatal EEG burst-suppression using time-frequency matching pursuit
| Content Provider | Semantic Scholar |
|---|---|
| Author | Awal, Abdul Khlif Dong, Shiying Azemi, Ghasem Colditz, Paul B. Boashash, Boualem |
| Copyright Year | 2015 |
| Abstract | The presence of burst suppression (B-S) in EEG predicts poor neurodevelopmental outcome in newborns. This paper presents a novel method to detect neonatal B-S from multichannel EEG using a time-frequency (TF) matching pursuit (TFMP) approach. This approach utilizes a dictionary of TF atoms for feature selection, resulting in a flexible, spontaneous and physically interpretable set of features. Two TFMP-based features that are able to discriminate between burst and suppression patterns are used in this study, i.e. the number of atoms needed to reconstruct the BS signal and the mean value of the expansion coefficients. For each feature, one way analysis of variance (ANOVA) was performed to show the discriminative capability between these two patterns. Results using EEG signals from 3 term neonates show that both features detect the B-S patterns well. A classifier using these features achieved an accuracy of 89.52%. |
| Starting Page | 6 |
| Ending Page | 6 |
| Page Count | 1 |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.abec.org.au/wp-content/uploads/2015/09/12.00-Awal-abstract-115212.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
