Obukhov K. Yu, Obukhov Yu. V., Maliuta I.A., Nikitov S. A., Sushkova O.S., Kershner I.A. Classification models application to quantitative diagnostics of early stage Parkinson’s disease. In: 11th German-Russian Conference on Biomedical Engineering (RGC 2015), 17–19 июнь 2015, Ахен, Германия , С. 127-130.
|
Другой (pdf)
25 Aachen German-Russian-Conference 2015.pdf - Опубликованная версия Загрузить (1MB) | Предварительный просмотр |
Аннотация
New approach to quantitative diagnostics of early stage Parkinson's disease was investigated. There are three main differences of EEG wavelet scalograms, which were considered as early stage PD features. The first is a disordering ridge of PD patient scalogram in frequency range more then ~6 Hz in comparison with that of normal volunteer. The second is a more powerful PD patient cortex electrical activity in frequency range ~4-6 Hz. And the third feature is a scalograms asymmetry of the left and right brain semi spheres. The logistic regression model is built with binary target variable – the disease factor. The importance of every feature was analyzed. The learning dataset consists of 5 major features, which were extracted from the EEG data: the maximum of amplitude of theta rhythm divided on the amplitude of alpha rhythm for C3 and C4 EEG electrodes, the amplitude of tremor for left hand divided on the amplitude of tremor for right hand (or otherwise), the correlation values calculated as in equation and the standard deviation of these correlations. There are 53 observations: 33 patients and 20 normal people.
| Тип объекта: | Доклад на конференции или семинаре (Доклад) |
|---|---|
| Авторы на русском. ОБЯЗАТЕЛЬНО ДЛЯ АНГЛОЯЗЫЧНЫХ ПУБЛИКАЦИЙ!: | Обухов К.Ю. Обухов Ю.В. Малюта И.А. Никитов С.А. Сушкова О.С. Кершнер И.А. |
| Подразделения (можно выбрать несколько, удерживая Ctrl): | 144 лаб. методов применения ЭВМ в научных исследованиях |
| URI: | http://cplire.ru:8080/id/eprint/2903 |
 |
Изменить объект |
