Title :
Spectral analysis of visual evoked potentials
Author :
Dobrowolski, Andrzej P. ; Okon, Marta
Author_Institution :
Fac. of Electron., Mil. Univ. of Technol., Warsaw, Poland
Abstract :
The paper presents a conception of classification method of Visual Evoked Potentials (VEP) to physiological or pathological case based on power spectral parameters. The authors have verified their concept through a series of numerical experiments performed using a dedicated application. As a result of experiments, the final method provided only 4 cases of wrong classification among training data (6%) and 13 among the testing data (43%).
Keywords :
medical signal processing; principal component analysis; signal classification; spectral analysis; support vector machines; visual evoked potentials; VEP; classification method; dedicated application; pathological case; physiological case; power spectral parameters; spectral analysis; testing data; training data; visual evoked potentials; Electric potential; Gravity; Pathology; Principal component analysis; Spectral analysis; Support vector machines; Visualization; Principal Component Analysis; Support Vector Machine; biomedical signal processing; spectral analysis; visual evoked potential;
Conference_Titel :
Signal Processing Symposium (SPSympo), 2015
Conference_Location :
Debe
DOI :
10.1109/SPS.2015.7168275
