Processing of brain stem auditory evoked potential for improving diagnosis of multiple sclerosis

Abstract:

Expert doctors use the shape of the principal components of the Brain Stem Auditory Evoked Potential (BSAEP) signal to determine if a person is sick or healthy. This suggests that the wavelet transform of the BSAEP could be used to capture the features that determine if a person is sick or healthy with the help of a neural network. The main obstacle in the application of this approach is the limited number of available signals. To solve this problem, some authors compress the data using different methods, and feed them to a multilayer perceptron; others transforms the signals using different wavelet basis and retain only 8 coefficients; finally a different approach is used by others where radial basis function artificial neural networks (ARBFNN) are involved with a combination of different clustering methods to obtain centers for the internal nodes of the ARBFNN. In all these approaches, the rate of success was always greater for the abnormal signals. In this paper we randomly expanded the training sets and the normal signals have in general a higher success rate.

Año de publicación:

2009

Keywords:

• Health Sciences
• Neural networks
• Multiple Sclerosis
• Radial Basis Functions
• clustering methods
• Signal processing
• wavelets

Fuente:

scopus