Estimación de potenciales evocados auditivos del tronco cerebral mediante descomposición modal empírica

Abstract

The brainstem auditory evoked potentials (BAEPs) are commonly used for diagnostic purposes; however, processing becomes difficult because they are immersed in a large amount of noise coming not only from external sources but also from other physiological sources. So far the most widely used and accepted method to obtain reliable recording is the coherent averaging, but this type of processing presents some drawbacks. The empirical mode decomposition (EMD) is a relatively new technique which is used for processing of non-stationary signal like almost physiological signals. This method separates a signal, extracting the energy associated with various intrinsic time scales, into a finite set of oscillatory modes. The purpose of the study was to asssess the EMD as a tool for improving the performance of the averaging coherent BAEPs seeking to reduce the amount of epochs needed to obtain a reliable register.To this end, we have analyzed the reconstruction of eight registers using only modes 2, 3 and 4 resulting from the EMD, the studies found that a reconstruction of 800 epochs is acceptable.