Use of Empirical Mode Decomposition and Hilbert-Huang Transform in the Analysis of Neuronal Signals
 

Arkansas State University

Abstract

Time-frequency and temporal analyses have been widely used in biomedical signal processing, in the analysis of EMG, EEG, neural signals, etc. These methods represent the temporal characteristics of the signals by their spectral components in the frequency domain. Historically, Fourier spectral analysis has been used for this purpose, but is valid only under extremely general conditions with some crucial restrictions. Imperfections of the probes or numerical schemes compound these complications. The proposed research project is based on the use of a new signal processing tool, namely the Empirical Mode Decomposition (EMD) and the Hilbert-Huang transform (HHT). This is an alternative approach to the analysis of non stationary and non linear signals, and based on the simple assumption that any signal consists of different simple intrinsic mode oscillations. The EMD and HHT techniques have been implemented and tested in the several different medical applications.

The application being considered in this project is the analysis of neuronal signals using EMD and HHT, which will result in the development of new algorithms to identify and decompose the rhythms in the central nervous system. Rhythms of the nervous system have been linked to important behavioral and cognitive states, including attention, memory, object recognition, sensory motor integration, perception, and language processing. Several critical questions can be addressed based on this analysis approach. The analysis will also lead to better understanding of other factors related to sensing and data acquisition, such as acquisition rates needed for optimum data resolution, a priori knowledge of what a "good signal" will look like, frequency content, etc. Validation of the algorithms will be done by applying it to artificially constructed signals to begin with, followed by real-life experimental data containing records of neuronal signals from normal rats. Preliminary analyses’ results and further tuning of the algorithms will help in looking at larger data sets as well as other applications in biology, medicine and related areas.

There is a lot of scope for undergraduate student participation as this project continues through the academic year. This will help ASU to better prepare students for the interdisciplinary research required in these fields today. The researches can continue further in terms of newer applications, further refinement of algorithms, and collaborations with researchers from other Arkansas-based institutions. This study will serve as a basis for providing preliminary data in preparing research grant proposals to agencies such as NIH, and help in obtaining research support through institutional internal research support grants.

Please contact Caroline Miller Robinson regarding questions or comments about this site or our program. For more information about the University of Arkansas for Medical Sciences visit http://www.uams.edu.