BSMART: a MATLAB/C toolbox for analysis of multi-channel neural data

Author: Lei Xu, M.S.

Primary Advisor: Hualou Liang, PhD (co-author)

Committee Members: Jiajie Zhang, PhD

Masters thesis, The University of Texas School of Health Information Sciences at Houston.

We have developed a free open source toolbox called Brain-SMART (System for Multivariate AutoRegressive Time series) for the analysis of multi-channel neural signals. The BSMART running under the common data analysis Matlab environment (The Mathworks, Inc.) is platform independent. Available functions include data visualization, data preprocessing (including trend removal and data standardization), Adaptive Multivariate AutoRegressive (AMAR) analysis, model validation and development of a suite of spectral quantities including spectral power, pair coherence and Granger causality. Graphic user interface (GUI) allows users to do data processing without the need of Matlab syntax. Experienced MATLAB users could take advantage of the MATLAB command line and write their own batch scripts for performing data analysis. The BSMART toolbox together with sample data and user tutorial is freely available under the GNU public license.