This project involves developing and applying desktop and mainframe computer based processing and analysis techniques to signals produced by devices extracting information in physiological contexts (e.g., ECG, EEG) or by laboratory apparatus (e.g., mass spectrometer). Many signal processing algorithms can be implemented via MATLAB on various computer systems, including IBM PCs, MacIntoshes, and the Convex mainframe, with interconnections via the NIH network. For continuous physiological signals a device for converting the analog signals to digital data is required; for other types of (digital) data, additional methods for transferring them in a compatible form to these computer systems will continue to be developed as needed. Major tasks in this project may involve the development of methods to analyze a very large data sets. These methods include data reduction and/or compression, noise suppression using sophisticated filtering algorithms and/or signal averaging, advanced techniques for pattern recognition, and statistically or mathematically based feature extraction, trend analysis, and construction of spectra where appropriate.
