Medical data analysis using microprocessor-based Walsh transformations

The use of Fourier transforms based on orthogonal sinusoidal functions is well established in the analysis of spectral components in noisy medical signals. In this paper the use of Walsh transformations based on orthogonal discrete signals is described for the analysis of ‘slow-wave’ rhythms in the gastro-intestinal tract of mammals. The Walsh transform approach has the advantage that very fast algorithms can be developed since multiplications involving sinusoidal functions are replaced by additions and subtractions. These algorithms are very suitable for simple microprocessors and it is shown that, for determining the frequency of gut rhythms, simple filtering and signal conditioning preceding the microprocessor dispenses with the need for analogue to digital conversion. Hence a very simple digital microprocessing system for the continuous monitoring of gut rhythms has been produced. It has also been extended to provide filtering of signals via forward and inverse Walsh transforms.