Evaluation of computerized analysis of the propagation of human duodenojejunal contractions

Abstract Detection of peristalsis in the human small intestine has been limited in the past by both the available measurement techniques and the complexity of this activity. Recent developments in ambulant recording have provided a means of monitoring the occurrence of intestinal contractions at multiple sites in the small bowel, but the problem of complexity remains. Using digital data recorded from an intra-luminal strain-gauge transducer in the proximal gut, an algorithm was implemented to identify and classify contractile events within the small bowel. By modelling propagated activity the effect of varying transducer spacing and the number of transducers used was assessed. The question of variability of apparent velocity of peristaltic contractions was examined using successive cross-correlation calculations to extract underlying phase differences between samples of 512 minutes of manometric recording over 150 mm of human small bowel. The effective velocity was found to have a median value of 14 mm sec^-1 and an inter quartile range of 12–18 mm see^-1). It is proposed that, in dynamically tracking variations in phase difference between adjacent recording sites, cross-correlation techniques should be used to control the parameters used for the recognition of propagated contractile events and thereby improve the specificity of this process.