Computer vision‐based diameter maps to study fluoroscopic recordings of small intestinal motility from conscious experimental animals

Abstract

Background: When available, fluoroscopic recordings are a relatively cheap, noninvasive and technically straightforward way to study gastrointestinal motility. Spatiotemporal maps have been used to characterize motility of intestinal preparations in vitro, or in anesthetized animals in vivo. Here, a new automated computerbased method was used to construct spatiotemporal motility maps from fluoroscopic recordings obtained in conscious rats. Methods: Conscious, non-fasted, adult, male Wistar rats (n=8) received intragastric administration of barium contrast, and 1-2 hours later, when several loops of the small intestine were well-defined, a 2 minutes-fluoroscopic recording was obtained. Spatiotemporal diameter maps (Dmaps) were automatically calculated from the recordings. Three recordings were also manually analyzed for comparison. Frequency analysis was performed in order to calculate relevant motility parameters. Key Results: In each conscious rat, a stable recording (17-20 seconds) was analyzed. The Dmaps manually and automatically obtained from the same recording were comparable, but the automated process was faster and provided higher resolution. Two frequenciesof motor activity dominated; lowerfrequency contractions (15.2±0.9 cpm) had an amplitude approximately five times greater than higher frequency events (32.8±0.7 cpm). Conclusions & Inferences: The automated method developed here needed little investigator input, provided high-resolution results with short computing times, and automatically compensated for breathing and other small movements, allowing recordings to be made without anesthesia. Although slow and/or infrequent events could not be detected in the short recording periods analyzed to date (17-20 seconds), this novel system enhances the analysis of in vivo motility in conscious animals