A radial basis function classifier for pediatric aspiration detection.

Silent aspiration presents a serious health issue for children with dysphagia. To date, there is no satisfactory means of detecting aspiration in the home or community. In an effort to design a practical device that could offer reliability, non-invasiveness, portability, and easy usability, radial basis functions based on cervical accelerometry signals were investigated. Vibration signals associated with safe swallows and aspirations, both identified via videofluoroscopy, were collected from over 100 children with neurologically-based dysphagia using a single-axis accelerometer. Three time-domain discriminatory mathematical features were extracted from the accelerometry signals. An exhaustive set of all possible combinations of the features was investigated in the design of radial basis function classifiers. The feature pairing of dispersion ratio and normality achieved an accuracy of 81.03 +/- 5.78%, a false negative rate of 9.06 +/- 4.84%, and a false positive rate of 9.91 +/- 5.03% for aspiration detection. The proposed classifier can be easily implemented in a hand-held device.