RehaIngest and automatic detection and evaluation of swallowing

Swallowing disorders are common and serious after stroke, brain injury, in ALS and a range of other conditions. For the diagnosis and treatment of swallowing disorders it is obviously important to have a reliable identification and evaluation of swallowing disorders. The imaging procedures used as the current gold standard (fiberoptic endoscopic evaluation of swallowing (FEES) and videofluoroscopy (VFSS)) are insufficiently reliable and are not often repeatable in case of the VFSS. The research literature suggests relatively low figures for interrater reliability of 40–69% for videofluoroscopy and 51% for video endoscopy..

For use in the everyday clinic, non-invasive procedures which would allow objective, reliable swallowing detection would represent a significant advance. This system is now available as RehaIngest created by Hasomed GmbH. This development combines EMG of critical muscles along with Bioimpedance.

Although EMG has been used before in the assessment of swallowing it is inherently unreliable as it is sensitive to a number of artefacts due to, for example, movement of the head. The addition of bioimpedance removes these artefacts and allows the effect of different food consistencies to be studied.

The system uses a reference electrode, two current sources with two current electrodes and two measurement electrodes. The current electrodes apply a sinusoidal current with a frequency of 50 kHz to measure respiratory tract closure. The amplitude of the current source spans the range 3.4–137 micro Amps. The measurement electrodes permit the simultaneous measurement of both bioimpedance and EMG.

The change in bioimpedance correlates with movement of the hyoid bone and larynx during the pharyngeal swallowing phase. While swallowing, the bioimpedance generates a typical curve. It is assumed that the bioimpedance changes depending on the tissue resistance (air, tissue, food). The measurement signal correlates with the elevation of the larynx, which allows the occlusion of the pharynx to be imaged.

The automatic swallowing detection was developed on the basis of standard data of healthy volunteers. There are data for saliva and three consistencies food (liquid, mashed and solid). For recognition accuracy, a value of over 90% could be achieved. Following this “learning process”, swallowing was tested in humans with swallowing disorders. The accuracy in this group was 87%. The great variability of the swallowing process causes the amount of missed or incorrectly recognized swallows. The accuracy of the classifier is designed to improve with an increasing number of swallows.

In 2019 the RehaIngest system was enhanced with the introduction of biofeedback training modules. These visually support the patient’s awareness of their swallowing process and allow performance to be improved. The use biofeedback reinforces improved swallowing performance and taps into the principles of neuroplasticity to encourage recovery.