Computational protocol: The Effect of Bolus Volume on Hyoid Kinematics in Healthy Swallowing

Similar protocols

Protocol publication

[…] In this paper, we report a retrospective analysis of data from a previously reported study [] of swallowing in 20 healthy young participants (10 male, 10 female; mean age 31.5 years, SD 5.7 years). Participants were recruited to represent the normal height distribution in the population. The data for the current analysis comprised 9 boluses of 22% w/v ultrathin liquid barium per participant, organized in blocks of 3 boluses at each of 3 target volumes (5, 10, 20 mL). Boluses were self-administered from 30 mL volume medicine cups and the order of bolus volume blocks was randomized. Precise methods for ensuring volumetric control and measuring sip volume have been described elsewhere []; 95% confidence intervals for actual sip volume were 3.4–3.7 mL for the 5 mL condition, 7.8–8.3 mL for the 10 mL condition, and 16.8–17.9 mL for the 20 mL condition. For simplicity, the bolus volumes will be referred to as 5, 10, and 20 mL throughout this paper. Swallows were captured in lateral view using a Toshiba Ultimax Fluoroscope (Toshiba America Medical Systems, Inc., Tustin, CA) at 30 pulses per second and recorded at 30 frames per second. Data processing involved splicing of the videofluoroscopy recordings into clips capturing a time interval beginning 30 frames prior to the first bolus of each block passing the mandibular ramus until 30 frames after the hyoid returned to rest after the 3rd bolus in the block. Data analysis involved frame-by-frame tracking of hyoid position in ImageJ freeware, using a coordinate system with an origin at the anterior-inferior corner of the C4-vertebrae and vertical axis defined by a line running from the origin upwards through the anterior-inferior corner of the C2-vertebrae (see ). The distance between the anterior-inferior corners of the C2 and C4 vertebrae also served as an anatomical scalar to enable us to control for differences in the size of the pharynx across participants during measurements of hyoid movement distance []. Using the frame-by-frame position histories, an algorithm in Excel VBA software was used to index the onset and end of the anterosuperior hyoid burst movement for each swallow. The onset of the hyoid burst movement was defined as the lowest position between the start of marking (10 frames before observed hyoid movement) and the peak hyoid position, calculated in both X and Y directions. The end of hyoid burst movement was defined as the peak of maximal hyoid position within the hyoid position history, calculated in both X and Y directions. This permitted the derivation of measures of maximum hyoid displacement (in anatomically normalized units, i.e., % of the C2–4 vertebral distance), distance travelled (maximum displacement minus onset position, in % C2–4 units), burst movement duration (ms), and velocity or speed (i.e., distance/duration in mm/s) for the X, Y, and XY movement directions. The frame (% of burst movement duration) and value (mm/s) of peak velocity or speed were also identified for each movement direction. Finally, the first frame showing laryngeal vestibule closure, defined as a seal between the laryngeal surface of the epiglottis and the arytenoids, was tracked for each swallow and its temporal location was calculated as a % of the duration of the hyoid burst movement (e.g., a value of 50% would indicate that the LVC occurred half-way through the hyoid burst movement, while measures of 30% and 70% would reflect earlier and later laryngeal vestibule closure, resp.). Strong inter- and intrarater reliability was obtained for all measurements based on repeat rating of a random selection of 10% of the recordings, as reported elsewhere [, ]. Statistical analyses were performed using IBM SPSS Statistics version 21. A mixed model analysis of variance (ANOVA) with a within-participant repeated factor of trial within bolus volume was performed to identify the impact of bolus volume on the study parameters. The correspondence between the timing of peak velocity and LVC, both expressed as % of the duration of the hyoid burst movement, was determined using scatter plots and correlation analysis. […]

Pipeline specifications

Software tools ImageJ, SPSS
Applications Miscellaneous, Microscopic phenotype analysis
Organisms Homo sapiens
Diseases Movement Disorders