Graduation Year


Document Type




Degree Granting Department


Major Professor

William E. Lee III, Ph.D.

Co-Major Professor

Peter J. Fabri, M.D.

Committee Member

Robert D. Frisina Jr., Ph.D.

Committee Member

Piyush Koria, Ph.D.

Committee Member

Michael A. Harrington, M.D.


Decubitus Ulcer, Bedsores, Spinal Cord Injury, Ischial Tuberosities


The three goals of this research were to investigate how normal subjects move while seated, how paraplegic patients move while seated, and whether seated movements can be modeled using a hemi-ellipsoid shape. Pressure readings were recorded at 11 Hz using a 36 by 36 sensor pressure map by XSENSOR. Subjects were instructed to move or perform pressure relief as they normally would while seated. Analysis was performed using Microsoft Excel with Solver and Matrix.xla add-ins and automated with VBA code. Major movements and time intervals between movements were calculated by locating the area of maximum pressure on each hemi-buttock for 20 normal and 6 paraplegic subjects. Statistical analysis revealed movements followed a normal distribution while time intervals followed a lognormal distribution. For both the normal (p=0.041) and paraplegic groups (p=0.007) the number of movements significantly increased from the first hour of recording to the second hour. The time interval between major movements decreased but not significantly for neither the normal subjects nor the paraplegics. No significant differences were identified between the normal and paraplegic groups over the first hour or second hour for number of movements or time intervals. Time series analysis with plotting, trend lines, ARIMA, and periodograms did not reveal patterns in the data. Preference for a side was shown. Next, all areas of identified major movements for one subject and one frame for each of the paraplegic patients were modeled as a hemi-ellispoid shape using minimization with Solver. Eigenvalues were calculated in order to obtain the lengths of the x, y, and z axis of the hemi-ellipsoid with an average error of 39.87% for the normal subject and an error range of 5.10% to 2701.81% for the paraplegic patients.