|M.Sc Student||Etgar Marcus|
|Subject||Computer Adaptive Testing of Knee Performance|
|Department||Department of Biomedical Engineering||Supervisors||Professor Emeritus Mizrahi Joseph|
|Dr. Avi Wiener|
The goal of this research was to develop an evaluation measuring system of joint torques which is more objective and less dependent of subjects’ cooperation level. This was achieved by combining Computer Adaptive Testing (CAT) methodology with real-time feedback.
The methodology consisted of conducting a series of isometric tests for knee joint performance. The test was divided into two separate parts: Algorithm test and Post-test Repeatability. In both tests, fatigue adjustments were taken into consideration by a linear interpolation, determined from Maximum Voluntary Contraction’s (MVC) measurements, before and after each test.
The results showed that, except in 2 subjects, the algorithm test scoring was higher than the repeatability test scoring. Additionally, only one subject was successful in “remembering” the force he was requested not to exceed. Non-parametric statistical test (wilcoxon) showed that the algorithm test scoring was significantly higher than the repeatability test scoring (P<0.05).
In conclusion, the developed system succeeded to evaluate subjects’ performance more objectively and with less dependency on the subjects’ cooperation level. The study demonstrated the feasibility of using CAT in physical evaluation. The real-time algorithm consisted of 6 tasks in each cycle, where usually the 6th caused the “effective distraction” of the subjects.
Future studies should examine expanding the present method for Isokinetic tests as well.