Evaluating upper-limb EMG-prosthesis user performance by combining psychometric measures and classification-rates

• Published in: 2013 6th International IEEE/EMBS Conference on Neural Engineering (NER) pp. 359 - 362
• Main Authors: Nan Ge, Goebel, Peter Michael, Amsuess, Sebastian, Paredes, Liliana, Pawlik, Roland, Farina, Dario
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.11.2013
• Subjects: Electromyography; Muscles; Prosthetics; Psychology; Sociology; Tracking; Training
• ISSN: 1948-3546
• DOI: 10.1109/NER.2013.6695946

The robustness of myo-electric prosthesis usage is largely influenced by user performance, where psychological factors (i.e. cognitive-skills, motor-skills and psychological status - such as motivation, will, and stress) play a prominent role. These factors become more important the more degrees of freedom (DOF) a multifunctional prosthesis provides. Despite the large amount of research efforts during the past decades on developing robust control and feedback methods, there has been limited attention on the importance of the aforementioned human factors on the usability of the prosthesis. Psychometric measures are necessary to get a better view on user-ability in prosthesis control. To achieve this aim, the work presented herein applies Item-Response Theory (IRT), which is a psychometric instrument that has been well established for testing human abilities, to introduce a novel score of user performance. This score can be utilized to judge the user's performance at different stages of training, which means measuring improvement or deterioration in movement muscle control according to training activities. As pattern recognition is the control method of choice, classification-rate is taken as second information on the discrimination and repeatability of the recorded movement related EMG patterns. It is used to update the IRT score by taking the joint probability, which combines both measures to determine then the user performance score in a more meaningful way. The score was calibrated on well-trained, able-bodied subjects, who act as a "gold" standard when their movement error was below a certain threshold. Then four amputees with different training experience were selected and it was verified that the score could distinguish between them.