Functional rehabilitation uses neural prostheses to replace motor, sensory or autonomous functions of the body. The necessary coupling of technical systems with the nervous system is mediated by interfaces that have to fulfill specifications with respect to recording of sensory and motor signals from downstream (efferent) or upstream (afferent) information, respectively. This work focuses on prostheses for upper extremities and contains three parts: Firstly, the state of the art will be described. So far, quite simple open-loop control systems with limited grasp functionality dominate the field of artificial limbs as well as upper limb neural prostheses. Secondly, results from animal models are taken to introduce and discuss possibilities and limitations of data acquisition and intelligent data analysis of afferent signals after mechanical stimulation of the skin. In the third part, perspectives of neural prostheses are discussed from the viewpoint of control theory. We recommend closed-loop control concepts and the use of natural, adequate signals for patient feedback to improve prostheses´ performance.
Print ISSN: 0178-2312
Volume: 54, 11/2006
Pages: 523 - 536