A critical review of EMG-controlled electrical stimulation in paraplegics.

This review presents a description and provides a comparative performance evaluation of EMG control vs. other approaches to controlling functional electrical stimulation (FES) in upper-motor-neuron paraplegics to provide them with a certain degree of walking ability with walker support. EMG control is considered in terms of a combination of above-lesion EMG control and below-lesion response-EMG control. The above-lesion EMG is to control the activation of different limb functions involved in standing up and walking via FES. This control is accomplished by analyzing raw surface EMG time-series patterns to discriminate between upper-trunk muscle contraction patterns which, in turn, are correlated with intended lower-limb functions involved in walking, so that natural and instinctive balance changes of paraplegics are reflected and controlled by the patient only above the lesion. The below-lesion response-EMG is the EMG produced in response to the FES pulses at the stimulation sites, for adjusting stimulation levels as needed when contractions weaken due to muscle fatigue. Above-lesion EMG is a stochastic (random-like) signal, being a response to unsynchronized motor-neuron firings, whereas the below-lesion EMG is a deterministic signal responding to synchronized firings that result solely from the FES pulses. The present review discusses the merits and difficulties of EMG control and attempts to give a critical evaluation of patient performance under such control, contrasted to other methods of control, noting that FES-activated walking without adequate and patient-responsive control is of very limited attraction and use to paraplegics. Of the various control methods, only foot- or hand-switch control comes close in overall performance to above-lesion EMG control. Although one cannot categorically prefer EMG control to foot- or hand-switch control performance-wise, psychologically, EMG control has the advantage in that concentration is not diverted to the patient's fingers, whereas foot switches are inadequate for patients lacking reasonable natural (non-FES) pelvic thrust control. However, for stimulation-level control, below-lesion response-EMG appears to be the clear answer. It is shown that below-lesion EMG control of stimulation levels can be used independent of above-lesion EMG control. Below-lesion EMG control can thus be combined also with hand-switch control or, in case of low level upper-motor-neuron lesions, with pelvis control. Hence, and since FES-activated walking is as good as its control, we conclude that EMG control should be given serious consideration in any FES walking.(ABSTRACT TRUNCATED AT 400 WORDS)