Ideas on sensory feedback in hand prostheses
P. Herberts *
L. Körner *
Abstract
Development of systems for sensory feedback in hand prostheses has not been as successful as that of modern prosthesis control systems. The discrepancy is partly caused by an insufficient analysis of the concept of sensory feedback and by neglect of knowledge on the physiology of kinesthesis. In the present paper modem theories on physiologic kinesthesis are briefly summarized and the implication of these theories on the development of prosthesis sensory feedback systems are discussed. It is concluded that the future development of sensory feedback systems for hand prostheses should be directed towards increased utilization of the physiologic kinesthesis resulting from operation of the prosthesis control systems. This can be obtained by further development of the control systems. One promising approach in this direction is the use of a proportional control signal based on signal acquisition through pattern recognition of multiple myoelectric signals. Development of artificial systems for feedback should be restricted to situations when feedback emerging from the prosthesis control is insufficient. The importance of simplicity and reliability of feedback systems is stressed as well as the necessity to maintain prosthesis self-containment even after application of a feedback system.
Introduction
The development of control systems for motorized prostheses based on detection of myoelectric signals has been rapid and successful following the first use of such systems in the late nineteen fifties (Battye el al., 1955). Clinical success with powered prostheses has been reported (Schmidl, 1973; Lewis et al., 1975; Herberts et al., 1978), but other authors have expressed doubts about the value of these devices (Mooney, 1976). Beyond doubt, many patients using myoelectrically controlled prostheses have been pleased with them, despite several limitations in the prosthetic function (Herberts et al., 1979). A generally recognized drawback in powered prostheses in comparison to conventional cable-operated devices is that the lack of feedback makes control outside the field of vision very difficult. In addition, the control of strength in the grip is insufficient. These findings have led to attempts to develop artificial feedback systems for use in myoelectric prostheses (Kato, 1970; Mann, 1973; Clippinger et al., 1974; Prior & Lyman, 1975; Rohland, 1975; Shannon, 1979). Most feedback systems described up to now have not reached a development stage allowing routine clinical use outside the laboratory. Clinical follow-up has been reported in only small series. Measurements of the performance of amputees using powered prostheses with feedback show that it is close to their performance with conventional cable-operated grips (Mann & Reimers, 1970). The purpose of this paper is to discuss some basic clinical and physiological principles relevant to feedback in externally energized prostheses. Possible fields for future research in this area are indicated.
The concept of sensory feedback
A hand prosthesis is mainly a machine to replace the prehension of the lost hand. The loss of a hand, however, also implies the loss of an important sensory function. The full sensation of a normal hand, as pointed out by Moberg (1964), is a very complex quality which does not lend itself easily to replacement by artificial devices. Sensory feedback is not intended to replace normal hand sensation. We think that sensory feedback in prosthetics should have the same meaning as feedback in control engineering; i.e. a way to compare the output of a machine with the input (Fig. 1 ). Feedback is an aid to increase the accuracy of the control system and of prehension, Therefore, the term "artificial touch" is inadequate and may lead development of prosthetic feedback systems in wrong directions.
In order to effectively enhance control of the prosthesis the sensory feedback should consist of several components. One component is kinesthetic information about position, movement and force in the prosthesis joints, i.e. proprioceptive feedback. Another component is information about the effects of the action of the prosthesis on the outside world; i.e. somatosensory feedback concerned with, for example, force in the grip and detection of slippage of handled objects. Most feedback systems developed today convey somatosensory information only. The somatosensory part of feedback information can not replace proprioceptive information and vice versa. The special importance of feedback concerning handling of objects in the grip has been pointed out by Forchheimer et al., (1978).
Existing feedback systems
Feedback systems for externally powered prostheses have been used since the early nineteen sixties (Tomovic & Boni, 1962). Most systems have employed some kind of artificial stimulation controlled by a transducer situated in the grip of the prosthesis. The block diagram of Fig. 2 illustrates some different solutions of the feedback problem as described in the literature. The most commonly used artificial feedback systems have worked with vibratory stimuli (Alles, 1970; Mann, 1973) and electrical stimulation (Kato, 1970; Clippinger et al., 1974; Reswick et al., 1975; Rohland, 1975; Anani et al., 1977; Shannon, 1979). The clinical application of artificial sensory feedback systems has often been unsuccessful (Reswick et al., 1975; Mooney, 1976; Reswick & Nickel, 1977). Rejection of systems has been caused by technical problems such as fragility, interference with the control systems and lack of miniaturization. It is obvious that in some applications the need for a feedback system has not been present, which of course, has led to rejection.
Kinesthetic mechanisms and prostheses feedback
Aids for the handicapped designed to take advantage of physiologic mechanisms will have a greater chance of being accepted by patients than devices based on entirely artificial grounds (Herman, 1973; Hirsch & Klasson, 1974). This is especially true of the complicated myoelectric prostheses that include a feedback system. In the design of such devices thorough knowledge of kinesthetic mechanisms in man is mandatory for success. Mann (1973) and Simpson (1974) described feedback for prostheses using the physiologic signals that result from the actions of the human body necessary to control the prostheses. Clinically this has resulted in the most attractive and successful systems so far described. Practical experience with such devices shows the need to take physiologic kinesthetic mechanisms more into consideration in the future development of feedback systems.
Physiologic kinesthesia and forearm amputation The knowledge of neural mechanisms behind the sensing of muscular effort and the sensing of position has been revised during the nineteen seventies. Previously, it was widely believed that the sense of position was based entirely upon information from joint and skin afferents (Rose & Mountcastle, 1959). Indirect evidence is now against this view. Investigation of cat knee joint receptors has not revealed the presence of any receptors capable of signalling absolute joint angles (Clark & Burgess, 1975). Sense of position is not affected by total joint replacement (Grigg et al., 1973). In addition, it was shown by Goodwin et al., (1972) that artificial stimulation of muscle spindles induces illusions of movements. This indirect evidence opposing the traditional views on kinesthesia led to a re-evaluation of the classical experiments concerning the sensing of movement and position when finger movements were performed with blocked skin and joint sensors. Through such experiments it has been convincingly shown by Gandevia & McCloskey (1977) that the sense of position has its neural mechanisms partly in muscle spindles. By performing weight-matching tests, McCloskey & Gandevia (1978) have shown that the estimation of heaviness and the sense of effort have their neural mechanisms largely in the central nervous system.
Substantial parts of the tendons and muscles executing movements of the normal hand and fingers are left intact after a hand amputation. The new evidence presented above leaves no doubt that these remaining structures contain receptors responsible for important components of the sensing of position and movement in the normal hand. The information from these receptors is also available to the amputee as is the sensing of effort.
These physiologic facts should be taken into consideration in the design of prosthesis feedback systems.
Sensing force and effort
The force proprioception system developed by Mann (1974) for the Boston arm works with the sensing of effort accompanying the muscle work at the control site necessary for generating myoelectric signals. A negative feedback signal proportional to the force resisting the movement is added to the myoelectric signal that controls the prosthesis. When the prosthesis is loaded an increased myoelectric signal is required to achieve movement. The result is an augmentation of the sensing of effort of the muscles at the control site. Such an augmentation is necessary to make the force clearly perceivable. Since the sensing of effort has its neural mechanisms in the central nervous system, it can be assumed to be influenced by signals originating from different muscles relevant to one specific movement (Herman, 1973). Such a convergence can explain why previous attempts to utilize the sensing of effort for prosthesis feedback purposes without strong augmentation have been disappointing. In these attempts subjective sensations resulting from activity in single muscles only have been applied. Single muscle activity is an unphysiologic phenomenon and probably lacks cortical representation (Radonjic & Long, 1970). Therefore, if the control of prostheses can be related to physiologic movements rather than to actions in single muscles, more feedback information can be expected to emerge from the control system.
Hand prostheses controlled by pattern recognition of multiple myoelectric signals have been described by, among others, Herberts et al., (1973). The pattern recognition approach permits control of the prosthesis through and integration of signals from several muscles relevant to one specific movement. It is therefore plausible to assume that a proportional control signal derived from signal processing according to the pattern recognition method should yield significant amounts of information about effort and force. Our preliminary data support this hypothesis.
Sense of position
Muscle afferents have been shown to play a significant part in the perception of movements (Mattews, 1977). In a clinical follow-up study of unilateral below-elbow amputees using myoelectric prostheses we found that all non-congenital amputees (35 patients) had a distinct phantom image (Table 1 ). All the patients except two stated that they could easily move the perceived phantom and that they could feel how the phantom was moved. The neural basis for the perception of movement of the phantom image is considered to lie in the muscle afferents from the distal part of the amputation stump (Henderson & Smyth, 1948). Control of prostheses using the pattern recognition method is based on the principle that specified movements of the phantom hand shall result in corresponding movements of the prosthesis. Our preliminary data show that it is to a certain extent possible to relate a proportional control signal achieved through the pattern recognition method to specified positions of the phantom hand. Therefore it seems reasonable to postulate that the sense of movement and even the sense of position of the phantom hand can be used to convey prosthesis feedback. Further development of the prosthesis control systems for feedback purposes is, however, necessary if this goal is to be achieved.
Extended physiological proprioception
A different way to utilize the physiologic actions at the prosthesis control site was described by Simpson (1974). In prostheses designed for amelic children the movements and the positions of the clavicle are translated to movements and positions of the prosthesis. Through this proportionality between the movements of the clavicle and the prosthesis the angle of the prosthesis in space can be determined by the normal kinesthetic mechanisms of the child. The brain will then very easily adapt to the length of the terminal segment (the prosthesis) in much the same way as the golfer will adapt to the length of his club. The phenomenon is called extended physiological proprioception (EPP) and gives the amputees significant amounts of feedback information in addition to excellent efferent control.
Areas for future research
Future approaches to feedback in hand prostheses should be directed towards the design of systems which can be included in self-contained prostheses. The importance of self-containment for patient acceptance is clearly documented (Childress, 1973). The experiences of Mann and Simpson indicate that feedback systems based on the physiologic actions necessary to control the prosthesis are consistent with high patient acceptance and with prosthesis self-containment. The new evidence on neural mechanisms underlying kinesthesia gives further support to the opinion that development of control systems providing feedback information as well will lead to success. One promising approach in this line of research is the use of pattern recognition of multiple myoelectric signals to generate a proportional control signal.
In accordance with the discussion above, a proportional control signal derived from several muscles with relevance to one specific movement should yield substantial amounts of force proprioception in addition to sensing movement and possibly also sensing position.
Even if important feedback information can be achieved from a properly designed control system, it is necessary to leave room for systems working with purely artificial stimulation as well. Such stimulation is needed to convey the somatosensory components of kinesthesis. In the design of purely artificial sensory systems it is equally necessary not to endanger self-containment and reliability of the whole prosthesis system. From this point of view electrical stimulation seems to have advantages over mechanical or auditory stimulation. Electrical stimulators can readily be miniaturized, their energy consumption is low, and they are reliable. If the electrical stimulation is applied to the nerves of the amputation stump the sensations will be felt by the amputee in specific parts of the phantom image (Anani et al., 1979). This creates a convergence of feedback and control functions to the phantom image. Such convergence can be expected to increase the accuracy of prosthesis control (Weissenberger & Sherridan, 1962).
Acceptance by the patient is the only important criterion to determine if an effort to replace a lost function with an artificial device is successful. Patient acceptance is a complicated concept which is not determined only by technical and cosmetic characteristics of the rehabilitation aids. Social and economic factors are equally important in addition to the psychologic attitude of the amputee towards his handicap (Hook, 1976). This means that complicated rehabilitation aids can never be prescribed without a thorough analysis of the psychologic and socio-economic situation of each individual patient. However, in order to meet the requirements of most patients, rehabilitation aids must be reliable, easy to use, and inconspicuous. Therefore, the aim in designing prosthesis feedback systems must be as much to maintain prosthesis self-containment and self-suspension as to provide significant amounts of feedback information.
References:
Alles, D. S. (1970). Information transmission by phantom sensations. IEEE Transactions on Man-machine Systems, MMS-11, 85-91.
Anani, A. B., Ikeda, K., & Körner, L. (1977). Human ability to discriminate various parameters in afferent electrical nerve stimulation with particular reference to prostheses sensory feedback. Med. & Bio. Eng. & Comp., 15, 363-373.
Anani, A. & Körner, L. (1979). Discrimination of phantom hand sensations elicited by afferent electrical nerve stimulation in below-elbow amputees. Medical Progress Through Technology, (in press).
Battye, C. K., Nightingale, A. & Willis, J. (1955). The use of myoelectric currents in the operation of prostheses. J. Bone & Jt. Surg. 37-B, 506-510.
Childress, D. S. (1973). Powered limb prostheses: Their clinical significance. IEEE Transactions on Biomedical Engineering, BME-20, 200-207.
Clark, F. J. & Burgess, P. R. (1975). Slowly adapting receptors in cat knee joint: Can they signal joint angle? J. Neurophys. 38,1448-1463.
Clippinger, F. W., Avery, R. & Titus, B. (1974). A sensory feedback system for an upper limb amputation prosthesis. Bull. Pros. Res. 10-22, 247-258.
Forchheimer, B., Klasson, B., Markström, G. & Rydell, N. (1978). Experiences from clinical evaluation of upper extremity prostheses: Proc. 6th Int. Symp. on External Control of Human Extremities. ETAN, 11-24, Dubrovnik.
Gandevia, S. C. & McCloskey, D.I. (1977). Sensations of heaviness. Brain, 100,345-354.
Goodwin, G. M,, McCloskey, D. I. & Matthews, P. B. C. (1972). The contribution of muscle afférents to kinesthesia shown by vibration induced illusions of movement and by the effects of paralysing joint afferents. Brain, 95, 705-748.
Grigg, P., Finerman, G. A. & Riley, L. H.(1973). Joint position sense after total hip replacement. J. Bone & It. Surg., 55-A, 1016-1025.
Herberts, P., Almström, C., Kadefors, R. & Law-ence, P. D. (1973). Hand prosthesis control via myoelectric patterns. Acta orth. Scand.. 44, 389-409.
Herberts, P., Almström, C. & Caine, K. (1978). Clinical application study of multifunctional prosthetic hands. J. Bone & Jt. Surg., 60-B, 552-560.
Herberts, P., Körner, L., Cainé, K. & Wensby, L. (1979). Acceptans av enfunktionella myoelektriska proteser hos en grupp unilateralt under-armsamputerade patienter. Manuscript in Swedish.
Herman, R. (1973). Augmented sensory feedback in the control of limb movement. Neural organization and its relevance to prosthetics. Ed. Fields, W. S. Intercontinental Medical Book Corporation, New York and London. 197-215.
Hirsch, C. & Klasson, B. (1974). Clinical aims and desires for to-day's arm prosthesis: The control of upper-extremity prostheses and orthoses. Ed. Herberts, P., Kadefors, R., Magnusson, R. and Petersen, I. Thomas, Springfield, Illinois. 58-62.
Höök, O. (1976). Neuropsychological aspects of motivation. Scand. J. Rehab. Med., 8,97-106.
Kato, I., Yamakawa, S., Ichikawa, K. & Sano, M. (1970). Multifunctional myoelectric hand prosthesis with pressure sensory feedback system. Waseda hand 4P. Proc. 3rd Int. Symp. on External Control of Human Extremities, ETAN, 155-170, Dubrovnik.
Lewis, E. A., Sheredos, C. R. & Sowell, T. T. (1975). Clinical application study of externally powered upper-limb prosthetic systems: The VA elbow, the VA hand, and the VA/NU myoelectric hand systems. Bull. Pros. Res., 10-24,51-136.
Mann,R. W. (1973). Prostheses control and feedback via noninvasive skin and invasive peripheral nerve techniques. Neural organization and its relevance to prosthetics. Ed. Fields, W. S. Intercontinental Medical Books Corp., New York and London. 177-195
Mann, R. W. (1974). Force and position proprioception for prostheses. The control of upper-extremity prostheses and orthoses. Ed. Herberts, P., Kadefors, R., Magnusson, R. and Petersen, I., Thomas, Springfield, III. 201-219.
Mann, R. W. & Reimers, S. D. (1970). Kinesthetic sensing for the EMG controlled Boston Arm. IEEE Transactions on Man-machine Systems, MMS-11, 110-115.
Matthews, P. B. C (1977.). Muscle afferents and kinaesthesia. Brit. Med. Bull. 33, 133-142.
McCloskey. D. I. & Gandevia, S. C. (1978). Role of inputs from skin, joints and muscles and of corollary discharges in human discriminatory tasks. Active Touch. The mechanism of recognition of objects by manipulation. Ed. Gordon, G., Pergamon Press. 177-187.
Moberg, E. (1964) Aspects of sensation in reconstructive surgery of the upper extremity. J. Bone & Jt. Surg., 46-A, 817-825.
Mooney, V. (1976). Sensory feedback in upper-extremity amputees. Clin. Orthop., 119,274-275.
Prior, R. E. & Lyman, J. (1975). Electrocutaneous feedback for artificial limbs. Bull. Pros. Res., 10-24, 3-37.
RadonjiC, D. & Long, C. (1970). Why myoelectric control is so difficult. Proc. 3rd Int. Symp. on External Control of Human Extremities. ETAN, 59-67, Dubrovnik.
Reswick, J., Mooney, V., Schwartz, A., McNeal, D., Su, N., Bekey, G., Bowman, B., Snelson, R., Irons, G., Schmid, P. & Sperry, C. (1975). Sensory feedback prosthesis using intraneural electrodes. Proc. 5th Int. Symp. on External Control of Human Extremities. ETAN, 9-24, Dubrovnik.
Reswick, J. B. & Nickel, V. L. (1977). Annual report of progress, Rehabilitation engineering centre at Rancho Los Amigos Hospital, Los Angeles, Dec. 1975-Jan. 1977.
Rohland, T. A. (1975). Sensory feedback for powered limb prostheses. Med. & Bio. Eng., 12, 300-301.
Rose, J. E. & Mountcastle, V. B. (1959). Handbook of Physiology, Section I: Neurophysiology. Amer. Phys. Society, Washington DC. Vol. 1,387-429.
Schmidl, H. (1973). The INAIL-CECA prostheses. Orth, and Pros., 27, 6-13.
Shannon, G. F. (1979). A myoelectrically-controlled prosthesis with sensory feedback. Med. & Bio. Eng. & Comp. 17, 73-80.
Simpson, D. C. (1974). The choice of control system for the multimovement prosthesis: extended physiological proprioception (epp). The control of upper-extremity prostheses and orthoses. Ed. Herberts, P., Kadefors, R., Magnusson, R. & Petersen, I. Thomas, Springfield, III. 146-150
Tomovic', R. & Boni, G. (1962). An adaptive artificial hand. IRE Transactions on Automatic Control. AC-7:3, 3-10.
Weissenberger, S. & Sherridan, T. B. (1962). Dynamics of human operator control systems using tactile feedback. J. Basic Eng., 84,297-301.
|
ISPO Home
