An electrode configuration technique using an electrode ... - CiteSeerX

Medical Engineering & Physics 28 (2006) 166–176

Technical note

An electrode configuration technique using an electrode matrix arrangement for FES-based upper arm rehabilitation systems S.B. O’Dwyer a , D.T. O’Keeffe a,∗ , Susan Coote b , G.M. Lyons a a

Biomedical Electronics Laboratory, Department of Electronic and Computer Engineering, University of Limerick, Ireland b Department of Physiotherapy, University of Limerick, Ireland Received 16 June 2004; received in revised form 1 February 2005; accepted 22 March 2005

Abstract An upper limb electrical stimulation technique has been developed which features a novel self-configuration approach, to obtain an ideal wrist response from the patient. The system uses an analogue de-multiplexer in conjunction with an electrode matrix so that different electrode sites can be tested using only one channel of stimulation. A twin axis goniometer is attached to the patient’s wrist and flex sensors are attached to the patient’s fingers so that the control algorithm can assess the wrist response. A data acquisition unit logs the data for further analysis. A clinical investigation on healthy subjects was conducted to test the proposed system. The results show a high variation in hand response across different subjects. In addition, for all subjects tested an ideal response was found which shows some justification for the use of the proposed technique. © 2005 IPEM. Published by Elsevier Ltd. All rights reserved. Keywords: FES; Upper arm; Hand extension

1. Introduction One of the consequences of hemiplegia is upper arm paralysis, which can seriously compromise a person’s autonomy. Conventional restoration of upper extremity function involves physiotherapy [1–3] with or without the assistance of orthotic devices [4–7]. Functional electrical stimulation (FES) is an alternative rehabilitation approach which is gaining popularity in the medical world and which can provide both orthotic and therapeutic benefits for persons with hemiplegia [8–10]. Both implanted and surface FES systems are available for upper arm applications. Upper limb FES devices provide therapeutic benefit by attempting to strengthen the muscles of the upper limb, reduce atrophy and increase muscle blood flow through the process of electrical stimulation. These devices provide orthotic benefit by eliciting artificial muscle contractions to enable some physical function such as a grasp or pinch to be performed. ∗

Corresponding author. Tel.: +353 61 202443; fax: +353 61 338176. E-mail address: [email protected] (D.T. O’Keeffe).

Depending on the actual orthotic benefit that the device provides, such as hand extension, palmar grasps, etc., the electrode positioning adopted with the device will vary. Of particular interest in this paper is the positioning of the electrodes adopted to restore hand extension movement. The recommended electrode positions used to elicit this form of movement [11] are illustrated in Fig. 1. Implantable FES devices such as the FreehandTM System developed by NeuroControlTM Corporation [12] are in use and clinical trials have shown a high rate of compliance and satisfaction with these devices [13]. However, implanted systems are costly and not all patients are comfortable with the idea of surgery. The use of noninvasive surface stimulation systems is therefore warranted in some cases for upper extremity electrical stimulation. Surface devices like the Handmaster [14], the Bionic Glove [15,16] and the ETHZ-ParaCare Neuroprosthesis [17] have already been developed for the upper limb with some success [18]. A common problem encountered by developers of surface stimulation devices for the upper limb is correct placement of the stimulation electrodes. In the case of the Bionic Glove and

1350-4533/$ – see front matter © 2005 IPEM. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.medengphy.2005.03.010

S.B. O’Dwyer et al. / Medical Engineering & Physics 28 (2006) 166–176

167

• finger flexion which is below a certain threshold (it was decided on a threshold of 25% of the maximum finger flexion angle); • hand adduction/abduction angles which are below a certain threshold (it was decided on a threshold of 25% of the maximum angle).

Fig. 1. The correct electrode positions for a balanced contraction of the wrist and finger extensors under electrical stimulation.

the ETHZ-ParaCare Neuroprosthesis, pad-electrodes must be placed by the clinician on the patient’s arm, before the glove is donned. This setup is unsatisfactory as even when the motor points for stimulation have been established, the user may find it difficult to independently place the electrodes. Having to place electrodes at correct stimulation sites also increases the time required for donning/doffing the device. This is especially true of the ETHZ-ParaCare Neuroprosthesis, which is reported to take between 7 and 10 min to don and doff [18]. Another problem encountered in such an upper arm FES setup is that movement of the electrode pads during operation can occur [15]. An alternative approach was adopted by the developers of the Handmaster system, who chose to fix the stimulation electrodes on the device, which means that the device is easily and quickly donned and doffed and can be done independently by the patient. However, a disadvantage of this setup is that it does not provide the user with sufficient flexibility to vary the position of the stimulation electrodes [18].

2. Methods This paper describes a technique for upper arm stimulation, which addresses the problems associated with both the Handmaster and Bionic Glove type upper arm FES systems. A key element of the technique is an algorithm, which selects from a matrix of electrodes, a pair of electrodes, which provide ideal orthotic performance for the subject. Current physiotherapeutic techniques advocate stroke rehabilitation exercise and training regimes that tend to be specific to task and context, i.e. related to the tasks to be learned [19]. Ideal orthotic performance was therefore defined as a hand response under electrical stimulation, which satisfies certain criteria particular to the mechanical attributes of a reaching motion. A reaching motion has been shown to have the following mechanical characteristics in relation to hand movement [20]: • hand extension angle of 15◦ ± 5◦ at the wrist joint;

The technique proposes that the electrode will be at fixed locations as in the Handmaster system. Once the device is donned, a configuration process can be initiated by the user, during which sensory data from a bi-axial accelerometerbased goniometer and several flex sensors attached to the hand and fingers of the user is recorded. An integrated accelerometer-based goniometer, the ADXL202 from Analog Devices1 measures the angle of extension of the hand while flex sensors measure the adduction and abduction of the hand and the flexion of the fingers. At the end of the configuration process the control unit uses an algorithm to evaluate the recorded sensory data. The electrode pair within the electrode matrix and the stimulation intensity that gives the hand response closest to ideal for the user is thus determined. This algorithm simply examines the data recorded from all tested electrode configurations and then returns the ‘best’ configuration found if one exists. A flow diagram of this algorithm is shown in Fig. 2. This approach has the dual advantage of giving sufficient flexibility to select from a variety of electrode positions whilst housing the electrodes within the glove so that the locations of the electrodes are fixed and donning/doffing times of the glove are reduced. In a typical application, the user could don the glove in the morning with ease without having to pay excessive attention to electrode placement. Once the system is booted it would automatically run the configuration program to determine which electrode pair to use and then this configuration would then be maintained until the user decides to doff the glove. Fig. 3 shows a schematic of the proposed device. A test unit was built so that the proposed technique could be evaluated; a block diagram of the test setup is shown in Fig. 4. The output of a commercial stimulator, the Neurotech NT20002 was channeled through an analogue de-multiplexer whose output channel was controlled by a single switch operated by the investigator. For each test iteration, the output of the de-multiplexer was connected to two of the six electrodes placed on the subject’s arm using the arrangement of Fig. 5. The proposed technique could facilitate an electrode matrix with a large number of electrodes at both extremities of the arm. For simplicity, six electrodes were positioned on the arm to illustrate the technique – two (electrodes A and B) at the proximal end of the wrist joint and a further four (electrodes C, D, E and F) at the distal end of the elbow joint. The 1 2

Analog Devices BV Ltd., Limerick, Ireland. BMR Ltd., Galway, Ireland.

168


electrodes were housed in two separate garment segments: one which held electrodes A and B and another which housed the remaining four. This garment division was necessary to accommodate subjects with varying arm lengths.

The movement sensors attached to the subject’s hand consisted of a twin axis accelerometer-based goniometer, used to measure the angle of extension of the hand and a number of flex sensors used to measure the angle of adduction and

Fig. 2. Flow diagrams of the electrode calibration algorithm. The first diagram shows the algorithm in its entirety, the second illustrates the isBestConfig algorithm subsection and the third diagram shows how the “Delta Value” used in the algorithm is calculated.


Fig. 2. (Continued )

169

170


Fig. 2. (Continued ).

Fig. 3. Concept diagram of the proposed calibration system. The sensors, control unit and electrodes are housed on the same glove, which can easily be donned and doffed.

abduction of the hand as well as the angle of flexion of the fingers. The complete set of sensors was housed in a glove so that all sensors could be attached or unattached by donning or doffing one garment. The glove was in turn attached to the sock garment which housed electrodes A and B. Photographs of part 1 and 2 of the garment are shown in Figs. 6 and 7, respectively. The analogue signals from these movement sensors were processed by a data acquisition system and then transmitted as digital data via an infrared link to a nearby PC, which recorded the sensory data at a sampling frequency of 50 Hz using customized software. This software also provided in-


171

Fig. 4. Test unit used for the clinical investigation. The test unit comprises of four main entities: the garment, the hardware unit, the Neurotech StimultorTM and the PC.

sity of Limerick Research Ethics Committee. The objective of the evaluation was: (1) to show that an ideal stimulation arrangement could be automatically detected from an array of electrodes and (2) to show that the hand response for similar electrode positions varies across different subjects.

Fig. 5. The approximate electrode locations used during the clinical investigation.

test feedback for each electrode configuration tested, such as the angle of extension of the subjects hand at the wrist joint, whether or not the hand response met the minimum criteria for an ideal configuration and if it not the reasons why? 2.1. Subjects To evaluate the technique a trial was conducted on 10 healthy subjects who gave written informed consent to participate in the evaluation, which was approved by the Univer-

For each of the 10 subjects the 6 electrodes were placed at approximately the same locations (Fig. 6), which are recommended to elicit hand extension [11]. Two electrodes (A and B) were placed adjacently to the base of the wrist while the remaining four were placed near the elbow joint on the extensor muscles. The electrodes used were 2.5 cm diameter PALS electrodes3 and the following stimulation parameters were utilized: • stimulation frequency 35 Hz; • an asymmetric biphasic pulse of width 350 ␮s; • ramp-up time 1 s and ramp-down time 0 s; 3

Axelgaard Manufacturing Co. Ltd., USA.

172


Fig. 6. Part 1 of the electrode garment which consists of the device sensors and the two distal electrodes which are housed in a garment which is pulled on over the subject’s hand.

• the “on-time” of the stimulation was 3 s and the “off-time” was 2 s; • for each electrode configuration, one stimulation duty cycle was delivered. Before the trial began the subject was first seated and instructed to clean his/her right arm with ethanol. The device was then donned with the help of the investigator. The flex sensitive resistors and the accelerometer-based goniometer used in the device were then calibrated. The subject was then told to rest his/her right arm on a provided bench with his/her hand dropping limp over the end of the bench as shown in Fig. 8. The subject was told to return to this start-

ing position at the end of each stimulation cycle during the trial. The complete trial involved four iterations of testing a predefined sequence of electrode configurations (Table 1). As previously discussed, the PC software, which logged the data from the movement sensors also provided in-test feedback for each electrode configuration tested. After each electrode configuration was tested the PC software displayed the angle of extension, adduction and abduction of the subject’s hand at the wrist joint and whether or not the hand response of the subject for the configuration met the minimum criteria for an ideal response. The investigator could then act on this information at the end of a test iteration to increase or decrease the


173

Fig. 7. Part 2 of the electrode garment which consists of the four distal electrodes housed in a garment which can be strapped onto the subject using the attached Velcro straps.

Fig. 8. The resting position assumed by the subjects when they were not undergoing electrical stimulation.

Table 1 Sequence of electrode configurations in chronological order

once again on the next iteration. This process was continued until either:

Sequence number

Anode

Cathode

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

A A A A B B B B C C D D E E F F

C D E F C D E F A D A B A B A B

stimulation intensity as he saw fit so that the subject’s hand response would approach an ideal response as previously defined. The stimulation intensity was subsequently changed to this new level and the same set of configurations was retested

(1) the subject requested that the trial be discontinued due to discomfort or other reasons and (2) a total of four test iterations had been completed.

3. Results Once the trial was completed, the evaluation algorithm was executed on the PC and it returned the stimulation arrangement (electrode configuration and stimulation intensity), which provided a hand response, which was closest to ideal. The algorithm also returned other useful statistics on all of the electrode configurations tested. The results are summarised in Table 2. Table 2 shows that an ideal stimulation arrangement was found for all subjects tested. The finger flexion shown is the average finger flexion of the subject’s four fingers. There seems to be a discrepancy when the percentage angles of adduction and abduction are both greater than 0 (e.g. subject 6). This is caused when the subject hand is adducted while

Table 2 Optimum stimulation arrangements for subjects Subject

Optimum intensity (maximum%)

Optimum configuration/ is it ideal?

Angle of extension (◦ )

Angle of abduction (maximum%)

Angle of adduction (maximum%)

Average angle of finger flexion (maximum%)

1 2 3 4 5 6 7 8 9 10

40 47 32 37 35 50 39 34 32 28

D → B/Yes B → D/Yes E → B/Yes E → A/Yes D → B/Yes E → A/Yes A → C/Yes B → E/Yes B → D/Yes E → B/Yes

16 11 11 19 17 18 15 10 15 19

10 6 1 15 6 13 12 13 4 9

0 7 3 0 0 14 0 0 0 0

15 22 1 14 13 3 21 9 21 12

174


Fig. 9. Graph of hand extension angles for subjects across electrode configurations.

Fig. 10. Graph of hand adduction angles for subjects across electrode configurations.

Fig. 11. Graph of hand/thumb abduction angles for subjects across electrode configurations.


175

Fig. 12. Graph of average finger flexion angles for subjects across electrode configurations.

Fig. 13. Standard deviation of hand response angles across electrode configurations.

the subject’s thumb is abducted because the abduction of the thumb also causes the hand abduction flex sensor to bend slightly. The individual subject hand responses in the categories of hand extension, hand adduction, hand/thumb abduction and finger flexion are shown in Figs. 9–12, respectively. The standard deviation of hand extension, adduction and abduction and finger flexion across every configuration were also computed and represented in graphical form (Fig. 13). This data was computed only from the test iteration in which the ‘best’ configuration was found. The data presented clearly shows how hand response fluctuates across subjects for similar electrode positions. Single factor ANOVA test on the data showed that the hand response was not similar across subjects at a significance level of 1%.

4. Discussion and conclusions This paper introduces a new electrode configuration system for FES based upper limb rehabilitation devices. The system identifies from a matrix of electrodes a stimulation arrangement, which provides an optimum hand response, which might be used in an FES-based exercise regime. A clinical evaluation of the technique produced positive results. The data presented shows that: (1) an ideal response was found for all the subjects tested in the trial (see Table 1) thus showing the effectiveness of the system and (2) hand extension and adduction/abduction and finger flexion angle responses fluctuate highly from subject to sub-


176

ject for similar electrode positions (see Fig. 13) and there is also high variation between different electrode positions for any given subject (see Figs. 9–12). Therefore, there is a strong justification for the proposed technique, which chooses from one out of a number of different stimulation arrangements to find the electrode pair and stimulation intensity, which precipitates movement that is closest to an ideal hand response. This stimulation arrangement might then be used to implement a FES-based rehabilitation exercise regime. The proposed technique has shown some success. A portable device capable of implementing the technique must now be developed fully and tested in a clinical setting on persons with hemiplegia.

Acknowledgments The authors would like to thank the Irish Research Council for Science Engineering and Technology (IRCSET) for their sponsorship of this project.

References [1] Bobath B. Abnormal postural reflex activity caused by brain lesions. London: H.M. Books; 1965. [2] Bobath B. Adult hemiplegia: evaluation and treatment. London: H.M. Books; 1970. [3] Brunnstrom. Movement therapy in hemiplegia. New York: H.A. Row; 1970. [4] Farrell J, et al. The effects of the functional tone management (FTM) arm training program on upper extremity motor control on chronic post-stroke individuals. J Stroke Cerebrovasc Dis 2003;12(5): 247. [5] Jean-Michel Gracies JEM, Roger R, Joseph S, Simon CG, David B. Short-term effects of dynamic Lycra splints on upper limb in hemiplegic patients. Arch Phys Med Rehabil 2000;81(12): 1547–55. [6] Lannin NA, et al. Splinting the hand in the functional position after brain impairment: a randomized, controlled trial. Arch Phys Med Rehabil 2003;84(2).

[7] Sodring KM. Upper extremity orthoses for stroke patients. Int J Rehabil Res (Internationale Zeitschrift Fur Rehabilitationsforschung, Revue Internationale De Recherches De Readaptation) 1980;3(1):33–8. [8] de Kroon JR, van der Lee J, IJzerman MJ, Lankhorst GJ. Therapeutic electrical stimulation to improve motor control and functional abilities of the upper extremity after stroke: a systematic review. Clin Rehabil 2002;16(4):350–60. [9] Cauraugh J, Light K, Kim S, Thigpen M, Behrman A. Chronic motor dysfunction after stroke: recovering wrist and finger extension by electromyography-triggered neuromuscular stimulation. Stroke 2000;31(6):1360–4. [10] Daly JJ, Marsolais E, Mendell LM, Rymer WZ, Stefanovska A, Wolpaw JR, et al. Therapeutic neural effects of electrical stimulation. IEEE Trans Rehabil Eng 1996;4(4):218–30. [11] Baker L, McNerl D, Bentin L, Bowman B, Waters R. Wrist extension: electrode placement in neuromuscular electrical stimulation—a practical guide. 3rd ed. California: Los Amigos Research and Education Institute; 1993. p. 117. [12] Smith B, Peckham P, Keith MW, Roscoe DD. An externally powered, multichannel, implantable stimulator for versatile control of paralyzed muscle. IEEE Trans Biomed Eng 1987;37(4):499–508. [13] Kilgore KL, Peckham P, Keith MW, Bryden AM, Wuolle KS. Restoration of grasp and release with an implanted neuroprosthesis (abstract). In: Proceedings of the first annual conference of the international FES society. 1996. [14] IJzerman M, Stoffers T, in ’t Groen F, Klatte M, Snoek G, Vorsteveld J, et al. The NESS handmaster orthosis: restoration of hand function in C5 and stroke patients by means of electrical stimulation. J Rehabil Sci 1996;9(3):86–9. [15] Prochazka A, Gauthier M, Wieler M, Kenwell Z. The bionic glove: an electrical stimulator garment that provides controlled grasp and hand opening in quadriplegia. Arch Phys Med Rehabil 1997;78(6):608–14. [16] Popovic D, Stojanovic A, Pjanovic A, Radosavljevic S, Popovic M, Jovic S, et al. Clinical evaluation of the bionic glove. Arch Phys Med Rehabil 1999;80(3):299–304. [17] Popovic MR, Keller T, Pappas IP, Dietz V, Morari M. Surfacestimulation technology for grasping and walking neuroprosthesis. IEEE Eng Med Biol Mag 2001;20(1):82–93. [18] Popovic MR, Popovic D, Keller T. Neuroprostheses for grasping. Neurol Res 2002;24(5):443–52. [19] Janet C, Roberta S. Guidelines for training, in stroke rehabilitation: guidelines for exercise and training to optimize motor skills. Butterworths/Heinemann; 2002. [20] Cynthia C, Norkin JW. Measurement of joint motion: a guide to goniometry. Davis FA; 1995.