Secondary Logo

Journal Logo

Original Article

Outcomes and Perception of a Conventional and Alternative Myoelectric Control Strategy

A Study of Experienced and New Multiarticulating Hand Users

Vilarino, Martin MSE; Moon, Jayet BTech; Rogner Pool, Kasey OTR, MOT; Varghese, Joby OTR, MSOT; Ryan, Tiffany OTR, MOT; V. Thakor, Nitish PhD; Kaliki, Rahul PhD

Author Information
Journal of Prosthetics and Orthotics: April 2015 - Volume 27 - Issue 2 - p 53-62
doi: 10.1097/JPO.0000000000000055
  • Free


In recent years, upper-limb myoelectric hands have been commercialized with more articulated joints than previously available. Often referred to as “multiarticulating hands,” these devices have the potential to increase functionality through the use of multiple hand grips. For instance, the i-limb™ ultrarevolution1 (Touch Bionics, Inc, Livingston, United Kingdom), the bebionic32 (RSL Steeper, Leeds, United Kingdom), and the Michelangelo hand3 (Otto Bock, Duderstadt, Germany) are capable of 24, 14, and 7 grip patterns, respectively. When prostheses include powered wrists and elbows, the number of degrees of freedom increases further. For the prosthesis user, these technological advancements hold the promise of restoring lost function.

However, technological capabilities have not directly translated into practical functional use. In the case of multiarticulating hands, users end up accessing only a fraction of all the grips available.4 One possible reason is that these devices have more output functions than input signals. Many myoelectric users employ two electromyography (EMG) sensors on antagonistic pairs of muscles (“dual-site control”),5 requiring a mode switch signal to switch among prosthesis modes, such as hand grips. Using a mode switch signal, such as a cocontraction (Figure 1), requires the mastery of complex muscle contractions and is considered slow, cumbersome, nonintuitive, and difficult for people to learn.6–8 In addition, using this mode switch signal can be cognitively demanding.9,10 These factors limit the user’s practical access to full functional use of the myoelectric hand. This is detrimental because lack of proper functionality is a cause of prosthesis rejection.11,12

Figure 1
Figure 1:
Conventional control strategy using electromyography (EMG) triggers to switch grips on a multiarticulating hand. When the user intends to switch hand grip, the user produces one of the following EMG triggers: cocontraction (both flex and extend signals rise above threshold at the same time), hold open (hand is stalled open and extend signal is held above threshold for a period), double impulse (hand is stalled open, and the extend signal sends two short pulses with relaxation between each pulse), or triple impulse (hand is stalled open, and the extend signal sends three short pulses with relaxation between each pulse). The electrode on the myoelectric user’s residual limb captures the user’s muscle activity and relays the signal to the hand.

In an effort to harness more functionality, new control strategies have been recently commercialized, such as pattern recognition,13 switching grips from a phone application,14 and wireless technology to detect objects and select grips. The latter has emerged in two forms: Bluetooth chips15 and radio frequency identification (RFID) tags.4

The RFID approach is the focus of this study. It includes an RFID controller, installed in the user’s prosthesis, and external RFID tags, small plastic items programmed with a grip switching signal. The tags can be placed in the user’s environment (i.e., at entrance to the living room) or carried on his or her person (i.e., in a pants pocket). When the user’s prosthesis approaches a tag, the RFID controller reads it and sends a signal to the hand that mimics an EMG trigger (i.e., cocontraction). This switches the grip on the hand (Figure 2).

Figure 2
Figure 2:
Alternative control strategy using radio frequency identification (RFID) tags to switch grips on a multiarticulating hand. When the user intends to switch hand grips, the user approaches the prosthesis to an RFID tag. The RFID tag is programmed with a trigger, and it transfers this information to the RFID controller in the prosthesis shell. The controller then sends the trigger signal to the hand. In this strategy, the user does not need to use electromyogram (EMG) to switch grips.

One advantage of this strategy is that the user does not need to engage any muscle to switch grips. As long as the user brings the prosthesis within range of an RFID tag, the controller automatically sends the signal. Unlike user generated signals, which can be affected by factors such as electrode contact and user fatigue, the controller’s signals, being digital, are consistent.

The tags come in many forms (i.e., keychain, business card), creating two possible scenarios. In the first, the user can tag his or her environment. For example, by placing a tag behind the closet door, the user can switch to a thumb park grip and slide the hand through a coat sleeve. In the second, the user can carry tags on his or her person (i.e., on clothing) to use them outside of tagged environments. When approaching an object of interest, the user can swipe the tag to switch grips on the way over. The study evaluates this second scenario.

The main purpose of this study was to compare the conventional use of EMG triggers (Figure 1)16 with the RFID strategy (Figure 2) for switching grips on the prosthesis. A preliminary case study suggested functional benefits of using RFID tags to access different modes.4 The proposed study includes a more comprehensive approach with more patients (n = 3). Subjects were evaluated on two performance metrics: trigger completion time and the percentage of triggers that were successful on first attempt (first attempt success rate). Subjects were also asked to rate the difficulty, effort, and frustration with each strategy. The goal of the study was to measure how quickly and successfully users can access grips and how satisfied users are with each mode switching strategy.


An optimal outcome measure has not been established for directly evaluating a dual-site myoelectric user’s ability to switch grips on a prosthesis. Many outcome measures often focus on task-oriented outcomes, such as moving or manipulating objects.17–19 In this study, the authors investigated subject performance and subjective experience using EMG triggers and RFID tags to switch grips on the i-limb. Performance metrics were recorded during the grip switching reliability test (GSRT), an outcome measure designed by the authors. Subjective metrics were quantified with the functional test questionnaire, a brief questionnaire designed by the authors and self-administered after the GSRT.

This work was approved by the Western Institutional Review Board (WIRB®). Written consent was obtained from each subject as per the submitted and approved protocol. The study was conducted at the Advanced Arm Dynamics Southwest Center of Excellence.


The GSRT measures how quickly and successfully a subject can execute the four i-limb grip switching triggers: cocontraction, hold open, double impulse, and triple impulse. The subject was evaluated using conventional EMG triggers (“EMG trial”) and RFID tags (“RFID trial”).


Each subject used an i-limb ultrarevolution. For all subjects, the i-limb was programmed with the grips shown in Table 1. These grips were selected because they are easily distinguishable and begin in the fully open position when triggered. In the EMG trial, the hand’s settings were set to the subject’s preferred trigger settings. In the RFID trial, the trigger duration and periods were changed to accommodate the RFID controller’s output signals. In the RFID trial, the subject was fit with suspenders. Each subject placed four RFID tags (one for each trigger) on the suspenders wherever the subject preferred, as shown in Figure 3. The assessor helped the subject place tags in accessible locations.

Table 1
Table 1:
Grips associated with each trigger during the grip switching reliability test (GSRT)
Figure 3
Figure 3:
Subject fit with the tag suspenders. Each subject wore four radio frequency identification (RFID) tags (one for each trigger) on suspenders during the RFID trial. The white disk RFID tags (indicated with black arrows) were attached to the suspenders with Velcro.

The subject was asked to sit down in front of a computer screen. The subject was given a maximum of 1 hr to practice using EMG signals and RFID tags to activate the different grips. The subject was allowed to view his or her EMG signals on BioSim (Touch Bionics) during the practice session only.


A series of triggers were displayed on the screen. The subject was instructed to attempt each trigger as soon as it was displayed. Only four unique triggers were displayed: cocontraction, hold open, double impulse, and triple impulse. However, each trigger was displayed multiple times. In the EMG trial, the subject was restricted to using muscle contractions to create each trigger. In the RFID trial, the subject was instructed to use RFID tags (e.g., if hold open was displayed, the subject approached the hold open tag).

As shown in Figure 4, assessor 1 monitored the subject’s screen and the subject’s i-limb hand. Every time the hand settled on the correct grip (grip associated with the displayed trigger), the assessor clicked one of two keys on the keyboard. One key was clicked if the subject executed the trigger on the first attempt, a different key if on a subsequent attempt. Assessor 2 monitored the subject’s signals on a second screen, occluded from the subject. In case the subject performed the trigger correctly on the first attempt, assessor 2 provided a cue to assessor 1. This process captured the first attempt success rate—the percentage of triggers that were successful on first attempt. If the subject was not able to produce the trigger within 30 seconds, it was considered a failed trigger attempt, and the subject was allowed to skip the trigger. Regardless of a success or fail, a 3-second time countdown separated each trigger. To reduce fatigue, the subject was allowed to rest after every set of four triggers. The subject was also allowed to rest between trials.

Figure 4
Figure 4:
Setup for the grip switching reliability test (GSRT). During the test, the subject attempted the trigger displayed on the screen directly in front of him or her. Assessor 1 observed the screen and the subject’s prosthesis, and was responsible for clicking a key when the hand settled on the correct grip. Assessor 2 monitored the subject’s signals on a second screen (occluded from the subject) and notified assessor 1 in case of a first attempt success.

Each trial contained 11 sets. Each set randomly presented each of the four triggers. Both trials contained the same 11 sets, but in different orders. To reduce user error based on adjustment to the test, the first set from each trial was discarded from the data. The order of the trials was determined by flipping a coin. The test software collected the data (trigger times and first attempt successes).


The functional test questionnaire was self-administered after the GSRT at the testing site. It consisted of 28 multiple choice options that prompted the subject to rate the difficulty, physical effort, and frustration while using EMG and RFID to execute each trigger. These items were reported on a Likert scale,20 similar to the Orthotics and Prosthetics User Survey Upper Extremity Functional Status (OPUS-UEFS) module.21 Each item was scaled from 1 to 5. For difficulty, 1 was defined as “very easy” and 5 as “very difficult.” For physical effort, 1 was defined as “low physical effort” and 5 as “high level of physical effort.” For frustration, 1 was defined as “low frustration,” and 5 was defined as “high level of frustration.” The questionnaire also prompted the subject to mark how he or she preferred executing each trigger, with or without tags. With a pencil, the subject circled his or her answers on the questionnaire sheet.


Three subjects were evaluated in this study. Subjects were recruited from a convenience sampling of patients presenting for prosthetic rehabilitation at Advanced Arm Dynamics (AAD) centers of excellence. Subject selection criteria included the following: 1) unilateral transradial amputation; 2) experienced dual-site control myoelectric prosthesis user; 3) no significant concomitant injuries; 4) no significant cognitive or psychological deficits; and 5) no previous knowledge or experience with RFID tags. Advanced Arm Dynamics occupational therapists screened subjects for criteria were listed previously.

Before testing, the subjects were asked to complete three published questionnaires: the Disability of the Arm, Shoulder and Hand (DASH),22 Trinity Amputation and Prosthesis Experience Scales—Revised (TAPES-R),23 and OPUS-UEFS.21 Their responses helped to develop the biographies below, summarized in Table 2.

Table 2
Table 2:
Study participant demographics


Subject 1 is a 34-year-old male who has been a bilateral transradial amputee since 2008. His amputations were a result of trauma/accident. Three months after amputation, he was fit with two body-powered prostheses on both arms. He was fit with myoelectric prostheses and i-limb hands on both arms around the same time, but due to complications, he did not wear his myoelectric devices until 2009. Therefore, at the time of testing, subject 1 had been wearing his i-limbs for 5 years. He reported that he wears his prosthetic devices an average of 16 hrs every day. In general, he rates his health as “good” and his physical capabilities as “very good.” He does not experience any residual limb pain but does experience phantom limb pain on both left and right limbs.


Subject 2 is a 35-year-old male who has been a unilateral transradial amputee since 2009. His amputation was a result of a work injury. Five months after amputation, patient was fit with the original i-limb and ETD (Motion Control, Salt Lake City, UT, USA). Four years after amputation, the patient was fit with the i-limb ultrarevolution and an activity-specific device. Therefore, at the time of testing, subject 2 had been using an i-limb hand for over four years. He reported that he wears his i-limb hand an average of 10 hrs every day. In general, he rates both his health and physical capabilities as good. He experiences both residual and phantom limb pain.


Subject 3 is a 49-year-old female who has been a unilateral transradial amputee since 1970. Her amputation was a result of an injury from a lawn mower as a child. Subject was initially fit with a body-powered prosthesis as a child. After 34 years, she was fit with a Vari Plus® hand (Otto Bock, Duderstadt, Germany). She reported that she wears her Vari Plus an average of 12 hrs every day. In addition, Subject 3 currently uses an electric wrist rotator, in which she is well trained. However, in her daily prosthesis use, she does not utilize the wrist often. Her primary prosthesis is programmed for cocontraction to activate the electric wrist. At the time of testing, she had never been fit with an i-limb hand or multiarticulating hand. In general, she rates her health as good and her physical capabilities as “fair.” She does not experience any residual limb pain or phantom limb pain.


The data were processed and analyzed using statistical package Minitab 16 (Minitab Inc, State College, PA, USA) and Microsoft® Excel 2013 (Microsoft Corporation, Redmond, WA, USA). Subject 1 performed the EMG trial first. Subjects 2 and 3 performed the RFID trial first. Subject 1 failed one trigger (hold open) in the EMG trial. Subject 3 failed one trigger in the EMG trial (double impulse) and one trigger in the RFID trial (hold open). Table 3 presents each subject’s average trigger completion times in the EMG and RFID trials. The failed trigger times were not included in the averages and SDs. Figure 5 presents these data graphically. The subjects’ first attempt success rates for all triggers in both trials are captured in Table 4.

Table 3
Table 3:
Average grip switching time and SD for electromyogram (EMG) and radio frequency identification (RFID) trials
Table 4
Table 4:
First attempt success rate for all subjects
Figure 5
Figure 5:
Average grip switching times for electromyogram (EMG) and radio frequency identification (RFID) trials. Each subject’s average grip switching time is shown for each trigger in the RFID trial (darker gray) and EMG trial (lighter gray). The time represents the average for 10 trigger attempts. SD bars are shown. A white X at the base of the bar indicates that a failed trigger attempt (time, 30 seconds) occurred in that set. Failed trigger attempts were not included in the calculations for the averages and SDs.

On average, subject 1 performed all four triggers faster in the EMG trial than in the RFID trial. Subject 1 attained a higher first attempt success rate for three triggers (cocontraction, double impulse, triple impulse) in the RFID trial and an equivalent rate for the fourth trigger (hold open) in both trials. In the RFID trial, the subject achieved a first attempt success rate of 100% for two of the triggers (double impulse and triple impulse).

On average, subject 2 also performed all four triggers faster in the EMG trial than in the RFID trial. Subject 2 attained a higher first attempt success rate for cocontraction in the RFID trial, a higher rate for double impulse in the EMG trial, and an equivalent rate for the other two triggers (hold open and triple impulse) in both trials. In both the EMG and RFID trial, the subject achieved a first attempt success rate of 100% for hold open.

On average, subject 3 performed two triggers (cocontraction and double impulse) faster in the RFID trial and two triggers (hold open and triple impulse) faster in the EMG trial. Subject 3 attained a higher first attempt success rate for three triggers (cocontraction, double impulse, triple impulse) in the RFID trial and a higher rate for the fourth trigger (hold open) in the EMG trial. The subject achieved a first attempt success rate of 100% for cocontraction in the RFID trial and a rate of 100% for hold open in the EMG trial.

Across all subjects in the EMG trial, subject 1 had the fastest trigger times, and subject 3 had the slowest. In the RFID trial, the subjects had more comparable times. Across all subjects in the EMG trial, subject 3 had the lowest first attempt success rates for double and triple impulse (30% and 40%, respectively). Across all subjects in the RFID trial, subject 3 had the highest first attempt success rate for cocontraction (100%).

In the EMG trial, subjects 1 and 2 performed the cocontraction the fastest and hold open the slowest. In the same trial, subject 3 performed the cocontraction and hold open the fastest, and the impulses the slowest.

Figure 6 displays the subjects’ rating of difficulty, effort, and frustration for each strategy. Table 5 displays each subject’s preferred strategy for performing each trigger.

Table 5
Table 5:
Subject preferred strategy for performing each trigger
Figure 6
Figure 6:
Subject rating of difficulty, physical effort, and frustration for electromyogram (EMG) and radio frequency identification (RFID) strategies. For each figure (1A–3C), each corner is labeled with a different trigger. Each figure contains a darker gray shape (subject ranking of EMG) and a lighter gray shape (subject ranking of RFID strategy). Smaller shapes, containing corners near the center of the figure, suggest lower levels of the specified category (difficulty, physical effort, frustration). All subjects rated all triggers using RFID with the lowest level of difficulty, effort, and frustration (1 of 5). The subjects rated the triggers using EMG with a variety of different values.

All subjects rated all triggers using RFID with the lowest level of difficulty, effort, and frustration (1 of 5).

Using EMG, subject 1 rated half of the triggers (hold open and triple impulse) with a moderate level (3 of 5) of difficulty, effort, and frustration, and the other half (cocontraction and double impulse) with the lowest level (1 of 5). He preferred using RFID for two triggers (hold open and triple impulse), preferred using EMG for one trigger (cocontraction), and had no preference for the last trigger (double impulse).

Using EMG, subject 2 rated all triggers with a low level (2 of 5) of physical effort. For difficulty and frustration, the subject rated one trigger (cocontraction) with a moderate level (3 of 5) and the other three triggers with the lowest level (1 of 5). He preferred using RFID for all four triggers.

Using EMG, subject 3 rated triple impulse with a high level (4 of 5) of difficulty, cocontraction with a moderate level (3 of 5), double impulse with a low level (2 of 5), and hold open with the lowest level (1 of 5). For physical effort, the subject rated cocontraction and double impulse with a low level (2 of 5) and hold open and triple impulse with the lowest level (1 of 5). For frustration, the subject rated cocontraction with a high level (4 of 5), double impulse with a low level (2 of 5), and hold open and triple impulse with the lowest level (1 of 5). She preferred using RFID for two triggers (cocontraction and double impulse) and EMG for the other two triggers (hold open and triple impulse).



Our result that subject 3 had the slowest EMG trigger times suggests that this new user had difficulty executing triggers efficiently. In addition, the subject’s low EMG first attempt success rates indicate that executing triggers consistently with EMG was another challenge for this user. Subject 3 exhibited performance benefits using the RFID approach. She was able to perform two triggers faster using RFID, particularly double impulse. With RFID, her first attempt success rate was greater for three of the four triggers. For these triggers, her RFID performance was comparable to that of subjects 1 and 2.

Subjects 1 and 2 had relatively fast trigger times and high first attempt success rates using EMG, suggesting that they were proficient with this strategy. There are two possible reasons why these experienced subjects performed the triggers faster with EMG than with RFID. First, they had been using EMG triggers for years and had received extensive and specialized occupational therapy training. In contrast, they had only been exposed to the RFID strategy on the day of the study. Therefore, they exhibited a higher risk of error during the RFID trial (i.e., approach wrong tag). Second, recall that in the RFID strategy, the user must approach a tag before the RFID controller sends the trigger signal to the hand. Although the trigger sent by the RFID controller is efficient and consistent, an experienced i-limb user may execute EMG triggers with a comparable time to that of the RFID controller. Therefore, the additional time necessary to approach a tag may have contributed to the time difference that was observed between EMG and RFID trials (typically 1–2 seconds). However, using RFID, there were advantages in the first attempt success rates for subject 1, likely because the RFID controller creates signals that are more consistent than EMG (digital signal compared with biological signal).

Because of the limited subject pool, it is difficult to derive generalized trends about the subjects’ performance. However, certain observation can provide useful insight. For example, unlike the EMG trigger times and first attempt success rates, the three subjects had comparable RFID times and rates. This suggests that prior experience with multiarticulating hands may not be necessary for a user to use the RFID strategy effectively.

In addition, the order of the subjects’ EMG trigger times can reveal how trigger characteristics affect new and experienced user performance with EMG triggers. The experienced users (subjects 1 and 2) exhibited the following order: cocontraction fastest, hold open slowest. This is likely a result of the intrinsic property of each trigger: for cocontraction, the user’s signals need pass a threshold once; for the impulses, the signal must cross more than once; for hold open, it must cross and stay above threshold for an extended time. For a user that has mastered the triggers (such as subjects 1 and 2), the performance is not affected by the complexity of the trigger, but rather the inherent delays in each trigger. However, trigger complexity had an effect on the new user. The pattern in her EMG trigger times (hold open fastest, impulses slowest) suggests that she mastered hold open fastest, but it took her longer to learn the impulses (also supported by her first attempt success rates). Based on these results, starting new multiarticulating hand users with only hold open and cocontraction may improve initial performance and reduce frustration.


For the experienced subjects, the true benefits of RFID were revealed in the subjective data. Despite RFID being a new approach for them, they both marked all RFID triggers with the lowest level of difficulty, physical effort, and frustration. Despite EMG being a very familiar approach for them, they marked some triggers as more difficult, tiring, and frustrating. Furthermore, when asked which approach they preferred for each trigger, subject 1 marked RFID for hold open and triple impulse. This suggests that subject 1’s weighted his perception of each strategy more heavily than his performance with each strategy; despite performing hold open and triple impulse faster with EMG, the subject preferred RFID strategy for both triggers. Subject 2 exhibited the same behavior, but more strongly; despite being faster with all triggers using EMG, he preferred using RFID for all triggers. For these subjects, the delay associated with approaching a tag was considered less of a burden than the physical and mental effort required for specific EMG triggers. It is also possible that the time difference between EMG and RFID was imperceptible to these users.

In addition, the subjective data for subject 3 offer valuable insight because she was new to both strategies. Like the experienced users, she considered RFID easy to adapt to. This was not the case for EMG. She considered most triggers more difficult and two of them more tiring and frustrating. However, unlike subjects 1 and 2, subject 3’s preference may have been driven by her performance. For example, despite her marking it easier to use RFID to create a triple pulse, her EMG preference for the triple pulse may have been driven by her faster performance in the EMG trial.


The fact that neither EMG nor RFID achieved a consistent first attempt success rate of 100% suggests that both technologies are not yet perfect and are still evolving. In the case of EMG, this inconsistency is likely because EMG patterns are difficult to replicate; users experience innate variations in their EMG due to many factors, such as fatigue. For RFID, it may be because the i-limb hand only accepts some triggers in a specific state (when the hand is stalled open). For example, the hand may not recognize the RFID controller’s signal if a user approaches a tag after accidentally closing the hand.

The RFID controller would be more effective if it could directly command the hand to switch grips. Instead of sending the hand a signal that mimics an EMG trigger, it would send an instantaneous, unique code that the hand associates with a grip. This would reduce the risk of the hand not recognizing the signal and would reduce the delay associated with the trigger signal (particularly for triggers such as the hold open). This could lead to a strategy that is faster than EMG, even for experienced users; the delay associated with the RFID strategy would be reduced to that of approaching a tag only. In addition, using triggers, i-limb users can only access five grips at a time. If the RFID controller could directly command the hand, it could open up access to all the grips available on the hand.


A limitation of the study is the small subject population. Based on the time differences observed in this initial study, a power analysis can be performed to determine how many subjects and trials would be necessary for statistically significant data. Future large scale clinical studies could then draw stronger conclusions from a larger subject pool. In addition, the subject pool could include experienced RFID users. This may help reveal the effect of prior experience for both strategies, as well as willingness to adopt new technology. Moreover, the study is not a perfect indicator of actual use because it represents a controlled environment. A future study would benefit from monitoring grip use at home using EMG triggers and RFID tags.

Despite the limitations, the results of this study offer valuable insight and suggest that EMG and RFID each offer benefits to experienced and new users. Therefore, we advocate synergistic use of both strategies. By allowing the user to choose which strategy to use for each trigger, the user is empowered with improved control and functionality of the device.


Our study reveals valuable information about control strategies. First, user preference for a control strategy can be based on performance with the strategy, perception of the strategy, or a combination of both. As observed, the experienced users tended to value perception of strategy more heavily when choosing which strategy they preferred for activating each trigger. On the other hand, the new user may have been biased by her performance. Regardless, it is clear that performance alone is not enough to evaluate the effectiveness of a control strategy. Second, for the EMG strategy, the performance of experienced users is limited by inherent trigger delays, whereas the performance of new users is limited by trigger complexity. For this reason, hold open was the slowest trigger for the experienced users but the fastest for the new user. However, it is important to note that these experienced users both received extensive training and were proficient with EMG triggers. Although it is preferred that all multiarticulating patients receive extensive training, this may not be the case for many users. Third, despite the subjects being exposed to RFID on the day of the study, their perception of RFID was very favorable. This suggests that for both experienced and new multiarticulating hand users, the RFID strategy is easy to adopt. Fourth, RFID yielded greater performance benefits for the new user. Because new users often exhibit the highest risk of prosthesis rejection, including RFID in new users’ prostheses could mitigate the risk of rejection. If available, this could occur in conjunction with training with EMG so that the user becomes proficient with both strategies. Finally, using RFID and EMG synergistically can yield beneficial results for both experienced and new users. It should be noted that our conclusions are drawn from a small number of subjects. Continued studies with a larger subject pool are necessary to determine factors influencing performance and patient preference to identify best strategies to access the full potential of new commercial devices. However, these results are promising and suggest that providing users with both control options empowers the user with more control and functionality of their prosthesis.


The authors would like to thank the study subjects and other amputee collaborators. In addition, the authors thank John Miguelez, Rob Dodson, Kerstin Baun, Mac Lang, Janice Hsu, Shelley Clem, and Brennan Hooper from Advanced Arm Dynamics. Finally, the authors thank Kevin Liu, Eugene Damiba, Goh Si Wei, Alex Diehl, and Rodolfo Finocchi from Infinite Biomedical Technologies.


1. Touch Bionics, Inc. i-limb™ Ultra Revolution. User Manual. Part MA01141, Issue 2. Sept 2014:9–13. Available at:
2. RSL Steeper Ltd. Bebionic3 Technical Information. Part RSLLIT317, Issue 2 2012:12–18. Available at:
3. Otto Bock Healthcare GmbH. The Michelangelo® Hand in Practice. Available at: 646D593=EN-01-1201:26–28.
4. Vilarino M. Technical note: a novel wireless controller for switching among modes for an upper-limb prosthesis. Academy Today 2014; 10(1): A12–A15.
5. Lake C, Miguelez JM. Evolution of microprocessor based control systems in upper extremity prosthetics. Technol Disabil 2003; 15(2): 63–71.
6. Li G, Geng Y, Tao D, Zhou P. Performance of electromyography recorded using textile electrodes in classifying arm movements. Conf Proc IEEE Eng Med Biol Soc 2011; 2011: 4243–4246.
7. Scheme E, Englehart K. Electromyogram pattern recognition for control of powered upper-limb prostheses: state of the art and challenges for clinical use. J Rehabil Res Dev 2011; 48(6): 643–659.
8. Kuiken Kuiken TA, Li G, Lock BA, et al. Targeted muscle reinnervation for real-time myoelectric control of multifunction artificial arms. JAMA 2009; 301(6): 619–628.
9. Soares A, Andrade A, Lamounier E, et al. The development of a virtual myoelectric prosthesis controlled by an emg pattern recognition system based on neural networks. J Intell Info Syst 2003; 21(2): 127–141.
10. Bongers RM, Kyberd PJ, Bouwsema H, et al. Bernstein’s levels of construction of movements applied to upper limb prosthetics. J Prosthet Orthot 2012; 24(2): 67–76.
11. Biddiss E, Chau T. Upper-limb prosthetics: critical factors in device abandonment. Am J Phys Med Rehabil 2007; 86(12): 977–987.
12. Mcfarland LV, Hubbard Winkler SL, et al. Unilateral upper-limb loss: satisfaction and prosthetic-device use in veterans and service members from Vietnam and OIF/OEF conflicts. J Rehabil Res Dev 2010; 47(4): 299–316.
13. Lock BA, Cummins FD. Search Engines for the World Wide Web: Simplified Approach to Myoelectric and Electrode Placement for Success in Clinical Pattern Recognition. Presented at Myoelectric Controls symposium MEC 2014; August 19-22, 2014; Fredericton, New Brunswick, Canada. Available at: Accessed January 15, 2015.
14. Touch Bionics, Inc. iPod Touch with biosimTM App: Quick Start Guide for i-limbTM ultra revolution, Part MA01124, Issue 1, April 2013:1–22. Available at:
15. Touch Bionics, Inc. Grip ChipsTM Datasheet, Part MA01254, Issue 1, May 2014. Available at: Available at:
16. Touch Bionics, Inc. i-limb™ ultra Clinician Manual, Part MA00003, Issue 2, March 2013: 39–43. Available at:
17. Light CM, Chappell PH, Kyberd PJ. Establishing a standardized clinical assessment tool of pathologic and prosthetic hand function: normative data, reliability, and validity. Arch Phys Med Rehabil 2002; 83(6): 776–783.
18. Jebsen RH, Taylor N, Trieschmann RB, et al. An objective and standardized test of hand function. Arch Phys Med Rehabil 1969; 50(6): 311–319.
19. Resnik L, Adams L, Borgia M, et al. Development and evaluation of the activities measure for upper limb amputees. Arch Phys Med Rehabil 2013; 94(3): 488.e4–494.e4.
20. Jamieson S. Likert scales: how to (ab)use them. Med Educ 2004; 38(12): 1217–1218.
21. Burger H, Franchignoni F, Heinemann AW, et al. Validation of the orthotics and prosthetics user survey upper extremity functional status module in people with unilateral upper limb amputation. J Rehabil Med 2008; 40(5): 393–399.
22. Hudak PL, Amadio PC, Bombardier C. Development of an upper extremity outcome measure: the DASH (disabilities of the arm, shoulder and hand) [corrected]. The Upper Extremity Collaborative Group (UECG). Am J Ind Med 1996; 29(6): 602–608.
23. Gallagher P, Franchignoni F, Giordano A, Maclachlan M. Trinity amputation and prosthesis experience scales: a psychometric assessment using classical test theory and Rasch analysis. Am J Phys Med Rehabil 2010; 89(6): 487–496.

prosthetics; upper limb; myoelectric; multiarticulating hand; prosthesis control strategy; electromyography; radio frequency identification; patient satisfaction

© 2015 by the American Academy of Orthotists and Prosthetists.