Mirror versus parallel bimanual reaching
© Abdollahi et al.; licensee BioMed Central Ltd. 2013
Received: 5 July 2012
Accepted: 14 June 2013
Published: 9 July 2013
In spite of their importance to everyday function, tasks that require both hands to work together such as lifting and carrying large objects have not been well studied and the full potential of how new technology might facilitate recovery remains unknown.
To help identify the best modes for self-teleoperated bimanual training, we used an advanced haptic/graphic environment to compare several modes of practice. In a 2-by-2 study, we compared mirror vs. parallel reaching movements, and also compared veridical display to one that transforms the right hand’s cursor to the opposite side, reducing the area that the visual system has to monitor. Twenty healthy, right-handed subjects (5 in each group) practiced 200 movements. We hypothesized that parallel reaching movements would be the best performing, and attending to one visual area would reduce the task difficulty.
The two-way comparison revealed that mirror movement times took an average 1.24 s longer to complete than parallel. Surprisingly, subjects’ movement times moving to one target (attending to one visual area) also took an average of 1.66 s longer than subjects moving to two targets. For both hands, there was also a significant interaction effect, revealing the lowest errors for parallel movements moving to two targets (p < 0.001). This was the only group that began and maintained low errors throughout training.
Combined with other evidence, these results suggest that the most intuitive reaching performance can be observed with parallel movements with a veridical display (moving to two separate targets). These results point to the expected levels of challenge for these bimanual training modes, which could be used to advise therapy choices in self-neurorehabilitation.
KeywordsBimanual training Bimanual coordination Self-telerehabilitation Upper limb
Bimanual training is particularly important in fostering recovery in neural injuries resulting in hemiparesis such as stroke, because the non-affected arm can potentially retrain the affected arm [1–3]. Upper extremity movements such as buttoning a shirt or zipping a jacket are simple but essential actions that need to be regained for making progress towards motor recovery and regaining activities of daily living (ADL). And although many studies focus on perfect performance, patients most care about task completion with proper coordination of both arms . Therefore, bimanual therapy, which gives the users the possibility of achieving their primary goal of retraining ADL’s, should be of major importance to the therapist.
Several studies have demonstrated that bimanual training improves coordination between the paretic and non-paretic arms [5–11]; while a few have shown unwanted outcomes, such as reductions in the Fugl-Meyer or Ashworth scales (See [12, 13] for a review). This discrepancy in performance may be due to the various ways the two limbs interact and move during bimanual training; for example, there are mirror-symmetric movements with respect to the body midline and asymmetric/alternating flexion-extension movements among others. Additionally, rigid coupling of the limbs (locking the actions of one limb to the other) often present in bimanual training could enable the paretic limb to act passively and depend mainly on the less affected limb for control, thus reducing its experience of the forces and motions associated with a particular movement. This may be why Kadivar et al., 2011 found no significant differences in bimanual performance (i.e. trajectory error) when one arm was rigidly forced to follow the other in parallel and mirror modes . Thus, it remains unclear which methods are optimal.
Placing a subject in an environment that manipulates the visual feedback may help to resolve this discrepancy in the literature. Virtual environments can allow each arm to perform independently while presenting novel visual feedback. This promotes active participation of both limbs, and hence each limb is making and learning from its own mistakes. Furthermore, using this paradigm may tell us which form of uncoupled bimanual practice provides superior results i.e. mirror vs. parallel mode. Previous robotic rehabilitation studies that used the mirror mode of bimanual practice showed a significant increase in brain activation in similar parts of both brain hemispheres as well as enhanced inter-hemispheric activation [1, 2, 9, 15–18]. However, to our knowledge, the intuitiveness, in terms of how fast and how accurate people can perform in this mode of practice, compared with the parallel mode, has not been studied in the uncoupled condition.
With two hands involved in practice, there are several approaches for coordinating both limbs. One is based on symmetry -- either transfer actions (parallel motions such as transferring a large object) or joint-similar actions (mirror motions such as opening a book). Lewis and Byblow reported that patients respond better to bimanual practices that involve in-phase and symmetric actions also denoted as mirror movements [19–21], which has been attributed to simultaneous brain activation of bilaterally homologous areas during these activities. However, these activations are not necessarily associated with functional gains, and the performance in a parallel mode was shown to be superior to mirror in a triangle drawing task . Hence, it remains to be seen whether parallel or mirror modes might be superior in terms of trajectory error and/or task completion time with the limbs decoupled.
Besides muscle grouping and coordination, visual attention also plays a role in task difficulty in targeted reaching. Virtual reality displays allow the possibility of transforming one of the hand’s feedback to the opposite side, so that subjects only need to attend to one side of their view. We hypothesize that such a “one-target” visual transformation might reduce task difficulty over managing a divided view to two targets.
The present study used healthy individuals in a virtual environment to examine how these modes of bimanual practice influence performance on a simple reaching task. We investigated how different modes, symmetry and feedback, might influence performance and rate of learning (change of performance across time). Specifically, we focused on differences in bimanual reaching due to mirror vs. parallel arm movements. We investigated the performance of uncoupled, bimanual point-to-point reaching under four conditions; mirror reaching to one target (the “one-target” visual transformation), mirror reaching to two targets, parallel reaching to one target, and parallel reaching to two targets. This study showed lowest completion times and trajectory errors for parallel movements reaching to two targets, identifying the least challenging mode for bimanual practice, which may suggest the most appropriate mode for self-therapy in future neurorehabilitation interventions.
Twenty healthy right-handed individuals (12 male, age range 19–53, mean age 28 ± 9) with corrected 20/20 vision were invited to participate and consented using approved Institutional Review Boards from both Rehabilitation Institute of Chicago and University of Illinois at Chicago guidelines for protection of human subjects Internal Review Boards according to the declaration of Helsinki. A pilot study determined the effect size and inter-group variance to be 1.21 and 1.41 seconds, leading us to a power estimate of 5 subjects in each of four treatment groups (described below) based on Cohen’s method for an ANOVAs with a targeted power of 0.8 and significance levels of 0.05. Participants were naïve to the apparatus and had no history of previous musculoskeletal or neurological injury. The handedness of each individual was assessed using the modified Edinburgh Handedness Inventory . Subjects were excluded if they scored less than 90 percent on the right-handedness test or if they had depth perception impairment of less than 8 out of 9 on graded circle test (Stereo Optical Company, Chicago, IL, USA).
For the purposes of familiarization and to check the effects of bimanual movements on unilateral performance, subjects began with 40 unilateral movements per hand to randomly placed targets before and after the bimanual task. Each session consisted of 200 bimanual practice trials; 100 of these trials were center-out reaches to randomly placed targets. These outward trials were evaluated for their performance.
Because movement speed and accuracy are believed to intimately tradeoff , we assessed both movement time and trajectory error as primary measures of interest. Each trial’s movement time was calculated from the time that both cursors left the home position until the time they both entered their target radius and remained there for 0.5 sec. Each trial’s trajectory error was summarized using the typical measure of maximum perpendicular distance to the ideal line to the target .
Learning curves were plotted for all trials, but the above-mentioned measures were calculated only for the first and last 20 movements in the practice phase (gray shaded area, Figure 3, bottom). Repeated measures ANOVA was performed on both measures with main (between) factors being movement type (mirror vs. parallel) and number of targets (one vs. two) and the within factors being location of targets and different evaluation times in each trial. Statistical alpha levels were 0.05 to detect significance.
Movement times changed least across practice for the parallel two-target group, indicated by a significant interaction amongst movement type, number of targets and practice (F(1,16) = 5.03, p < 0.05). Movement time changed an average of 1.43 seconds less across practice for the groups reaching in parallel, indicated by a significant interaction between movement type and practice (F(1,16) = 12.86, p < 0.01) (Figure 4, top). Furthermore, movement time changed an average of 1.49 seconds less across practice for groups reaching to two targets, as indicated by a significant interaction between number of targets and practice (F(1,16) = 14.07, p < 0.01). As Figure 4 (top) shows, the parallel two-target group begins with low movement time and exhibits a “floor effect” where there is little opportunity for improvement beyond their initial movement time .
Trajectory error results differed from movement time results. There was no significant change in trajectory error across practice for 16 of the 20 subjects from the beginning to the end of trials within each group (individual t-test, Figure 4, indicated by dashed lines). The right hand trajectory errors changed an average of 8 mm less across practice for the groups reaching in parallel, indicated by a significant interaction between movement type and practice (F(1,16) = 5.16, p < 0.05). The parallel two-target group showed a lower average error for both hands compared to all other groups even after training (Figure 4, middle and bottom). Finally, different target locations did not significantly affect movement time or trajectory errors.
This work shows that there are significant differences between how subjects perform bimanual targeted-reaching tasks under differing visual feedback modes. Among the four groups tested, the mode that involved parallel reaching to two targets clearly showed the lowest errors and shortest completion times at the beginning and throughout the experiment. In repetitive practice, errors did not tend to change across trials. Subjects tended to maintain lower left hand than right hand errors while increasing the average speed (and reducing completion times) of the left. While this paper does not attempt to understand any underlying neurophysiological processes, it reveals behavioral evidence that can inform choices in future bimanual applications.
The small change in both completion time and trajectory error observed in the group performing parallel reaching to two targets suggested that there was little (if any) learning. This lack of change maybe due to a floor effect because error was low at the start and throughout the trials. In addition, we speculate that this is the most familiar or intuitive mode of bimanual activity making it easiest for subjects to execute. This is consistent with “directional compatibility,” where limbs are more coordinated when endpoint directions agree . We also speculate that other modes were more difficult, making them initially less familiar, fostering learning, and leading to improvement across training. This was especially dramatic for the mirror transformation, which had the largest errors, slowest completion times, but showed the largest amount of change (learning) across practice. Nevertheless, no other groups’ final errors were as low as the mode involving parallel reaching to two targets, suggesting that this mode is, by far, the most intuitive.
One issue not investigated in this initial study is the persistence of any learning effects. Depending on the bimanual training application, retention may be required at different times. Hence, the appropriate time for follow-up tests and the durability of learning should be evaluated in future use of our results in a particular application.
Our results differ from a related study by Kadivar et al., 2011, in which no difference between bimanual parallel and mirror modes was found. Our results, which detected significant differences, may have been due to differences in task between these two studies -- their task coupled the limbs through a robotic interface, while ours allowed each hand to move independently. Our study also calculated error differently -- we used maximum perpendicular distance from the line to the target, while Kadivar and colleagues averaged this distance and divided by path length. Such dividing by path length can mask error. For example, a movement with several reversals might result in a deceptively low value if divided by its long length. Our data showed similar trends for both of these measures, but normalized average error produced more variable results.
Contrary to our assumption that attending to only one visual target area would simplify the task, we observed longer movement times in the “one target” modes that involved cursor transformations. This poor performance may result from the subject’s need to reinterpret or mentally transform the conflict between the movements of the hands and its associated visual feedback [28–30]. Such conflict may place a further attentional burden that lengthens completion time. Furthermore, of these two transformed modes (one target), mirror feedback showed longest completion times of all and hence was deemed the most difficult. Although participants in this group significantly improved in movement time, the final performance was still not as good as other groups even after 200 trials. The remaining “two target” feedback modes performed significantly better, which suggests that attending to two different visual areas is easier than mentally transforming visual cues. This separation of targets to different areas of visual space may also involve parallel computations in separate somatotopic areas of visual cortex that require less competing neural resources [31, 32]. Also, such visual transformations are not commonly encountered in the physical world, while simultaneous attention to two areas is a frequent ecological challenge to humans in tasks such as typing, drawing, and playing video games . Parallel modes, now possible with such virtual reality technology, may provide the most intuitive feedback for training environments.
Our mirror (one target) approach also differed from approaches that use physical mirrors to display limb actions [33–36]. In previous mirror approaches, reflection of one hand replaced the visual feedback of the other. Here, we transformed the right hand cursor so that it appeared on the same side as the left cursor, which we speculate to be more challenging. Such a mirror transformation could provide a “feedback puzzle” that may promote learning. Such complex challenges may encourage recovery better than intuitive ones , but these more challenging tasks might also be discouraging to some individuals. Hence, the results of this study serve merely as a guide to identify training modes that are either challenging or intuitive.
Nearly all participants kept error constant across training while decreasing completion times (with the exception of one subject with very high initial error). Participants improved speed rather than accuracy, which is one choice in the scheme of speed-accuracy tradeoff (Fitts’ law) . Some have shown increasing speed in the course of learning a skill , while others have shown error reduction . Therefore, changes in speed or accuracy may depend on the task. Interestingly, each group’s error was maintained at a different level. We speculate that each bimanual task requires its own level of information processing until a competent strategy is learned. Therefore, subjects hold error constant, begin slowly and speed up as they train. It remains to be seen whether these error levels reflect physiological limits in sensorimotor pathways or simply a different “tolerance” for error in each feedback condition.
These results have implications in rehabilitation, where bimanual interactions can assist a person in re-learning movement skills . Our results suggest that parallel reaching to two targets may be the optimal method for such self-telerehabilitation because it is the most familiar (least challenging) mode of practice. To the patient, however, improvement in bilateral symmetry may not be as important as completing a functional tasks, some of which are asymmetric. It is also possible that intuitive modes for healthy may not be equally intuitive for brain injured individuals. It remains to be seen whether these results translate effectively to neurorehabilitation. Nevertheless, the initial findings presented here in healthy subjects can help identify environments for rehabilitation or in any training situation requiring bimanual practice.
Activities of daily living
Virtual reality and robotic optical operations machine
Personal augmented reality immersive system.
This work was supported by the National Institutes of Health (NIH) under grant 1 R01 NS053606 and the National Institute for Disability and Rehabilitation Research (NIDRR) RERC H133E0700 13. We thank Daniel Evestedt of SenseGraphics for his technical programming assistance.
- Lum P, Burgar C, Shor P, Majmundar M, Van der Loos M: Robot-assisted movement training compared with conventional therapy techniques for the rehabilitation of upper limb motor function following stroke. Arch Phys Med Rehabil 2002, 83: 952-959. 10.1053/apmr.2001.33101View ArticlePubMedGoogle Scholar
- Hesse S, Schulte-Tigges G, Konrad M, Bardeleben A, Werner C: Robot-assisted arm trainer for the passive and active practice of bilateral forearm and wrist movements in hemiparetic subjects. Arch Phys Med Rehabil 2003, 84: 915-920. 10.1016/S0003-9993(02)04954-7View ArticlePubMedGoogle Scholar
- Macclellan LR, Bradham DD, Whitall J, Volpe B, Wilson PD, Ohlhoff J, Meister C, Hogan N, Krebs HI, Bever CT Jr: Robotic upper-limb neurorehabilitation in chronic stroke patients. J Rehabil Res Dev 2005, 42: 717-722. 10.1682/JRRD.2004.606.0068View ArticlePubMedGoogle Scholar
- Rose DK, Winstein CJ: Bimanual training after stroke: Are two hands better than one? Top Stroke Rehabil 2004, 11: 20-30.View ArticlePubMedGoogle Scholar
- Whitall J, McCombe Waller S, Silver KH, Macko RF: Repetitive bilateral arm training with rhythmic auditory cueing improves motor function in chronic hemiparetic stroke. Stroke 2000, 31: 2390-2395. 10.1161/01.STR.31.10.2390View ArticlePubMedGoogle Scholar
- Cunningham CL, Stoykov ME, Walter CB: Bilateral facilitation of motor control in chronic hemiplegia. Acta Psychol (Amst) 2002, 110: 321-337. 10.1016/S0001-6918(02)00040-9View ArticleGoogle Scholar
- Chang JJ, Tung WL, Wu WL, Huang MH, Su FC: Effects of robot-aided bilateral force-induced isokinetic arm training combined with conventional rehabilitation on arm motor function in patients with chronic stroke. Arch Phys Med Rehabil 2007, 88: 1332-1338. 10.1016/j.apmr.2007.07.016View ArticlePubMedGoogle Scholar
- Cauraugh JH, Kim SB, Summers JJ: Chronic stroke longitudinal motor improvements: cumulative learning evidence found in the upper extremity. Cerebrovasc Dis 2008, 25: 115-121. 10.1159/000112321View ArticlePubMedGoogle Scholar
- Cauraugh JH, Coombes SA, Lodha N, Naik SK, Summers JJ: Upper extremity improvements in chronic stroke: coupled bilateral load training. Restor Neurol Neurosci 2009, 27: 17-25.PubMed CentralPubMedGoogle Scholar
- Stinear JW, Byblow WD: Rhythmic bilateral movement training modulates corticomotor excitability and enhances upper limb motricity poststroke: a pilot study. J Clin Neurophysiol 2004, 21: 124-131. 10.1097/00004691-200403000-00008View ArticlePubMedGoogle Scholar
- Mudie MH, Matyas TA: Responses of the densely hemiplegic upper extremity to bilateral training. Neurorehabil Neural Repair 2001, 15: 129-140. 10.1177/154596830101500206View ArticlePubMedGoogle Scholar
- Latimer CP, Keeling J, Lin B, Henderson M, Hale LA: The impact of bilateral therapy on upper limb function after chronic stroke: a systematic review. Disabil Rehabil 2010, 32: 1221-1231. 10.3109/09638280903483877View ArticlePubMedGoogle Scholar
- Cauraugh JH, Lodha N, Naik SK, Summers JJ: Bilateral movement training and stroke motor recovery progress: A structured review and meta-analysis. Hum Mov Sci 2010, 29: 853-870. 10.1016/j.humov.2009.09.004PubMed CentralView ArticlePubMedGoogle Scholar
- Kadivar Z, Sung C, Thompson Z, O'Malley M, Liebschner M, Deng Z: Comparison of Reaching Kinematics During Mirror and Parallel Robot Assisted Movements. Medicine Meets Virtual Reality 2011, 18: 247-253.Google Scholar
- Luft AR, McCombe-Waller S, Whitall J, Forrester LW, Macko R, Sorkin JD, Schulz JB, Goldberg AP, Hanley DF: Repetitive bilateral arm training and motor cortex activation in chronic stroke: a randomized controlled trial. JAMA 2004, 292: 1853-1861. 10.1001/jama.292.15.1853PubMed CentralView ArticlePubMedGoogle Scholar
- Summers JJ, Kagerer FA, Gary MI, Hiraga CY, Loftus A, Cauraugh JH: Bilateral and unilateral movement training on upper limb function in chronic stroke patients: A TMS study. J Neurol Sci 2007, 252: 76-82. 10.1016/j.jns.2006.10.011View ArticlePubMedGoogle Scholar
- Burgar C, Lum P, Shor P, Van der Loos H: Development of Robots for Rehabilitation Therapy: The Palo Alto VA/Stanford Experience. J Rehabil Res Dev 2000, 376: 663-673.Google Scholar
- Hesse S, Werner C, Pohl M, Rueckriem S, Mehrholz J, Lingnau ML: Computerized arm training improves the motor control of the severely affected arm after stroke: a single-blinded randomized trial in two centers. Stroke 2005, 36: 1960-1966. 10.1161/01.STR.0000177865.37334.ceView ArticlePubMedGoogle Scholar
- Lewis GN, Byblow WDB: Bimanual coordination dynamics in poststroke hemiparetics. J Mot Behav 2004, 36: 174-188. 10.3200/JMBR.36.2.174-188View ArticlePubMedGoogle Scholar
- Staines WR, McIlroy WE, Graham SJ, Black SE: Bilateral movement enhances ipsilesional cortical activity in acute stroke: a pilot functional MRI study.[see comment]. Neurology 2001, 56: 401-404. 10.1212/WNL.56.3.401View ArticlePubMedGoogle Scholar
- Byblow WD, Stinear CM, Smith MC, Bjerre L, Flaskager BK, McCambridge AB: Mirror symmetric bimanual movement priming can increase corticomotor excitability and enhance motor learning. PLoS One 2012, 7: e33882. 10.1371/journal.pone.0033882PubMed CentralView ArticlePubMedGoogle Scholar
- Bogaerts H, Swinnen SP: Spatial interactions during bimanual coordination patterns: the effect of directional compatibility. Mot Control 2001, 5: 183-199.Google Scholar
- Oldfield R: The assessment and analysis of handedness: the Edinburgh Inventory. Neuropsychologia 1971, 9: 97-113. 10.1016/0028-3932(71)90067-4View ArticlePubMedGoogle Scholar
- Scharver C, Evenhouse R, Johnson A, Leigh J: Pre-surgical cranial Implant design using the PARIS_prototype. IEEE Virtual Reality 2004; MARCH 27, 2004 through Wednesday, MARCH 31, 2004; Chicago 2004.Google Scholar
- Fitts PM: Perceptual-motor skill learning. In Categories of human learning. Edited by: MeHou AW. New York: Academic; 1964:243-245.Google Scholar
- Thoroughman KA, Shadmehr R: Learning of action through adaptive combination of motor primitives. [see comments]. Nature 2000, 407: 742-747. 10.1038/35037588PubMed CentralView ArticlePubMedGoogle Scholar
- Everitt BS (Ed): In Cambridge Dictionary of Statistics. CUP. 2nd edition. 2002.Google Scholar
- Krakauer JW, Pine ZM, Ghilardi MF, Ghez C: Learning of visuomotor transformations for vectorial planning of reaching trajectories. Journal of Neuroscience (Online) 2000, 20: 8916-8924.Google Scholar
- Tong C, Flanagan JR: Task-specific internal models for kinematic transformations. J Neurophysiol 2003, 90: 578-585. 10.1152/jn.01087.2002View ArticlePubMedGoogle Scholar
- Miles FA, Eighmy BB: Long-term adaptive changes in primate vestibuloocular reflex I: Behavioral observations. J Neurophysiol 1980, 43: 1406-1425.PubMedGoogle Scholar
- Swinnen S, Wenderoth N: Two hands, one brain: cognitive neuroscience of bimanual skill. Trends Cogn Sci 2004, 8: 18-25. 10.1016/j.tics.2003.10.017View ArticlePubMedGoogle Scholar
- Debaere F, Wenderoth N, Sunaert S, van Hecke P, Swinnen SP: Internal vs external generation of movements: differential neural pathways involved in bimanual coordination performed in the presence or absence of augmented visual feedback. NeuroImage 2003, 19: 764-776. 10.1016/S1053-8119(03)00148-4View ArticlePubMedGoogle Scholar
- Altschuler EL, Wisdom SB, Stone L, Foster C, Galasko D, Llewellyn DM, Ramachandran VS: Rehabilitation of hemiparesis after stroke with a mirror. Lancet 1999, 353: 2035-2036. 10.1016/S0140-6736(99)00920-4View ArticlePubMedGoogle Scholar
- Sathian K, Greenspan AI, Wolf SL: Doing it with mirrors: a case study of a novel approach to neurorehabilitation. Neurorehabil Neural Repair 2000, 14: 73-76. 10.1177/154596830001400109View ArticlePubMedGoogle Scholar
- Yavuzer G, Selles R, Sezer N, Sütbeyaz S, Bussmann JB, Köseoğlu F, Atay MB, Stam HJ: Mirror therapy improves hand function in subacute stroke: a randomized controlled trial. Arch Phys Med Rehabil 2008, 89: 393-398. 10.1016/j.apmr.2007.08.162View ArticlePubMedGoogle Scholar
- Nojima I, Mima T, Koganemaru S, Thabit MN, Fukuyama H, Kawamata T: Human motor plasticity induced by mirror visual feedback. J Neurosci 2012, 32: 1293-1300. 10.1523/JNEUROSCI.5364-11.2012View ArticlePubMedGoogle Scholar
- Guadagnoli MA, Lee TD: Challenge point: a framework for conceptualizing the effects of various practice conditions in motor learning. J Mot Behav 2004, 36: 212-224. 10.3200/JMBR.36.2.212-224View ArticlePubMedGoogle Scholar
- Atkeson G, Hollerbach JM: Kinematic features of unrestrained vertical arm movements. J Neurosci 1985, 5: 2318-2330.PubMedGoogle Scholar
- Balasubramanian R, Howe RD, Matsuoka Y: Task performance is prioritized over energy expenditure. IEEE Trans Biomed Eng 2009, 56: 1310-1317.View ArticlePubMedGoogle Scholar
- Lum SP, Lehman SL, Reinkensmeyer DJ: The bimanual lifting rehabilitator: an adaptive machine for therapy of stroke patients. IEEE Trans Rehabil Eng 1995, 3: 166-174. 10.1109/86.392371View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.