The authors have declared that no competing interests exist.
Conceived and designed the experiments: PKM RLS. Performed the experiments: PKM. Analyzed the data: PKM. Contributed reagents/materials/analysis tools: PKM KYH RLS. Wrote the paper: PKM KYH RLS.
Motor lateralization in humans has primarily been characterized as “handedness”, resulting in the view that one arm-hemisphere system is specialized for all aspects of movement while the other is simply a weaker analogue. We have proposed an alternative view that motor lateralization reflects proficiency of
The modern view of brain lateralization, built upon early work in patients with unilateral brain damage and more recent work in split-brain patients
Studies of lateralization of cognitive and perceptual processes have supported the notion that each hemisphere contributes unique mechanisms to the control of a given function. For example, language comprehension recruits the left hemisphere for lexical, semantic and syntactic processing, and the right hemisphere for processing its emotional and non-verbal features such as prosody
Unfortunately, motor lateralization, which may be defined as a difference in motor performance between the two arms, has yet to be appreciated from this perspective. Instead, motor lateralization has primarily been conceptualized in terms of “handedness”, assessed in two ways – first, by noting the preferred arm for doing tasks and second, by examining performance differences between arms on a particular task. Both these approaches have led to a view that one arm (at a population level, often the right arm) is specialized for all aspects of movement, while the other is simply a weaker analogue. Theories postulating the origin of such “lateralization” have ranged from cultural and social influences during development
Prior studies have attempted to describe the mechanistic basis of interlimb performance differences based primarily on the distinction between feedforward and feedback modes of control. For instance, Roy, Elliott and colleagues noted that the dominant arm system demonstrates an advantage in the speed of processing visual and proprioceptive feedback, and proposed a specialization for feedback control for this system
We have introduced an alternative hypothesis that motor lateralization is the result of a specialization of each arm-hemisphere system for distinct and complementary motor control mechanisms
Here we provide a clear test of our hypothesis by using a task that directly probes the control mechanisms that result in interlimb differences in motor performance. In this task we occasionally and covertly shifted the starting location of the hand as healthy right-handed adults performed targeted reaching movements in a virtual-reality environment
The institutional review board of the New Mexico Veterans Affairs Healthcare System approved the study. All participants gave written informed consent prior to testing according to the principles expressed in the Declaration of Helsinki.
Participants were young, healthy, right-handed adults (n = 14, mean age = 23.35 yrs). Handedness was determined using a 10-item version of the Edinburgh inventory
Subjects sat facing a table with their forearm supported over the table using an air sled system. A cursor (diameter = 0.8 cm) representing the position of the index finger tip, a start circle (diameter = 1 cm) and targets (diameter = 2 cm) were projected using a horizontally mounted HDTV onto a mirror placed beneath it. The mirror blocked direct vision of the subjects arm, but reflected the visual display to give the illusion that the display was in the same horizontal plane as the fingertip. Subjects performed reaching movements on the tabletop below the mirror. Position and orientation of the forearm and upper-arm segments were sampled using a Flock of Birds system (Ascension Technology). The positions of the index finger tip, the lateral epicondoyle of the humerus and the acromion were computed using two Flock of Birds markers per arm and recorded using custom software, with the X-Y plane parallel to the tabletop. We used the computed X-Y coordinates of the fingertip to define the projected cursor position. A bib running from the subjects’ neck to the edge of the mirror was used to block the view of the shoulder and upper arm.
Each participant performed the task with both the left and the right arm; the starting arm was counterbalanced across subjects. Two blocks of movements were performed with the same arm in succession and then the arm was switched. Each block consisted of 200 movements to a single target located 15 cm from the start circle, either 45 degrees (“Lateral” movement direction) or 135 degrees (“Medial” movement direction) relative to the horizontal. The task was done in a blocked order to ensure the consistency of movements to every target.
The first 40 movements of each block were baseline movements, during which the on-screen cursor position matched the position of the hand. Prior to each trial, the start circle was displayed on the screen and subjects were asked to bring their hand (cursor) into it. After a brief delay, the target for that trial appeared along with an audio-visual “go” signal, which served as the cue for subjects to reach to the target in a single, uncorrected, rapid motion. Cursor feedback was eliminated at this time. Velocity feedback was provided and subjects were encouraged to attain a peak speed of at least 0.5 m/s. One, three or ten point(s) were given based on movement accuracy if this speed requirement was met. Between trials, the cursor was shown only when the index fingertip was within 4.5 cm from the center of the start circle. This was done to prevent subjects from consciously perceiving the altered conditions during probe trials (see below).
After the 40 baseline movements at the beginning of each block, we altered the relationship between the hand and cursor position on occasional “probe” trials. On these trials, which were pseudo-randomly interspersed within the remaining baseline trials, the location of the on-screen cursor was displaced from that of the hand. Importantly however, visually, the task remained exactly identical to the baseline trials in that subjects still brought the cursor into the start circle to initiate the trial. However, on these trials, unbeknownst to subjects, their hand was positioned outside the start circle. Post-testing, subjects reported being unaware of this manipulation. Maximum points were awarded on these trials regardless of movement accuracy.
For each target direction, four different probe start locations were used (
Note that shifted start positions were “shared” between the targets. For example, the anterior start position for the lateral movement served as the top start position for the medial movement. The baseline start position was the same for the lateral and medial movement directions.
All recorded data were low-pass filtered at 12 Hz (third order dual pass Butterworth), and angular kinematic data were differentiated to yield velocity values. The first 20 movements of each block were considered practice and were not analyzed. Probe trials on which subjects failed to move or made extremely curved (almost circular) movements were also excluded after they were identified using outlier Box plots. In total, 8 trials (4 each from the top and bottom start locations) were excluded from a total of 448 probe trials across arms, movement directions and subjects.
We identified movement onset by noting the time of peak velocity and searching backwards in time for the first minimum in velocity below 8% of peak tangential velocity. Movement end was determined by searching forward from peak velocity to find the first minimum below 8% of the peak. Our primary measure of interest was direction error at movement end. For baseline movements, this was defined as the angular difference between the line connecting the center of the start circle and the target, and the line connecting movement start and end points. For the orthogonal probe trials, direction error was calculated relative to a straight line originating from the shifted start location parallel to the baseline movement direction. Counterclockwise direction errors were considered positive, while clockwise errors were considered negative. For each subject, we normalized the direction errors on probe trials by subtracting out the mean baseline direction error. We similarly calculated initial direction errors, defined as the angular difference between the line connecting the start position and the target, and the line connecting the hand locations at movement start and at peak acceleration. For probe trials, these errors were normalized by subtracting the mean initial direction error on baseline movements. We also calculated the position error perpendicular to the baseline target direction. This measure gave us a better estimate of closeness to the baseline target than absolute final position error, which also takes into account the overshoot or undershoot in movement along the target direction. We also used the absolute final position error measure for comparison, and calculated it as the distance between the finger position at movement end and the center of the target. In addition, we computed movement distance as the straight-line distance between movement start and end points. We normalized movement distance for each subject by subtracting the mean extent of baseline movements. For averaging of hand trajectories, the following method was used: first, the X and Y hand-displacement profiles were time normalized, then decimated to 100 points. Then, each series of X and Y displacement profiles were point averaged to yield a mean and SE value for each consecutive point. The mean X and Y values were plotted against each other to yield a mean handpath profile. The SE for X and Y displacements were displayed as horizontal, and vertical error bars respectively.
For statistically comparing the performance of the two arms, we used paired Wilcoxon signed rank tests, pooled across perturbed start positions (top and bottom). For direction errors, absolute values were used during this comparison. Our choice of non-parametric tests was motivated by the fact that the number of probe trials per start location per subject was small (by design) and therefore, the data tended not to be normally distributed.
In order to numerically compare the arms in terms of their trajectory differences, we computed the direction of the trajectories at the beginning and end of movement as an angular error relative to a straight line parallel to the baseline target direction, but originating from the perturbed start positions. There were no significant differences between the arms in terms of the mean initial movement direction (p = 0.3596) or its variability (p = 0.1311). However, we found significantly greater direction errors at movement end (
A similar pattern of results was observed for movements made away from the body midline (“Lateral” movement direction). The mean baseline handpaths of both arms (thick black lines) for the subject in
While initial movement direction was not different between the arms (p = 0.3808), its variability was larger in the non-dominant arm (p = 0.0190). However, our critical measure of direction error at movement end, shown in
Initial movement direction was not significantly different between the arms for probe trials from the anterior and posterior start positions for either the medial (p = 0.7966) or the lateral (p = 0.1261) movement direction. Further the variability in initial movement direction was also not significantly different between the arms in both the medial (p = 0.2409) and the lateral (p = 0.10) directions. This interlimb similarity in movement direction was maintained at movement end, as shown in the left panels of
This study was motivated by the fact that prior studies have not provided a clear understanding of the neural mechanisms underlying lateralization within the motor system. Several previous studies have quantified performance differences between the two arms
It is important to first discuss the lack of interlimb differences in movements from the anterior and posterior start positions. Movement directions at the beginning and end of motion, as well as the magnitude of shortening and lengthening of movement extent in response to these shifts in start location, were similar between the arms. Our prior work using a similar paradigm
In contrast to movements from the collinear start positions, our results indicated clear differences in movements initiated from the orthogonally shifted start locations. It is important to point out that these findings cannot be explained solely on the basis of a distinction between feedforward or open loop and feedback or closed loop control mechanisms. Applied to our study design, this hypothesis does not make specific predictions regarding the mean final positions of the two arms on probe trials, but suggests larger variability for the non-dominant and dominant arms at the beginning and the end of the movement respectively. We did not find this to be the case. Mean final positions were clearly distinct between the arms on orthogonally perturbed probe trials. In addition, non-dominant arm variability was slightly (but statistically significantly) greater at the beginning of the movement only for movements made in the lateral direction, and contrary to the feedforward/feedback hypothesis, continued to be so even at movement end. Further, variability of movements made in the medial direction was not different between the arms either at the beginning or at the end. Thus, our data suggests a mode of control that appears to be distinct from a general feedforward versus feedback specialization for the dominant and non-dominant arm-hemisphere systems.
We made two key observations in the movement patterns on these probe trials. First, we did not observe any differences in movement direction during the early phases of the movement. Second, we did not observe complete convergence of non-dominant arm trajectories onto the baseline target. The angular deviation was about 60% of that required to land the arm exactly on that target. Similarly, dominant arm movements, particularly in the medial direction, were not completely parallel to baseline trajectories. A recent computational model proposed by Yadav and Sainburg
Our framework thus emphasizes bilateral hemispheric contribution to movements of each arm. Such bi-hemispheric control is consistent with observations from previous neuroimaging studies that cortical areas of both brain hemispheres are active during unilateral arm movements
Another important observation in the current study was that non-dominant arm movements in the medial direction tended to be longer than those of the dominant arm on probe trials, but not on baseline trials. One possible explanation for these findings is that the misalignment between the visually and proprioceptively signaled start positions lead to an incorrect estimate of the actual hand position, which, in turn produces errors in movement extent
Prior studies have almost always viewed motor lateralization as “handedness” (either hand/arm preference or greater “skill” in one hand/arm across tasks) rather than arising from specialization for each arm-hemisphere system for different aspects of control within “right-” or “left-handed” individuals. Our findings question this traditional view of handedness, because specialized mechanisms for
Thanks to Melissa Daniels and Lee Stapp for help with data collection, and Sierra Widmer for assistance with data analysis.