Effects of spatial-memory decay and dual-task interference on perturbation-evoked reach-to-grasp reactions in the absence of online visual feedback

Human Movement Science 32 (2013) 328–342

Contents lists available at SciVerse ScienceDirect

Human Movement Science journal homepage: www.elsevier.com/locate/humov

Effects of spatial-memory decay and dual-task interference on perturbation-evoked reach-to-grasp reactions in the absence of online visual feedback Kenneth C. Cheng a,b,c, Jay Pratt b,d, Brian E. Maki a,b,c,e,f,⇑ a

Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Canada Toronto Rehabilitation Institute, University of Toronto, Toronto,Canada c Institute of Medical Science, University of Toronto, Toronto, Canada d Department of Psychology, University of Toronto, Toronto, Canada e Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, Canada f Department of Surgery, University of Toronto, Canada b

a r t i c l e

i n f o

Article history: Available online 29 April 2013 PsycINFO classification: 2330 Keywords: Arm movements Postural balance Triggered reactions Dual-task interference Spatial working memory

a b s t r a c t Recent findings suggest that rapid perturbation-evoked reach-tograsp balance-recovery reactions can be (and often are) guided by visuospatial information stored in working memory. To further our understanding, the present study examined the influence of memory-decay and concurrent cognitive-task performance on the speed, accuracy and effectiveness of these reactions by using liquid-crystal goggles to initiate occlusion of vision at various ‘‘recall-delay’’ times prior to perturbation-onset, in ten healthy young-adults. A small handhold was moved unpredictably to one of four locations 2 s prior to vision-occlusion; reactions to recover balance by grasping the handhold were evoked by unpredictable antero-posterior platform-translation perturbations. Recall-delay time (0 s/2 s/5 s/10 s) was randomized, and subjects performed a spatial- or non-spatial-memory task during the delay-time in a subset of trials. Consistent with studies of volitional reach-tograsp, recall-delay led to some reduction in endpoint accuracy; however, unlike those studies, the present results showed no evidence that recall-delay led to slowing of the arm movement. Both spatial and non-spatial cognitive tasks had similar effects (slowing of movement initiation and execution), suggesting these effects were related to generic attentional demands rather than competition for specific resources related to spatial working

⇑ Corresponding author. Address: Toronto Rehabilitation Institute, Room 12-121, 550 University Avenue, Toronto, Canada M5G 2A2. Tel.: +1 (416) 597 3422x7808. E-mail address: [email protected] (B.E. Maki). 0167-9457/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.humov.2012.11.001

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

329

memory. Further work is needed to determine effects of agerelated impairments in visuospatial memory and attentional capacity. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction Rapid reach-to-grasp reactions are often executed in response to sudden loss of postural balance (Bateni, Zecevic, McIlroy, & Maki, 2004; Maki & McIlroy, 1997, 2005; Maki, Perry, & McIlroy, 1998; Marigold & Misiaszek, 2009; McIlroy & Maki, 1995). Visuospatial information (VSI) regarding the ‘‘target’’ (handhold) location is required for successful reach-to-grasp (Jeannerod, 1988). Studies of volitional goal-directed arm movements have shown that visual fixation of the target is typically used to guide the reaching movement during natural behavior (Abrams, 1992; Carnahan & Marteniuk, 1991; Haycoe & Ballard, 2005; Land, 2006; Prablanc, Echallier, Jeannerod, & Komilis, 1979). When reacting to sudden unpredictable loss of balance, however, the urgent need to grasp the handhold very rapidly in order to prevent a fall imposes temporal constraints that may limit the capacity to identify a suitable target and to execute a saccade to that target in ‘‘real time’’, i.e., after the onset of the balance perturbation (King et al., 2011; Maki & McIlroy, 2005). Recent studies of arm reactions evoked by a truly-unexpected balance perturbation while ambulating in an unfamiliar environment have, in fact, shown that the reach-to-grasp movements were invariably executed in the absence of concurrent visual fixation of the handrail (King et al., 2011, 2009). Most subjects did, however, fixate briefly on the handrail one or more times upon first entering the environment. This has led to suggestions that the CNS may guide these arm reactions using VSI stored in spatial working memory. The needed VSI, regarding the location of potential handholds (as well as other salient objects), would be acquired and stored automatically as a contingency, through natural exploratory gaze behavior, as the person enters or moves through the environment (King et al., 2011, 2009). This stored-VSI, in combination with multi-sensory feedback regarding the perturbationinduced body motion, would then allow the hand to be moved very rapidly toward the nearest available handhold, if and when an unexpected balance perturbation occurs. The capacity to use stored-VSI to guide effective reach-to-grasp reactions is supported by recent studies in which liquid-crystal goggles were used to occlude vision at time of balance-perturbation onset (Cheng, McKay, King, & Maki, 2012a, 2012b; Ghafouri, McIlroy, & Maki, 2004). Although it was found that reach-accuracy and grip-formation were somewhat impaired when dependent on stored-VSI, subjects were generally well able to achieve a functionally-adequate grasp and prevent themselves from falling (Cheng et al., 2012a, 2012b). However, those studies did not examine the situation where the VSI must be retained in visuospatial memory for some interval of time prior to perturbation onset. Such a situation could occur in daily life, for example, when potential handhold locations are mapped via natural exploratory gaze behavior upon first entering an environment but the balance perturbation occurs several seconds later. Studies of volitional arm movements have indicated that the accuracy of the VSI stored in visuospatial working memory can decay rapidly (Elliott & Madalena, 1987; Hesse & Franz, 2010; Hu, Eagleson, & Goodale, 1999). In support of this, even modest demands for memory retention (e.g., a recall-delay time of 2 s) led to undershoot error and endpoint variability, as well as slowed initiation and execution of the arm movement (Lemay & Proteau, 2002; Westwood, Heath, & Roy, 2003). In addition, such effects are likely to be exacerbated when memory retention and retrieval are disrupted by a concurrent visuospatial cognitive task (Fougnie & Marois, 2006; Hale, Myerson, Rhee, Weiss, & Abrams, 1996; McAfoose & Baune, 2009). Although the findings related to endpoint accuracy are also likely to apply to reach-to-grasp reactions evoked by sudden unpredictable balance perturbation, the aforementioned temporal constraints that govern these rapid reactions could well preclude any significant slowing of the response. The consequent speed-accuracy trade-off would be expected to further reduce endpoint accuracy, and could ultimately compromise the capacity to achieve a functional grasp.

330

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

The objectives of the present study were to: 1) examine the effects of memory decay on the capacity to use stored-VSI to help guide rapid reach-to-grasp balance-recovery reactions, and 2) determine whether the memory-decay effects were exacerbated when concurrent cognitive tasks were performed. As in previous studies (Cheng et al., 2012a, 2012b), we used liquid-crystal goggles to force reliance on stored-VSI and randomly varied the location of the handhold between trials; however, rather than simply occluding vision at time of perturbation onset, we now randomly varied the length of time (‘‘recall-delay time’’) that the handhold target location had to be retained in memory by occluding vision either 0 s, 2 s, 5 s or 10 s prior to perturbation onset. In a subset of trials, subjects performed a concurrent spatial or non-spatial cognitive memory task during the memory-retention interval prior to perturbation onset. Similar to effects previously observed in studies of volitional reaching, we hypothesized that increase in recall-delay time would lead to increase in both variable and systematic error in the landing position of the hand, and that this would be accompanied by an increased frequency of impaired prehension (e.g., hand-handhold collision, incomplete grasp formation). However, in view of the temporal constraints that govern balance-recovery reactions, we hypothesized that previously-observed slowing in reaction-onset time, movement-time and handhold-contact time would not occur as a consequence of increase in recall-delay time. We further hypothesized that the adverse effects on the accuracy of the grasping reaction would be exacerbated when performing a secondary task that interfered with spatial working memory, in comparison to trials involving a non-spatial secondary task or no secondary task at all. 2. Materials and methods 2.1. Participants Ten right-handed healthy young-adults (five male, five female; aged 22–30, mean = 26; mass 47– 98 kg, mean = 68 kg; height 155–184 cm, mean = 169 cm) participated after signing informed consent to comply with ethics approval granted by the institutional review board. They were required to have a minimum Snellen visual acuity of 20/40 (uncorrected, or corrected with contact lenses; volunteers who needed to wear spectacles were excluded, due to the need to wear the liquid-crystal goggles described below). None of the subjects had participated in previous balance studies, and none reported any neural, sensorimotor or musculoskeletal impairments, medical conditions or medication use affecting control of balance or limb movement. 2.2. Protocol Balance-recovery reach-to-grasp reactions were evoked by sudden forward (0.12 m, 0.41 m/s, 1.4 m/s2) or backward (0.18 m, 0.6 m/s, 2.0 m/s2) translation of a 2 m  2 m computer-controlled motion-platform (Maki, McIlroy, & Perry, 1996). Each platform translation comprised an approximatelysquare 300 ms acceleration pulse followed immediately by an equal and opposite deceleration pulse. In each trial, a motor-driven device (Cheng et al., 2009) mounted on the platform controlled a cylindrical handhold (length = 20 cm, diameter = 3.8 cm) to move along a transverse axis in front of the subject (distance from handhold to back of heels = 33% of body height; handhold height = 60% of body height) and to stop unpredictably at one of four locations (0%, 33%, 67% or 100% of shoulder width (SW) to the right of the mid-sagittal plane; Fig. 1). The focus of the study was on the grasping reactions evoked by forward platform translation (backward falling motion), in trials where the handhold was positioned at 33% of shoulder-width from the mid-sagittal plane. Forward platform translations with other handhold positions (21% of trials), backward translations (32% of trials) and motionless ‘‘catch’’ trials (16% of trials) were randomly interspersed throughout the testing session, so as to increase unpredictability and deter anticipatory reactions or other proactive strategies, but these trials were not analyzed. Subjects were instructed to recover balance by grasping a marked (with red tape) ‘‘target’’ section of the handhold (length = 125% of hand-width) as quickly as possible after perturbation-onset with an

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

331

Fig. 1. Motion platform and moveable handhold systems. Dashed lines in the insert indicate the four handhold positions that were tested. Analysis focused on the highlighted handhold position (33% shoulder-width⁄) and anterior platform translations (backward falling motion).

overhand grasp, using only their right hand. Instructions not to move the feet, plus foam-rubber barriers (height = 30 cm), were used to deter foot motion and force reliance on reach-to-grasp reactions. To provide moderate visual contrast, the handhold was covered with medium-gray grip tape, while the visual background (floor and wall) was black, with an ambient room-illumination level of 230 lux. For safety, subjects wore a harness designed to prevent impact between body and floor without restricting movement or providing somatosensory feedback that could aid in control of balance. Padded gloves and wrist guards were worn to reduce impact to the hands and wrist if and when collision to the back of hand or wrist occurred. Translucent liquid-crystal goggles (Translucent Technologies Inc., Toronto, ON) were used to occlude vision during a portion of each trial. The goggles were custom-modified to hold the liquid-crystal element flush against the orbit of each eye, so as to completely block both central and peripheral vision when activated (opaque) and to permit near-complete field-of-view when de-activated (Scovil, Zettel, & Maki, 2008). At the start of each trial, vision was first occluded for 2 s (interval T1 in Fig. 2A) while the movable handhold moved to and stopped at one of the four handhold positions. Vision was then allowed for 2 s (interval T2), during which subjects were instructed to look at the handhold without turning their head. Vision was then occluded again, for a recall-delay interval (T3) of 0 s, 2 s, 5 s or 10 s prior to perturbation-onset, and continued to be occluded until the end of the trial (i.e., until 2 s after perturbation-onset). For the trials with recall-delays of 5 s and 10 s, subjects performed either a spatial-memory task, a non-spatial-memory task, or no secondary task during the recall-delay interval (T3). The non-spatial (mental arithmetic) task (Logie, Gilhooly, & Wynn, 1994) required subjects to sequentially add a series of auditorily-presented random numbers and to report the final sum after the end of the trial (Fig. 2B), whereas the spatial-task (a modified Brooks spatial task (Brooks, 1967)) required subjects to imagine sequential movements (up, down, left, right) of a highlighted square within an N  N matrix and to report the final position of the cell within the matrix (Fig. 2C). A different instruction sequence of numbers (to add) or movements (to imagine) was randomly generated in each trial, with a new instruction delivered (via headphones) every 1.25 s, and the last instruction beginning 1.25 s prior to perturbation onset. Each instruction took 400– 600 ms to deliver. Subjects practiced the cognitive tasks for 20–30 min prior to balance-perturbation testing, while sitting in front of a computer. Starting from the easiest level (adding random numbers ranging from

332

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

Fig. 2. A. Sequence and timing of events: liquid-crystal goggles occluded vision for 2 s (interval T1) at the start of each trial, while the movable handhold moved to and stopped at one of four possible locations. Vision was then allowed for 2 s (T2) and subjects were instructed to look at the handhold during this viewing period. Vision was then occluded again, for a recall-delay interval of 0 s, 2 s, 5 s or 10 s (T3) prior to perturbation-onset (PO), and continued to be occluded for 2 s after PO (T4), i.e., until the end of trial. For trials with recall delays of 5 s or 10 s, subjects performed either a spatial-memory task, a non-spatialmemory task or no secondary task during the recall-delay interval (T3). B. For the non-spatial-memory task, subjects were asked to add a series of auditorily-presented random numbers and to report the final sum after the end of the trial. Difficulty of the task was adjusted by changing the range of numbers (i.e., 1-to-3, 1-to-5, 1-to-9, or 4-to-12) to be added. Starting at the number ‘‘3’’, subjects were instructed to sequentially add a series of random numbers delivered (via headphones) every 1.25 s. As an example, the figure shows the correct response (3 + 5 + 3 + 2 + 4 = 17) to the sequence of numbers shown (5, 3, 2, 4). C. For the spatial-memory task, subjects were instructed to imagine a highlighted square moving around in an N  N matrix and to report the final position of the highlighted square within the matrix. Starting at or near the center cell, a random verbal command to move up, down, left or right was given every 1.25 s. As an example, the gray arrows in the figure show the correct responses to a sequence of commands to move up, right, right and down, after starting at cell ‘‘F’’. Subjects were shown the matrix after each trial and asked to identify the correct final response (cell ‘‘H’’ in the example shown).

one to three, or imagining a highlighted box moving around in a 33 matrix), task difficulty level was gradually increased (i.e., adding larger numbers or imagining a larger matrix). The objectives were to: 1) find task conditions that were challenging for each individual, but were not so difficult that subjects would ‘‘give up’’; 2) match the difficulty of the two tasks within each subject; and 3) match task difficulty between subjects. To this end, the task difficulty was adjusted for each subject so that the correct ‘‘answer’’ was reported in 70–90% of trials. After finding comparable difficulty levels, baseline error-rates were established, for task durations of 5 s and 10 s (see Table 1). At the start of each trial, subjects were informed of the secondary task, and were instructed to face forward, with arms resting at sides and the hands forming a relaxed fist with thumb ‘‘on top’’. To motivate subjects to reach for the handhold as quickly as possible in response to the platform motion, they

Table 1 Cognitive task performance. Subject

1 2 3 4 5 6 7 8 9 10  

Non-spatial-memory task

Spatial memory task

Difficulty 

Single–

Dual–

D Error %§

Difficulty 

Single–

Dual–

D Error %§

1-to-5 1-to-5 1-to-5 1-to-5 1-to-5 4-to-12 1-to-5 1-to-5 1-to-3 1-to-9

2/10 3/10 0/10 1/10 3/10 2/10 3/10 1/10 1/10 2/10

11/20 3/20 2/20 5/20 3/20 5/20 5/20 2/20 3/20 7/20

+35 15 +10 +15 15 +5 5 0 +5 +15

66 44 66 66 77 66 66 66 66 66

2/10 2/10 1/10 2/10 2/10 2/10 1/10 0/10 1/10 1/10

5/20 1/20 3/20 4/20 6/20 2/20 5/20 4/20 4/20 5/20

+5 15 +5 0 +10 10 +15 +20 +10 +15

Non-spatial task difficulty was adjusted by changing the range of possible numbers to be summed (see Fig. 2B); spatialmemory task difficulty was adjusted by changing the N  N matrix size (see Fig. 2C). – Proportion of trials with error during single-task trials (no balance perturbations) and dual-task trials. § Difference in percentage error rate (dual-task error rate minus single-task error rate).

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

333

were told that a $50 prize would be awarded to the subject who achieved the quickest average response time. They were also told that premature initiation of arm movements (prior to onset of platform movement) or errors in performing the cognitive task would result in a penalty that reduced their chances of winning the $50 prize. For the no-secondary-task trials, the protocol included three ‘‘focus’’ trials (forward platform translation, handhold positioned at 33% of body-height from midline) at each of the four different recalldelay times (0 s, 2 s, 5 s and 10 s). For the spatial and non-spatial dual-task trials, testing was limited to recall-delays of either 5 s or 10 s (since shorter recall-delay intervals would not allow sufficient time to perform the cognitive task), and three ‘‘focus’’ trials were performed for each of the four combinations (two tasks  two delay times). In addition, as noted earlier (and detailed in Table 2), trials involving a number of other perturbation and handhold locations were also used to increase unpredictability. To minimize mental fatigue and adaptive effects, the order of testing the trials was completely randomized for each subject. To further reduce adaptive effects, ten practice perturbation trials were performed prior to the start of the experiment (three spatial-task, three non-spatial-task, and four no-secondary-task trials). 2.3. Data collection and analysis Video recordings from four cameras were used to determine which arm was used to grasp the handhold, whether a full grasp was achieved (all digits wrapped around the handhold), whether a collision error occurred (contact with wrist, or back of hand or digits), and whether the subject attempted to step (by kicking the foam-rubber barriers) or fell into the safety harness (confirmed by load cell on harness cable). Surface electrodes were used to record electromyographic (EMG) activity in the anterior deltoid, lateral deltoid and biceps muscles (band-pass filtered, 10–500 Hz; samplingrate = 1000 Hz). Reaction-time was defined as the earliest EMG onset latency in any of these muscles, as determined by a computer algorithm (McIlroy & Maki, 1993) and confirmed by visual inspection. Contact-time was detected by force-sensing resistors mounted on the front, back and top of the

Table 2 Distribution of trials (tested in random order).

Note: ‘‘focus trials’’ = trials included in the analyses: forward platform translation (backward falling motion); handhold at 33% shoulder-width from mid-line Additional trials included to increase unpredictability: ‘‘other fwd-translation trials’’ = forward platform translation (backward falling motion); handhold at one of the three other handhold positions (see Fig. 1 inset) ‘‘bwd-translation trials’’ = backward platform translation (forward falling motion); handhold at any of the four handhold positions ‘‘catch trials’’ = no platform motion; handhold at any of the four handhold positions Trials included in the delay-time analyses are enclosed by the thick black border; whereas trials included in the secondary-task analyses are enclosed by the thick gray border

334

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

handhold (sampling-rate = 200 Hz), and confirmed using a hand-velocity criterion (<5% peak resultant velocity), as determined by the motion-analysis system described below. Movement-time was defined as the difference between contact- and reaction-time. Reaction- and contact-time were defined relative to onset of platform acceleration (>0.1 m/s2), determined by an accelerometer mounted on the motion platform. A three-dimensional motion-analysis system (Vicon-Peak Performance; Englewood, CO) collected kinematic data (sampling rate = 200 Hz; displacement data low-pass filtered at 6 Hz (Gage, Zabjek, Hill, & McIlroy, 2007) using a dual-pass fourth-order Butterworth filter (99% of marker signal power was found to be <6 Hz in unfiltered trials)). Analysis focused on the location and trajectory of the right hand (marker on third-metacarpal knuckle), relative to markers mounted at each end of the handhold. The hand-marker data were used to determine the maximum resultant hand velocity in the transverse plane, time-to-peak-velocity (relative to the movement-onset time (vertical hand velocity >5% of peak)), and time-after-peak-velocity (time from peak velocity to handhold contact). The location of the hand marker relative to the center of the handhold target area, at time of handhold contact, was used to define the reach error in each coordinate direction. Data from the hand marker were also used to describe the trajectory of the reach in the transverse plane. Trajectory data were used to determine the deviation from the ‘direct-path’ to the handhold (using a method adapted from studies of volitional reaching (Khan et al., 2006; Khan, Lawrence, Franks, & Elliott, 2003)), as well as the maximum lateral direct-path deviation and maximum hand elevation (see insets in Fig. 5). The direct-path was defined as the straight-line path connecting the hand position at movement onset to the center of the target region of the handhold. Orthogonal deviation from this direct-path was calculated at increments of 5% of the direct-path distance. Repeated-measures analysis of variance (ANOVA) and post-hoc Tukey multiple comparisons were performed to test the hypotheses. The primary dependent variables were: 1) reach timing (reactiononset, movement-time and contact-time); 2) reach velocity (peak velocity, time-to-peak-velocity and time-after-peak-velocity); 3) reach accuracy and variability in reach accuracy; and 4) grasp formation (frequency of full-grasp and frequency of hand-handhold collision errors). Other variables analyzed included: 1) orthogonal deviation from a direct-path trajectory in the transverse plane; 2) maximum vertical elevation and lateral deviation of the hand during the trajectory; and 3) cognitive-task error rate. The reach accuracy and variability variables were analyzed separately in each coordinate

Fig. 3. Effect of recall-delay (A) and cognitive-task (B) on reach-to-grasp timing (EMG-onset, movement and handhold-contact times). Note the slower timing during no-delay and cognitive-task trials. ⁄indicates a significant difference due to recall-delay in A or cognitive-task in B (a = .05); whiskers indicate standard deviations.

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

335

direction, and were expressed as a proportion of subject height prior to analysis (to control for variation related to differences in body size). The criterion level of significance for all analyses was .05. For each dependent variable, the entire set of observations (from individual trials) was rank-transformed (smallest observation assigned rank 1, second smallest assigned rank 2, etc; average ranks assigned in case of ties) prior to analysis, to avoid errors arising from violations of the assumptions (i.e., normality of residuals and homogeneity of variance) underlying the ANOVA (Conover & Iman, 1981). This rank-transform approach is equivalent to performing a non-parametric test, and allows for nonparametric analysis of experimental designs (e.g., repeated-measures) that cannot be analyzed using standard non-parametric tests (e.g., Kruskal-Wallis test) (Akritas & Arnold, 1994; Conover & Iman, 1981; Thompson, 1991b). Analyses focussed on effects of: 1) recall delay time (DT = 0 s, 2 s, 5 s or 10 s); and 2) secondary cognitive task (no-task, spatial-task or non-spatial-task). The delay-time analyses were restricted to the no-secondary-task trials, since these were the only trials where all four delay times were tested. Similarly, the secondary-task analyses were restricted to the trials with DT = 5 s or DT = 10 s, since these were the only trials where all three secondary tasks were performed. Task  Delay-Time interaction was included in the statistical model for the secondary-task analyses; however, no significant interactions were found. (It has been reported that analysis of rank-transforms may not provide a robust test for interactions (Thompson, 1991a); however, analysis of the untransformed data also showed no significant interactions). For the frequency variables, the proportion of trials in which the event was observed was calculated within each subject, for each of the experimental conditions, and the ANOVA was performed on the rank-transformed proportions. To analyze accuracy variability, the standard deviation of the grasp error was determined within each subject, for each experimental condition, and the ANOVAs were performed on the rank-transformed standard deviations. To analyze deterioration in cognitive-task performance during the balance-perturbation trials, the percentage of trials with incorrect responses was determined within each experimental condition, for each subject, and the change relative to the subject’s baseline (single-task) incorrect-response rate was determined; the ANOVA was performed on the rank-transformed changes in error rates. For all of the other variables, rank-transformed data from individual trials were used. Each subject performed 24 ‘‘focus’’ trials; therefore, there were a total of 240 trials available for analysis.

Fig. 4. Effect of recall-delay (A) and cognitive-task (B) on hand velocity (peak resultant velocity in the transverse plane, time-topeak velocity and time-after-peak-velocity). Note the slightly longer time-after-peak-velocity during spatial-task trials. ⁄ Indicates a significant difference due to cognitive-task in B (a = .05); whiskers indicate standard deviations.

336

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

3. Results All subjects were able to execute reach-to-grasp reactions that were adequate to restore postural equilibrium despite having to ‘‘remember’’ the handhold location for up to 10 s, even when performing a concurrent spatial or non-spatial cognitive task. They never fell into the safety harness (maximum harness loading was below 5% of body-weight in all trials) or stepped to recover their balance (i.e., by kicking the foam barriers out of the way), and were always able to make contact with the handhold (using the right hand, as instructed).

3.1. Effect of delay time (in trials with no secondary cognitive task) 3.1.1. Timing and speed of motion There was no evidence that recall delay time (DT) led to slower responses. In fact, the opposite was the case. Although recall delay had no significant effect on movement-time (p = .17), both reactionand contact-time were slowest in the no delay trials (mean reaction-time of 135 ms vs. 119– 126 ms for DT = 2–10 s; contact-time of 444 ms vs. 418–434 ms; Fs (3, 27) > 5.56, ps < .0042; Fig. 3A). There was no effect due to recall delay on peak hand velocity, time-to-peak-velocity, or time-after-peak-velocity (ps > .35; Fig. 4A).

Fig. 5. Effect of recall delay and cognitive task on mean hand trajectory (A and B, respectively) and on maximum hand displacement (C and D, respectively). The trajectories and maximum lateral hand displacement are defined relative to the ‘‘direct path’’ connecting the starting hand position and the center of the handhold in the transverse plane (see inset schematic drawings). Note the increased tendency to transport in a more lateral path with greater maximum vertical elevation during longer recall delay trials, as well as the lack of cognitive task effect on hand trajectory. ⁄Indicates a significant difference due to recall delay in C (a = .05); whiskers indicate standard deviations. (Note: the data were expressed as a proportion of body height in the analyses to reduce variation related to differences in body size; however, the data are shown here in mm to make it easier to interpret the magnitude of the differences in hand displacement.)

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

337

Fig. 6. Effect of recall delay (A) and cognitive task (B) on reach-to-grasp accuracy (systematic error) for each coordinate axis. Note the increased lateral error at the longest recall delay time (10 s) and a near significant trend for increased lateral error when performing either cognitive task. ⁄ Indicates a significant difference due to recall delay in A (a = .05); whiskers indicate standard deviations. (See Fig. 5 caption for note re: presentation of data in mm.)

3.1.2. Reach trajectory, accuracy and variability Analysis of the hand trajectory revealed no significant effects of recall delay on deviation from a direct path (ps > .072; Fig. 5A); however, there was an effect on maximum lateral deviation and maximum vertical elevation (Fs (3, 27) > 5.20, ps < .0058; Fig. 5C). Mean maximum vertical elevation increased from 107–118 mm for DT = 0–2 s to 134–138 mm for DT = 5–10 s, whereas mean maximum lateral deviation increased from 16–20 mm for DT = 0–2 s to 28–37 mm for DT = 5–10 s. Medio-lateral landing accuracy was also affected by recall delay, with largest lateral errors occurring for DT = 10 s (10.6 mm, vs. –6.5 to 2.7 mm for DT = 0–5 s; F(3, 27) = 4.18, p = .015 (Fig. 6A). However, delay time did not affect mean accuracy in the antero-posterior or vertical directions (p’s > .51; Fig. 6A), nor did it have any effect on accuracy variability, in any direction (ps > .33; Fig. 7A). 3.1.3. Prehension Recall delay had no significant effect on the frequency of hand-handhold collision (3–7% of trials; p = .82) or the frequency of full grasp (83% of trials for DT = 0 s, 70% for DT-2–5 s, and 60% for DT = 10 s; p = .12). 3.2. Effect of secondary task (in trials with delay-time of 5 s or 10 s) 3.2.1. Cognitive task performance As a result of our efforts to control the level of task difficulty, there was relatively little variation between subjects in baseline (single-task) cognitive-task performance (Table 1). Relative to baseline performance, there was little consistent evidence of change in the dual-task (perturbation) trials. For both spatial- and non-spatial-tasks, the rate of incorrect response during dual-tasking increased by 15% or less (with respect to single-task trials performed at baseline) in nine of ten subjects, and the mean error rate only increased by 3–4% during dual tasking (p = .78 for main effect due to task). 3.2.2. Timing and speed of motion Cognitive task had a significant main effect on reaction-, movement- and contact-time, Fs (2, 18) > 5.30, ps < .016 (Fig. 3B). In each case, the fastest times were observed during the no-task trials

338

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

Fig. 7. Effect of recall delay (A) and cognitive task (B) on reach-to-grasp accuracy (variable error) for each coordinate axis. Bars indicate means; whiskers indicate standard deviations. (See Fig. 5 caption for note re: presentation of data in mm.)

(reaction-time of 123 ms vs. 135–139 ms for the two dual tasks; movement-time of 305 ms vs. 317– 326 ms; contact-time of 428 ms vs. 452–465 ms). Peak hand velocity and time-to-peak-velocity were not affected by task condition (ps > .33; Fig. 4B). There was, however, a tendency for time-after-peakvelocity to be slightly prolonged during cognitive-task trials (72 ms for spatial-task, 64 ms for nonspatial-task, 57 ms for no-task; F(2, 18) = 4.97, p = .019; Fig. 4B). 3.2.3. Reach trajectory, accuracy and variability Analysis of the reach trajectory revealed no significant effect of cognitive task on direct-path deviation, maximum lateral hand deviation, or maximum vertical hand elevation (ps > .18; Fig. 5B and D). Analysis of systematic (mean) reach accuracy showed no significant cognitive-task effect in the antero-posterior and vertical directions (ps > .15; Fig. 6B); however, there was some evidence of a possible trend for greater lateral error when performing either cognitive task (12.9–19.5 mm, vs. 2.0 mm in notask trials; F(2, 18) = 3.00, p = .075; Fig. 6B). Similarly, variability of hand landing position was not affected by cognitive task in the antero-posterior or vertical directions (ps > .11), but there was a possible trend toward increased medio-lateral variability when performing either cognitive task (22.6–24.5 mm, vs. 16.5 mm for no-task; F(2, 18) = 2.91, p = .081); see Fig. 7B. 3.2.4. Prehension Cognitive task had no significant effect on the frequency of hand-handhold collisions (2–5% of all trials; p = .47) or the frequency of full grasp (63–65% of no-task and spatial-task trials, 52% of non-spatial-task trials; p = .12). 4. Discussion Subjects were generally well able to achieve a functionally-adequate grasp and prevent themselves from falling in the absence of online visual feedback, despite the challenges imposed by: (1) substantially prolonging the length of time that visuospatial information (VSI) about the handhold location had to be retained in working memory prior to perturbation onset, and (2) performing a concurrent spatial-memory task that was expected to interfere with memory-retention.

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

339

Nonetheless, it is clear that some decay in accuracy occurred, based on the findings that lateral deviation in hand trajectory and lateral error in hand-landing position were significantly larger during the 10 s-delay trials. Importantly, however, there was no evidence that the reach-to-grasp responses were slowed as a consequence of this memory decay. The findings thus support our expectation that the CNS would act to preserve rapid speed of response to prevent falling, and would not trade off speed in order to increase accuracy as long as a functional grip could be achieved. While the present findings are consistent with the reductions in endpoint accuracy that have been demonstrated in previous studies of effects of recall-delay on volitional targeted arm movements (Hu et al., 1999; Lemay & Proteau, 2002; Westwood et al., 2003), the absence of any slowing of the perturbation-evoked responses due to recall delay is a clear departure from the volitional-movement literature. Presumably, the slowing that has been observed during volitional arm movements (Bradshaw & Watt, 2002; Hesse & Franz, 2009; Hu et al., 1999; Lemay & Proteau, 2001, 2002; Westwood, Heath, & Roy, 2001; Westwood et al., 2003) serves to increase the time available to formulate and execute online trajectory corrections (Khan et al., 2006) based on proprioceptive feedback (Sarlegna et al., 2003, 2004). If so, then the absence of any slowing due to recall delay in the present study may have reduced the capacity for such trajectory corrections and thereby exacerbated the reduction in end-point accuracy arising from memory decay. Interestingly, the delayed-recall perturbation responses not only failed to show any slowing compared with no-delay trials, responses at delays of 2 s, 5 s and/or 10 s were actually faster than in the no-delay trials in terms of reaction- and contact-time. Possibly, subjects were unprepared to respond so shortly after visual-occlusion onset in the no-delay trials (Gottsdanker, 1979). Conversely, the longer recall delay times may have reduced temporal uncertainty of perturbation onset, leading to faster reaction time in those trials (Polzella, Ramsey, & Bower, 1989). Alternatively, the faster reactions in recall-delay trials might be due to increased arousal (Carpenter, Frank, Adkin, Paton, & Allum, 2004). One might well expect the more prolonged delay times to lead to a heightened build-up of arousal (as participants ‘‘wait’’ longer for the perturbation to occur). Further work is needed to examine whether arousal levels were actually influenced by recall-delay time. Previous studies of volitional reach-to-grasp movements have also shown that recall-delay leads to adaptations such as increase in the hand aperture (Berthier, Clifton, Gullapalli, McCall, & Robin, 1996; Hesse & Franz, 2009; Jakobson & Goodale, 1991; Schettino, Adamovich, & Poizner, 2003). Although we did not measure hand aperture in the present study, we did see possible evidence of another adaptation. Specifically, we found a tendency for greater elevation of the hand (above the handhold) prior to handhold contact, in trials with increasing recall-delay time (see Fig. 5C). This may reflect a strategic response to avoid possible collision with the handhold, which could otherwise occur due to the decay in the egocentric visuospatial-memory of the precise handhold location. Our finding that lateral landing error was increased by recall-delay could possibly reflect the manner by which the CNS maintains a representation of the environment. In the no-delay trials, the CNS has access to the high resolution visual representations that are briefly held in iconic memory (Adam, Paas, Ekering, & van Loon, 1995; Binsted, Rolheiser, & Chua, 2006; Elliott & Madalena, 1987; Lemay & Proteau, 2002); hence, end-point control is likely to be most accurate. However, in recall-delay trials, with the recall time exceeding the temporal limits of iconic memory, the CNS would presumably have to rely on the more long-lasting, but lower resolution, representations that occur in the ventral-stream areas (Goodale & Milner, 1992; Milner & Goodale, 2008). The observed lateral undershoot error (and associated lateral shift in hand trajectory; Fig. 5A, C) is also consistent with the ‘‘contraction of working space’’ that appears to occur when there is recall-delay (Chieffi, Allport, & Woodin, 1999; McIntyre, Stratta, & Lacquaniti, 1998). Contrary to our hypotheses, concurrent performance of spatial and non-spatial cognitive tasks had similar effects on the perturbation-evoked reach-to-grasp reactions. Both tasks caused a similar delay in reaction-, movement-, and contact-time, and there was also some evidence that both systematic and variable end-point error in the medio-lateral direction were larger when either cognitive task was performed. Previous dual-task studies of balance-recovery reactions have tended to focus on lower-limb reactions (see reviews by (Maki & McIlroy, 2007; Woollacott & Shumway-Cook, 2002)). We are aware of only two studies that have examined dual-task effects on perturbation-evoked

340

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

reach-to-grasp. The present finding that reaction-onset time was delayed when performing a concurrent cognitive task is consistent with one of these studies, which showed that performing a mentalarithmetic task led to a 40 ms delay in deltoid latency (Quant, McIlroy, Verrier, & Maki, 2000); however, the other study found that performing a visual vigilance task had no effect on reach-to-grasp reaction-onset time (King, McKay, Cheng, & Maki, 2010). Although the cognitive-task instructions were delivered only during the recall-delay interval in the present study, it is possible that our subjects continued to attend to each cognitive task for at least some interval of time subsequent to perturbation onset, and that the primary effect of both tasks was due to competing demands for attentional resources. These resources would be required to retrieve handhold location from visuospatial-memory (Oberauer, 2002), and to plan and execute the stabilizing limb movement (Maki & McIlroy, 2007). We expected that the primary effect of the visuospatial-memory task would be to interfere with accurate retention of the handhold location in spatial working memory, due to a capacity-limited store (Fougnie & Marois, 2009). The failure to see any dual-task interference effects (on reach accuracy or cognitive-task performance) would therefore suggest that the spatial-working-memory capacity of our healthy young-adult cohort was not exceeded by the demands of the current protocol. Despite our efforts to use recall-delay and concurrent cognitive tasks to challenge the ability to use stored VSI to guide reach-to-grasp reactions, the healthy young-adult subjects in the present study were always able to achieve a functionally-adequate grasp and prevent themselves from falling. However, it should be noted that the perturbations used in this study were relatively small in magnitude, and it may well be that the presently observed errors and delays in contacting the rail may have a greater impact when responding to larger perturbations. For example, the need to respond as rapidly as possible, in order to counter a greater degree of instability, could potentially limit the time available for the CNS to process the proprioceptive feedback needed to correct the arm trajectory after initiation. In addition, more challenging handhold locations, greater variation in these locations, and variation in handhold shape, size and orientation could require more accurate encoding and storage of target position. Another limitation was that unlike many real-life loss-of-balance situations, the perturbations in this study, though unpredictable in timing and direction, were fully expected by the subjects. It is possible that the effects observed in the present study may have greater functional significance when responding to balance perturbations in the complex environments and unpredictable situations of daily life. 5. Conclusion The results of this study indicate that healthy young adults were able to execute functionally-adequate reach-to-grasp balance-recovery reactions in the absence of online visual feedback, even when they were required to store the requisite visuospatial target information in memory for a prolonged interval prior to perturbation onset and to perform a challenging concurrent spatial-memory task during this ‘‘recall-delay’’ interval. Consistent with studies of volitional reach-to-grasp, recall-delay did lead to some reduction in endpoint accuracy; however, unlike those studies, the present results showed no evidence that recall delay led to slowing of the arm movement, although slowing did occur as a consequence of performing a concurrent cognitive task. Both spatial and non-spatial cognitive tasks had similar effects, which suggests that these effects were related to generic attentional demands rather than competition for specific cognitive resources related to spatial working memory. Further work is needed to establish whether the observed memory-decay and/or dual-task interference effects are exacerbated when responding to larger perturbations with more challenging handhold locations, and to determine the effects of age-related impairments in visuospatial-memory and attentional capacity. Funding This work was supported by the Canadian Institutes of Health Research (MOP-13355); the Ontario Neurotrauma Foundation (2009-PREV-INT-785); and scholarships from the Ontario Graduate Scholarship Program.

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

341

Acknowledgments Assistance from Stephanie Middleton, Simon Jones, Sandra McKay and Carmen Ho is gratefully acknowledged.

References Abrams, R. A. (1992). Coordination of eye and hand for aimed limb movements. In L. Proteau & D. Elliot (Eds.), Vision and motor control (pp. 129–152). Amsterdam: Elsevier. Adam, J. J., Paas, F. G., Ekering, J., & van Loon, E. M. (1995). Spatial localization: Tests of a two-process model. Experimental Brain Research, 102, 531–539. Akritas, M. G., & Arnold, S. F. (1994). Fully nonparametric hypotheses for factorial designs I: Multivariate repeated measures designs. Journal of the American Statistical Association, 89, 336–343. Bateni, H., Zecevic, A., McIlroy, W. E., & Maki, B. E. (2004). Resolving conflicts in task demands during balance recovery: Does holding an object inhibit compensatory grasping? Experimental Brain Research, 157, 49–58. Berthier, N. E., Clifton, R. K., Gullapalli, V., McCall, D. D., & Robin, D. J. (1996). Visual information and object size in the control of reaching. Journal of Motor Behavior, 28, 187–197. Binsted, G., Rolheiser, T. M., & Chua, R. (2006). Decay in visuomotor representations during manual aiming. Journal of Motor Behavior, 38, 82–87. Bradshaw, M. F., & Watt, S. J. (2002). A dissociation of perception and action in normal human observers: The effect of temporal delay. Neuropsychologia, 40, 1766–1778. Brooks, L. R. (1967). The suppression of visualization by reading. Quarterly Journal of Experimental Psychology, 19, 289–299. Carnahan, H., & Marteniuk, R. G. (1991). The temporal organization of hand, eye, and head movements during reaching and pointing. Journal of Motor Behavior, 23, 109–119. Carpenter, M. G., Frank, J. S., Adkin, A. L., Paton, A., & Allum, J. H. (2004). Influence of postural anxiety on postural reactions to multi-directional surface rotations. Journal of Neurophysiology, 92, 3255–3265. Cheng, K. C., McKay, S. M., King, E. C., & Maki, B. E. (2012a). Does aging impair the capacity to use stored visuospatial information or online visual control to guide reach-to-grasp reactions evoked by unpredictable balance perturbation? Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 67, 1238–1245. Cheng, K. C., McKay, S. M., King, E. C., & Maki, B. E. (2012b). Reaching to recover balance in unpredictable circumstances: Is online visual control of the reach-to-grasp reaction necessary or sufficient? Experimental Brain Research, 218, 589–599. Cheng, K. C., McKay, S. M., King, E. C., Tung, J. Y., Lee, T. A., Scovil, C. Y., et al (2009). The moveable handhold: A new paradigm to study visual contributions to the control of balance-recovery reactions. Gait & Posture, 29, 339–342. Chieffi, S., Allport, D. A., & Woodin, M. (1999). Hand-centred coding of target location in visuo-spatial working memory. Neuropsychologia, 37, 495–502. Conover, W. J., & Iman, R. L. (1981). Rank transformations as a bridge between parametric and nonparametric statistics. The American Statistician, 35, 124–133. Elliott, D., & Madalena, J. (1987). The influence of premovement visual information on manual aiming. Quarterly Journal of Experimental Psychology A, Human Experimental Psychology, 39, 541–559. Fougnie, D., & Marois, R. (2006). Distinct capacity limits for attention and working memory: Evidence from attentive tracking and visual working memory paradigms. Psychological Science, 17, 526–534. Fougnie, D., & Marois, R. (2009). Dual-task interference in visual working memory: A limitation in storage capacity but not in encoding or retrieval. Attention, Perception, and Psychophysics, 71, 1831–1841. Gage, W. H., Zabjek, K. F., Hill, S. W., & McIlroy, W. E. (2007). Parallels in control of voluntary and perturbation-evoked reach-tograsp movements: EMG and kinematics. Experimental Brain Research, 181, 627–637. Ghafouri, M., McIlroy, W. E., & Maki, B. E. (2004). Initiation of rapid reach-and-grasp balance reactions: Is a pre-formed visuospatial map used in controlling the initial arm trajectory? Experimental Brain Research, 155, 532–536. Goodale, M. A., & Milner, A. D. (1992). Separate visual pathways for perception and action. Trends in Neurosciences, 15, 20–25. Gottsdanker, R. (1979). A psychological refractory period or an unprepared period? Journal of Experimental Psychology: Human Perception and Performance, 5, 208–215. Hale, S., Myerson, J., Rhee, S. H., Weiss, C. S., & Abrams, R. A. (1996). Selective interference with the maintenance of location information in working memory. Neuropsychology, 10, 228–240. Haycoe, M., & Ballard, D. (2005). Eye movements in natural behavior. Trends in Cognitive Sciences, 9, 188–194. Hesse, C., & Franz, V. H. (2009). Memory mechanisms in grasping. Neuropsychologia, 47, 1532–1545. Hesse, C., & Franz, V. H. (2010). Grasping remembered objects: Exponential decay of the visual memory. Vision Research, 50, 2642–2650. Hu, Y., Eagleson, R., & Goodale, M. A. (1999). The effects of delay on the kinematics of grasping. Experimental Brain Research, 126, 109–116. Jakobson, L. S., & Goodale, M. A. (1991). Factors affecting higher-order movement planning: A kinematic analysis of human prehension. Experimental Brain Research, 86, 199–208. Jeannerod, M. (1988). The neural and behavioural organization of goal-directed movements. Oxford: Clarendon. Khan, M. A., Franks, I. M., Elliott, D., Lawrence, G. P., Chua, R., Bernier, P. M., et al (2006). Inferring online and offline processing of visual feedback in target-directed movements from kinematic data. Neuroscience and Biobehavioral Reviews, 30, 1106–1121. Khan, M. A., Lawrence, G. P., Franks, I. M., & Elliott, D. (2003). The utilization of visual feedback in the control of movement direction: Evidence from a video aiming task. Motor Control, 7, 290–303. King, E. C., Lee, T. A., McKay, S. M., Scovil, C. Y., Peters, A. L., Pratt, J., et al (2011). Does the ‘‘eyes lead the hand’’ principle apply to reach-to-grasp movements evoked by unexpected balance perturbations? Human Movement Science, 30, 368–383.

342

K.C. Cheng et al. / Human Movement Science 32 (2013) 328–342

King, E. C., McKay, S. M., Cheng, K. C., & Maki, B. E. (2010). The use of peripheral vision to guide perturbation-evoked reach-tograsp balance-recovery reactions. Experimental Brain Research, 207, 105–118. King, E. C., McKay, S. M., Lee, T. A., Scovil, C. Y., Peters, A. L., & Maki, B. E. (2009). Gaze behavior of older adults in responding to unexpected loss of balance while walking in an unfamiliar environment: A pilot study. Journal of Optometry, 2, 119–126. Land, M. F. (2006). Eye movements and the control of actions in everyday life. Progress in Retinal and Eye Research, 25, 296–324. Lemay, M., & Proteau, L. (2001). A distance effect in a manual aiming task to remembered targets: A test of three hypotheses. Experimental Brain Research, 140, 357–368. Lemay, M., & Proteau, L. (2002). Effects of target presentation time, recall delay, and aging on the accuracy of manual pointing to remembered targets. Journal of Motor Behavior, 34, 11–23. Logie, R. H., Gilhooly, K. J., & Wynn, V. (1994). Counting on working memory in arithmetic problem solving. Memory and Cognition, 22, 395–410. Maki, B. E., & McIlroy, W. E. (1997). The role of limb movements in maintaining upright stance. The ‘‘change-in-support’’ strategy. Physical Therapy, 77, 488–507. Maki, B. E., & McIlroy, W. E. (2005). Change-in-support balance reactions in older persons: An emerging research area of clinical importance. Neurologic Clinics, 23, 751–783. Maki, B. E., & McIlroy, W. E. (2007). Cognitive demands and cortical control of human balance-recovery reactions. Journal of Neural Transmission, 114, 1279–1296. Maki, B. E., McIlroy, W. E., & Perry, S. D. (1996). Influence of lateral destabilization on compensatory stepping responses. Journal of Biomechanics, 29, 343–353. Maki, B. E., Perry, S. D., & McIlroy, W. E. (1998). Efficacy of handrails in preventing stairway falls: A new experimental approach. Safety Science, 28, 189–206. Marigold, D. S., & Misiaszek, J. E. (2009). Whole-body responses: Neural control and implications for rehabilitation and fall prevention. The Neuroscientist, 15, 36–46. McAfoose, J., & Baune, B. T. (2009). Exploring visual-spatial working memory: A critical review of concepts and models. Neuropsychology Review, 19, 130–142. McIlroy, W. E., & Maki, B. E. (1993). Changes in early ‘automatic’ postural responses associated with the prior-planning and execution of a compensatory step. Brain Research, 631, 203–211. McIlroy, W. E., & Maki, B. E. (1995). Early activation of arm muscles follows external perturbations of upright stance. Neuroscience Letters, 184, 177–180. McIntyre, J., Stratta, F., & Lacquaniti, F. (1998). Short-term memory for reaching to visual targets: Psychophysical evidence for body-centered reference frames. Journal of Neuroscience, 18, 8423–8435. Milner, A. D., & Goodale, M. A. (2008). Two visual systems re-viewed. Neuropsychologia, 46, 774–785. Oberauer, K. (2002). Access to information in working memory: Exploring the focus of attention. Journal of Experimental Psychology: Learning, Memory, and Cognition, 28, 411–442. Polzella, D. J., Ramsey, E. G., & Bower, S. M. (1989). The effects of brief variable foreperiods on simple reaction time. Bulletin of the Psychonomic Society, 27, 467–469. Prablanc, C., Echallier, J. E., Jeannerod, M., & Komilis, E. (1979). Optimal response of eye and hand motor systems in pointing at a visual target. I. Spatio–temporal characteristics of eye and hand movements and their relationships when varying the amount of visual information. Biological Cybernetics, 35, 113–124. Quant, S., McIlroy, W. E., Verrier, M. C., & Maki, B. E. (2000). Effect of attention on compensatory upper limb balance reactions. Physiotherapy Canada, 52(Suppl.), 98. Sarlegna, F. R., Blouin, J., Bresciani, J. P., Bourdin, C., Vercher, J. L., & Gauthier, G. M. (2003). Target and hand position information in the online control of goal-directed arm movements. Experimental Brain Research, 151, 524–535. Sarlegna, F. R., Blouin, J., Vercher, J. L., Bresciani, J. P., Bourdin, C., & Gauthier, G. M. (2004). Online control of the direction of rapid reaching movements. Experimental Brain Research, 157, 468–471. Schettino, L. F., Adamovich, S. V., & Poizner, H. (2003). Effects of object shape and visual feedback on hand configuration during grasping. Experimental Brain Research, 151, 158–166. Scovil, C. Y., Zettel, J. L., & Maki, B. E. (2008). Stepping to recover balance in complex environments: Is online visual control of the foot motion necessary or sufficient? Neuroscience Letters, 445, 108–112. Thompson, G. L. (1991a). A note on the rank transform for interactions. Biometrika, 78, 697–701. Thompson, G. L. (1991b). A unified approach to rank tests for multivariate and repeated measures design. Journal of the American Statistical Association, 86, 410–419. Westwood, D. A., Heath, M., & Roy, E. A. (2001). The accuracy of reaching movements in brief-delay conditions. Canadian Journal of Experimental Psychology, 55, 304–310. Westwood, D. A., Heath, M., & Roy, E. A. (2003). No evidence for accurate visuomotor memory: Systematic and variable error in memory-guided reaching. Journal of Motor Behavior, 35, 127–133. Woollacott, M., & Shumway-Cook, A. (2002). Attention and the control of posture and gait: A review of an emerging area of research. Gait and Posture, 16, 1–14.