Comparing three portable,tablet-based visuomotor tasks to laboratory versions: An assessment of test validity

Abstract

The assessment of visuomotor function can provide important information about neurological status. Many tasks exist for testing visuomotor function in the laboratory, but the availability of portable, easy-to-use versions that allow reliable, accurate, and precise measurement of movement timing and accuracy has been limited. We developed a tablet application that uses three laboratory visuomotor tests: the double-step task, interception task, and stop-signal task. We asked the participants to perform both the lab and tablet versions of each task and compared their response patterns across equipment types to assess the validity of the tablet versions. On the double-step task, the participants adjusted to the displaced target adequately in both the lab and tablet versions. On the interception task, the participants intercepted nonaccelerating targets and performed worse on accelerating targets in both versions of the task. On the stop-signal task, the participants successfully inhibited their reaching movements on short stop-signal delays (50–150 ms) more frequently than on long stop-signal delays (200 ms) in both versions of the task. Our findings suggest that the tablet version of each task assesses visuomotor processing in the same way as their respective laboratory version, thus providing the research community with a new tool to assess visuomotor function.

Keywords

Assessment tablet motor skills vision

Introduction

Concussions are the most common form of brain injuries, with prevalence reports of approximately 110 people per 100,000 in Canada.^1–3 Several methods are typically used to assess concussion including self-report symptom checklists and neuropsychological testing. Checklists offer a quick and relatively comprehensive survey of unresolved symptoms but are susceptible to underreporting of symptoms.^4,5 Neuropsychological assessment offers a more objective and in-depth evaluation of memory, attention, and processing speed^6,7 and may be repeated to determine when patients have returned to normal levels of function.^5,7,8

One neurological process that can be affected by concussion is visuomotor control, usually assessed by analyzing the speed, accuracy, and adaptability of visually guided upper-limb movements. Heitger et al.⁹ asked concussed patients and age-matched controls to undergo a battery of neuropsychological and visuomotor tests. The neuropsychological battery focused on assessment of attention, memory, processing speed, and executive performance, and the visuomotor tests focused on oculomotor assessment (reflexive saccades, anti-saccades, memory-guided saccades, and smooth-pursuit) and assessment of upper-limb movement speed and stability during target-tracking tasks. In the oculomotor domain, concussed participants showed higher saccadic reaction times (RT), decreased saccadic accuracy, and more directional errors compared to healthy controls. Upper-limb assessment revealed that concussed participants had lower peak speeds than healthy controls, which resulted in significantly longer lag times on tracking tasks. By contrast, neuropsychological test performance did not differ between the two groups, suggesting that visuomotor assessment may be more sensitive to concussion than traditional cognition-focused neuropsychological assessment.

Heitger et al.¹⁰ extended this finding by using both neuropsychological and visuomotor assessments to examine recovery from concussion one week, three months, six months, and one year post-injury. They found that neuropsychological assessment detected differences between concussed patients and sex- and age-matched controls reliably at one week post-injury but did not detect differences between groups at later time points (with the exception of performance on a test of verbal memory). By contrast, visuomotor assessment revealed differences between groups on tests of both oculomotor and upper-limb performance at the early time points and up to six months post-injury. Together, the Heitger et al.^9,10 studies demonstrate that concussion negatively influences the visuomotor system and that deficits in oculomotor and visuomotor domains take longer to resolve than other neuropsychological deficits. Indeed, visuomotor assessments may be more sensitive to concussions than other components of neuropsychological testing and may be used to track recovery over the weeks following an injury.

Given the importance of visuomotor assessment when determining concussion recovery, we recognized the need for valid and easy-to-use portable versions of common visuomotor tasks for assessment purposes. The portable versions of these tasks must provide reliable, accurate, and precise measurement of movement timing and accuracy. Our proposed battery focuses on three ubiquitous visuomotor tasks—the double-step task, the interception task, and the stop-signal task—chosen because they measure visual, cognitive, and motor processes that are crucial to the performance of any activity that requires the coordination of one person’s (player’s) movements with those of other people (both teammates and opponents). Another key feature of the chosen tests is that they have been used to assess visuomotor dysfunction after brain insult. The double-step task^11,12 measures the ability of the visuomotor system to track the location of the hand relative to a visual target and to use that information to adjust ongoing movements to an unexpected change in target location. This online spatial updating depends critically on the posterior parietal cortex (PPC) as patients with lesions to the PPC fail to adapt to unpredictable changes in target position,^13–15 and healthy participants fail to adapt to a perturbation of target location when online single-pulse transcranial magnetic stimulation (TMS) is applied to the PPC.¹⁶

Successful interception of a moving target depends on the participants’ ability to use target motion information to predict the target’s future position so that they can time their movement to hit the target.^17,18 Performance on this task is dependent on brain regions, such as the middle temporal area (MT, also known as V5), that are specifically tuned for detecting target motion, target speed, and direction of motion.^19,20 Patients with lesions that include MT perform poorly when asked to intercept moving targets,^21,22 and when TMS is applied to MT in healthy participants, they are unable to accurately time their interceptive movements.²³

Finally, the stop-signal task measures how well and how quickly participants can interrupt a planned and signaled response.²⁴ Participants are presented with a simple-choice RT task but are also instructed that on some trials (50%), the presentation of a subsequent tone indicates that they need to interrupt and withhold the response. The medial prefrontal cortex and frontal gyrus are two cortical areas involved in the execution of the stop-signal task.^25,26 Patients with lesions in these areas cannot inhibit their responses as well as healthy controls.²⁷

Visuomotor assessment with these three tasks is typically conducted in laboratory facilities using nonportable equipment that is not commonly available in clinical settings. The goal of the current study was to create such a test battery on a tablet and to determine the validity of the tablet version by comparing patterns of test performance on the tablet to that on laboratory versions. A common paradigm for evaluating motor learning—the visuomotor rotation paradigm—was successfully implemented in a tablet application (PoMLab²⁸), providing evidence that our goal is reasonable. If these patterns are similar across equipment type, the validity of the tablet-battery tasks will be supported. A reliable, easy-to-use, portable, and standardized version of common visuomotor tasks may aid in the clinical assessment of visuomotor performance. While our current focus is on developing a tablet-battery for assessing and monitoring recovery, it is possible that these tests could be used at earlier stages of assessment (in the sporting arena) and/or used in an assessment or rehabilitation training capacity by patients who are recovering at home.^28,29 Our tablet tasks were coded in Xcode (https://developer.apple.com/xcode/), and the code for each of the tasks can be downloaded here: https://github.com/lianabro/TabletTasks.

Task 1: The double-step task

This task requires participants to reach for targets that vary in distance (but not direction) from a fixed start location. On the majority of trials, participants simply need to reach for the target quickly and accurately to successfully complete each trial. On a proportion of trials (25% here), the target moves a short time after it is presented to a position that is slightly beyond or slightly short of the original target. The “target jump” is timed to coincide with the onset of the movement to the initial target. Measures of movement accuracy and movement speed capture how well participants are able to adjust their movement to this perturbation. Past research shows that participants typically adjust to the perturbation seamlessly such that there are no consequences for movement accuracy or time.^11,12,30

Methods

Participants

Sixty-seven healthy undergraduate students were recruited from Trent University with 32 participants (19.9 ± 3.19 years, 91% female) completing the laboratory version of the task and 35 participants (25.1 ± 9.8 years, 83% female) completing the tablet version. All participants were compensated monetarily or by receiving bonus grades.

The Dutch Handedness Questionnaire³¹ was used to assess handedness so that the dominant hand was used during the task. Of the 32 participants who completed the laboratory version, 28 (88%) used their right hand during the task, and four (12%) used their left hand. Of the 35 participants who completed the double-step task on the tablet, 33 participants (94%) used their right hand, and two participants (6%) used their left hand.

Part of the intake package asked the participants to report if they had ever experienced an epileptic seizure, a stroke, a head injury, or if they had any visual disorders. Eight participants stated that they had received at least one form of head injury in their life; however, all participants except two had received these head injuries greater than three years previously. None of the participants who had injuries reported lingering complaints, and all had recovered from any associated symptoms, at the very least, several months previously. All participants reported having either normal or corrected-to-normal vision. The Trent University Research Ethics Board approved all procedures, and each participant gave written informed consent before participation.

Apparatus

The participants were tested on one of two versions of the double-step task: the laboratory-based version or the tablet version. For the lab-based version, each participant entered a dimly lit room and was seated in a chair adjusted in height so that they were able to sit comfortably at the testing table. Figure 1(a) shows a schematic side view of the testing table, which consisted of a three-layer visual display system. A projector (Sharp PG-LX2000, Sharp Electronics of Canada Ltd., Mississauga, ON) hung from the ceiling of the testing room above a white cloth screen, a one-way mirror, and a platform that was the working surface of the experiment where participants kept their hands. The cloth screen was 170 cm below the projector and reflected the light that was projected downward, allowing the image of the display to be presented on the one-way mirror (27.5 cm below the cloth screen). This arrangement allowed the participants to see the projected display (60.5 cm × 47.5 cm) of the double-step task on the surface of the one-way mirror but not to have vision of their hands underneath the mirror on the working surface. Attached to the middle of the one-way mirror was a chin rest so that the participants could view the entire surface of the mirror when sitting down. Because the one-way mirror was set above the working surface where the participants kept their hands (27.5 cm), vision of the hands was restricted. In fact, the only visual feedback the participants had while completing this task occurred between trials when a red dot appeared on screen that coincided with where their index finger landed after the reaching movement was completed.

Figure 1.

(a) Apparatus arrangement for the double-step task in the laboratory condition. (b) View of starting location (black dot) and the positions of the five final targets (the three initial target locations are outlined in black). This arrangement was displayed on the surface of the mirror for the laboratory condition. It was also displayed on the screen of the tablet for the tablet condition. Images not to scale.

The working surface that the participants made their reaching movements on was a 183 cm × 107 cm glass plate 85 cm from the floor. A motion tracker (Polhemus LIBERTY 240/8, Polhemus Inc., Colchester, VT) was attached to the index finger of the participants’ dominant hand to track their reaching movements through space over time. The single marker was attached to the index finger of the dominant hand with tape and secured to the rest of the arm to avoid extraneous movement. The motion tracking data were collected at a sampling frequency of 100 Hz via the ProkLiberty toolbox for MATLAB (http://www.prokopenko.org/liberty.html) and saved on a computer for offline analysis.

The participants completed the double-step task tablet version on tablet (iPad 2, Apple Inc., Cupertino, CA) in the same quiet and dimly lit room as the one used for the laboratory version. The screen display on the tablet (19.3 cm × 14.7 cm) was smaller than that of the laboratory version, and the participants could see their hand movements. Further, the double-step task on the tablet utilized a touch screen so a motion tracker was not used with this equipment type. We limited additional laboratory equipment to duplicate how the tablet application would be used portably outside of the laboratory.

Task

The displayed images presented on the laboratory version were the same as those presented on the tablet version. The experiment began with the screen displaying a gray/black random-pixel noise background with a red dot that indicated the starting location for the participants to place their index finger (Figure 1(b)). The trial could not begin unless the participant had placed and held their index finger of their dominant hand on the starting dot. After a pseudo-randomized amount of time from 500 ms to 2 s, the red starting dot disappeared, and a gray target appeared on the screen above where the starting dot was at one of three initial target locations. For the laboratory version of the double-step task, initial target 1 was 23 cm from the starting location, with each of the other two initial locations 3 cm further up the screen. On the tablet version, initial target 1 was 11 cm from the starting location, with each target 2 cm further up the screen. The difference of distances between the two apparatus types occurred due to the difference of display size.

There were five possible final locations of the targets. On jump-absent trials, the final location would be the same as the initial location, as the target would not move. On jump trials, the final location would be a position either one target above or one target below the initial target. The jump trials were randomly interspersed throughout the jump-absent trials, and the participants were not informed that the experimental blocks would contain jump trials. All displays for the laboratory task were created in MATLAB (MathWorks, Natick, MA) using the Psychophysics Toolbox^32,33 version 3.0.8, and the programming for the tablet was done with Xcode (https://developer.apple.com/xcode/).

Design

We used a two-equipment (laboratory version, tablet version) × three-target (final target location) × two-jump-presence (jump, jump-absent) mixed design. Jump presence and final target location were the within-subject independent variables, and equipment type was the between-subject independent variable. The task was presented in five blocks, with the participants initially completing a practice block of 24 jump-absent trials, which was used to help acquaint the participants with the remaining four experimental blocks. After the practice block was concluded, the participants completed the remaining four experimental blocks that consisted of 48 trials each (36 jump-absent trials and 12 trials jump trials). During each jump-absent trial, the participants made 12 reaches to each of the three middle-target locations. On jump trials, the target appeared at one of the three middle-target locations initially (four times to each location) and then jumped to the final location (twice to the target above and twice to the target below) from each initial target location. All jump conditions were interspersed randomly throughout each experimental block.

The jump trials occurred 250 ms after the initial target appeared on screen. This stimulus-onset asynchrony (SOA) was determined from a pilot study in which 20 participants were tested with one of four different SOA times: 175 ms, 200 ms, 225 ms, and 250 ms. We recorded how many jumps each participant perceived in each block to find an ideal SOA where the participants perceived the fewest jumps (as was described in Pélisson et al.¹²). The participants perceived the fewest number of jumps during trials with a SOA of 250 ms, so this SOA was used during experimental testing and data collection on both task versions.

Procedures

The participants entered the testing room and were positioned at the respective apparatus. For the laboratory version, the participants were situated in a chair such that their head fit into a chin and forehead rest, and they were able to make reaching movements to the far end of the projected display on the table with ease. The participants were then instructed verbally how to complete the task, with the same instructions for each equipment type. The participants who completed the lab-based version of the experiment began by completing a calibration to ensure that the motion tracker was measuring the participant’s movements correctly. During the calibration, the participants could view their hand and reaching movements by means of an illuminated lamp beneath the mirror. Once calibration was completed, the practice and experimental blocks commenced, and the lamp remained off for the remainder of the experiment.

After calibration, the participants completed the practice block followed by the experimental blocks. They were instructed to complete reaching movements toward the target as quickly as possible. Upon trial completion, the curser and start location reappeared on screen, and the participants were instructed to return the curser to the starting location and proceed with the remainder of the trials. The short duration between the disappearance of the target and the reappearance of the fingertip cursor may have provided the participants with feedback about how closely they stopped to the target.

The participants who completed the experiment on the tablet were given the same instructions and completed the double-step task with the same experimental process as the participants who completed the task on the laboratory equipment. The only procedural difference was that the participants sat at a wooden table with the tablet placed on top and had continuous visual feedback of their hands at all times. After completion of the experiment, all of the participants were debriefed and given information of the study.

Data analysis

Data processing for the double-step task used both MATLAB software (MathWorks, Natick, MA) and SPSS (Version 20, SPSS Inc.). All trials conducted with the motion tracker in the laboratory group were processed using MATLAB to determine the beginning and end of the pointing movement, chosen automatically based on hand velocity profile. If the automatic process failed to locate the beginning and end of movement, then the start and end were chosen manually.

Movement accuracy was assessed by measuring the distance between the participants’ finger location at movement end and the final target location in depth (depth error). A positive value if their finger landed beyond the target and a negative error if it landed short of the target. Absolute error values were used for the analysis of movement accuracy; it was more important to determine distance from the target rather than the direction of error. Movement accuracy was assessed using the motion tracker for the laboratory version and with the touch screen on the tablet. Movement time was defined as the duration between movement initiation and movement end. On the tablet, movement time was calculated as the duration between the point in time that the participants’ finger left the start location and the time it touched back down. All values of this variable were measured in milliseconds.

Movement accuracy and movement time were the main variables analyzed for the double-step task. The accuracy data were trimmed such that trials with an error greater than four standard deviations above the group means were removed. In total, this resulted in approximately 1% of all trials being omitted from the analysis. These two variables were submitted to a two-equipment (laboratory version, tablet version) by three-target (final target location) by two-jump-presence (jump, jump-absent) mixed analyses of variance (ANOVAs). The source of main effects was determined by conducting pairwise (least significant difference) comparisons of the group means. Interactions were decomposed by conducting simple main effects analyses and pairwise comparisons.

Results

The analysis focused on the participants’ ability to adjust to the presence of the jump and on jump direction. The laboratory version was presented on a larger surface than the tablet version, and we compensated for this difference by scaling response measurements on jump trials to the responses made on jump-absent trials. Consequently, measurements of error and movement time on jump trials were calculated as a percentage of mean absolute error and mean movement time, respectively, incurred on jump-absent trials within each participant for each equipment type. First, we determined if measures on jump trials were greater (greater error, increased movement time) from those on jump-absent trials using one-sample, one-tailed t-tests against a perfect 100%. Then, both measures were submitted to an equipment type (lab, tablet) by jump direction (forward, backward) mixed ANOVA.

Movement accuracy: Absolute error in depth

T-tests revealed that the requirement to adjust to the jump perturbation resulted in greater error on jump-present than jump-absent trials for both the lab version (107.5 ± 4.0%), t(30) = 2.11, p = .022, and the tablet version (117.4 ± 3.7%), t(34) = 4.27, p < .001. The ANOVA, however, revealed no significant difference between equipment types, F(1, 64) = 3.26, p = .076 (Figure 2(a)). The proportion of error incurred on the tablet version was not significantly greater than that measured on the lab version. The ANOVA also revealed that the effect of jump direction was not statistically significant, F(1, 64) =2.54, p = .116. The proportion of error incurred on jump trials that required an adjustment backward from the original target (115.9 ± 3.9%) was not significantly greater than that on jumps that required an adjustment forward (109.0 ± 2.9%). Finally, there was no equipment by jump direction interaction, F(1, 64) = .46, p > .502, suggesting that the adjustments made in response to target jump perturbations did not depend on equipment type.

Figure 2.

(a) Absolute error in depth and (b) movement time on jump-present trials shown as a percentage of performance on jump-absent trials for both the tablet and lab versions of the task. Shown as a function of target distance and jump presence. All comparisons between equipment and jump presence are not significant, except where indicated. *p < .05.

Movement time

T-tests revealed that while there was no significant movement time adjustment made in response to target jump perturbation for the lab version (100.1 ± .8%), t(30) = 0.19, p = .604, movement times were significantly lengthened on jump trials on the tablet version (105.0 ± .8%), t(34) = 5.61, p < .001. The ANOVA also revealed a significant difference between equipment types, F(1, 64) = 19.51, p < .001 (Figure 2(b)). The adjustment to movement time experienced on the tablet version was significantly greater than on the lab version. The ANOVA also revealed that the effect of jump direction was statistically significant, F(1, 64) = 578.89, p < .001, such that jump trials that required an adjustment forward lengthened movement time (111.6 ± .5%), whereas backward adjustments shortened movement time (93.5 ± .8%). Despite these differences, there was no equipment by jump direction interaction, F(1, 64) = 1.48, p > .228, suggesting that the adjustments made in response to target jump perturbations did not depend on equipment type.

Task 2: The interception task

The interception task measures how well people can perceive target visual acceleration and velocity, and how well people can use that information to (1) predict the future motion of the target as it approaches a target zone and (2) plan and execute a movement that will terminate at the time and location of the target zone.^17,18,34 This task has been shown to be sensitive to disturbances in brain function.^23,35 We developed a tablet application of the interception task for use in our battery of visuomotor tasks.

The participants’ percentage of correctly intercepted targets and temporal error (whether the person’s hand arrived in the interception zone earlier or later than the target) was measured. It was predicted that the participants would find accelerating targets particularly difficult^17,18,34,36 and that interception success and timing would improve as targets moved more slowly. It was also predicted that this pattern of responding would be preserved across equipment types.

Methods

Participants

Thirty-five healthy undergraduate students (24.5 ± 5.9 years, 69% female) were recruited from Trent University. The Trent University Research Ethics Board approved all procedures, and each participant gave written informed consent before participation. All were compensated monetarily or by receiving bonus grades. The Dutch Handedness Questionnaire³¹ revealed that 29 participants (83%) were right-hand dominant, and six participants (17%) were left-handed. All the participants reported having either normal or corrected-to-normal vision and hearing. Six participants stated that they had suffered at least one head injury in their life; however, all but one had been injured more than three years prior. None had lingering complaints and all reported a full recovery.

Apparatus

For the laboratory-based version, the participant was seated at the same apparatus as described above for the double-step task (Figure 1(a)). The projected display used for this task was a black background measuring 88 cm laterally and 67 cm in depth. The participants could not see their hands but received feedback about fingertip location from a display cursor driven by data collected from a motion-tracking marker attached to the fingertip of the dominant hand. The motion tracking data were collected at a sampling frequency of 100 Hz via the ProkLiberty toolbox for MATLAB and saved on a computer for offline analysis.

The lab version of the task was programmed in MATLAB using the Psychophysics Toolbox^32,33 version 3.0.8. Each trial began with the display of a blue square that indicated the starting location was positioned on the bottom right side of the screen; the trial did not proceed until participants moved the fingertip cursor into this start location. White crosshairs appeared above the start location and defined the interception zone. The interception target (a green oval) appeared 49 cm to the left side of the interception zone and then moved from left to right across the screen.

The tablet version was programmed with Xcode (https://developer.apple.com/xcode/). The participants completed the iPad tablet version in the same quiet, dimly lit room as was used for the laboratory version. The participants sat comfortably with the tablet placed in front of them on the table. The screen display on the tablet was 19.3 cm × 14.7 cm, and the interception target was a goldfish that appeared 18 cm to the left of the interception zone. Participants had full view of their reaching hand for the duration of the task (see Figure 3).

Figure 3.

Display used for the tablet version of the interception task. The fish target appeared on the left of the screen and then moved rightward across the screen. Participants placed their dominant-hand index fingertip on the red dot and made a quick reaching movement to intercept the fish as it moved through the interception zone (the light-gray region where the muted fish is seen). The target could move with three different speeds and with or without acceleration.

Design

We used a two-equipment (laboratory version, tablet version) by three-velocity (slow, medium, and fast) by two-acceleration (acceleration, no acceleration) within-subject design. Each participant completed the task on both equipment types. Equipment type order was counterbalanced. On the lab version, the levels of target velocity were 20 cm/s, 30 cm/s, and 40 cm/s, and on the tablet, the levels of target velocity were 13.5 cm/s, 21.4 cm/s, and 26.5 cm/s. Target viewing time was equalized to control for differences in screen size (see Table 1). As such, the participants viewed the targets for the same amount of time, and thus had the same amount of time to intercept it, for each equipment type. Both versions of the task had six blocks of 60 trials such that each combination of velocity and acceleration was repeated 60 times. The participants received feedback between each trial, which informed them if they had hit or missed the target.

Table 1.

Viewing times for the interception task on each equipment type.

	Acceleration present			Acceleration absent
Velocity	Slow	Medium	Fast	Slow	Medium	Fast
Laboratory (ms)	1120	707	560	746	600	507
Tablet (ms)	1120	706	568	810	599	507

Procedures

The participants entered the testing room and were positioned at the apparatus to which they had first been assigned. Each participant was instructed both verbally and in writing on how to complete each task, with identical instructions for both apparati. The participants were instructed to time their movement to the interception zone so that they hit the target on as many trials as they could. For each equipment type, they began by completing a practice block of 18 randomized trials to acquaint them with the task. The data from these trials were discarded. After completing testing on the initial apparatus, the participants were moved to the other side of the room where the second apparatus was set up. Upon completion of both equipment types, the participant was debriefed by the experimenter and was given pertinent study information and experimenter contact information.

Data analysis

Data processing for the interception task used both MATLAB software and SPSS. All data collected for the laboratory task were processed using MATLAB to determine whether the participant hit or missed the target and movement time error. On the tablet, the touch screen allowed measurement of the same variables.

A hit was coded as any trial on which the fingertip and any part of the target were in the interception zone at the same time. Movement time error was the difference between the time of the target’s arrival and the time of the finger’s arrival in the interception zone, with early reaches coded as a negative and late reaches coded as positive. Movement time error was only calculated on trials in which the participant missed the target.

Before these analyses were conducted, the data within each equipment type were trimmed based on participants’ timing error such that trials with an error greater than four standard deviations above the group mean were removed (1% of the data set). These variables were submitted to a two-order (laboratory first, tablet first) × two-equipment (laboratory, tablet) × three-velocity (slow, medium, and fast) × two-acceleration (acceleration, no acceleration) mixed ANOVA. The source of the main effects was determined by conducting planned pairwise (least significant difference) comparisons of means. Interactions were decomposed by conducting simple main effects analyses and then pairwise comparisons.

Results

Movement accuracy: Percentage of targets hit

The ANOVA revealed a main effect of equipment, F(1, 34) = 5.34, p < .05, indicating that the participants successfully intercepted more targets on the laboratory version of the task (M = 68% ± 2.7%) than on the tablet version (M = 62% ± 2.7%; Figure 4).

Figure 4.

Percentage of correct interception responses on slow, medium, and fast trials, as well as acceleration and non-acceleration trials. Both the laboratory equipment and the tablet data are shown. All comparisons between equipment, acceleration, and velocity are significant (p < .05).

The ANOVA also revealed a main effect of acceleration, F(1, 34) = 316.50, p < .001; nonaccelerating targets (M = 78% ± 2.0%) were hit more frequently than accelerating targets (M = 52% ± 2.7%). A significant interaction of equipment, acceleration, and order was also observed, F(1,34) = 5.93, p = .021. Simple main effect analysis revealed an equipment difference on acceleration trials, F(1, 34) = 18.66, p < .001, such that the participants intercepted more accelerating targets on the lab version (M = 59% ± 3.1%) than on the tablet (M = 45% ± 3.2%) and that this difference was bigger (but in the same direction) for people who performed the iPad task first. There was no difference between the lab (M = 77% ± 2.7%) and tablet (M = 79% ± 2.4%) on nonaccelerating trials, F(1, 34) = 0.15, p = .70.

Finally, the ANOVA revealed a significant main effect of velocity, F(2, 68) = 215.20, p < .001, with interception rate decreasing significantly with each step increase in target velocity (slow: 80% ± 1.9%, medium: 71% ± 2.4%, and fast: 44% ± 3.0%). There was also a significant interaction of acceleration and velocity, F(2, 68) = 15.48, p < .001. Simple main effect analysis revealed that the decline in success with increased velocity was steeper for accelerating trials (slow: 71% ± 2.9%, medium: 60% ± 3.3%, fast: 25% ± 2.7%) than on nonaccelerating trials (slow: 89% ± 1.7%, medium: 82% ± 2.1%, fast: 63% ± 3.7%). The interaction of equipment and velocity was not statistically significant, F(2, 68) = .62, p = .54, nor was the three-way interaction of equipment, acceleration, and velocity, F(2, 68) = .12, p = .89.

Movement time error

The two-equipment × three-velocity × two-acceleration ANOVA revealed a significant main effect of equipment, F(1, 34) = 79.41, p < .001, such that, on average, participants responded to the target 52 ms ± 6.39 ms late on the tablet version of the task, and 9 ms ± 6.01 ms early on the laboratory version. This effect of equipment interacted with order, F(1, 34) = 7.49, p = .01, such that this difference between the lab and tablet versions was 33 ms larger (but in the same direction) when participants completed the tablet version first. The ANOVA also revealed a significant main effect of velocity, F(2, 68) = 265.10, p < .001, such that the participants were more likely to be late when targets were moving at faster speeds (slow: 74 ms ± 5.21 ms, medium: 15 ms ± 5.69 ms, fast: –24 ms ± 6.24 ms).

While the equipment × acceleration did not reach significance, F(1, 34) = 4.00, p = .054, the equipment × velocity interaction was statistically significant, F(2, 68) = 14.47, p < .001. Simple main effects analysis revealed a significant effect of velocity on both the laboratory version of the task, F(2, 68) = 163.44, p < .001, and the tablet version, F(2, 68) = 190.92, p < .001. The differences between velocity levels were smaller on the tablet (26 ms between slow and medium and 57 ms between medium and fast) than on the laboratory version (54 ms between slow and medium and 60 ms between medium and fast; see Figure 5). A equipment × velocity × order interaction, F(2, 68) = 7.21, p < .01, revealed that this difference was more pronounced (but in the same direction) when participants completed the tablet task first.

Figure 5.

Mean time error for all speed and acceleration conditions of the interception task for both the laboratory equipment and the tablet. All comparisons between equipment, acceleration, and velocity are statistically significant (p < .05).

The ANOVA also revealed a main effect of acceleration, F(1, 34) = 795.57, p < .001. The participants were 14 ms ± 5.42 ms early on nonaccelerating targets and 58 ms ± 5.22 late on accelerating targets. A significant interaction of acceleration × velocity, F(2, 68) = 51.41, p < .001, was also present. The participants’ timing became increasingly late as velocity increased with the interaction driven by the finding that differences between velocity levels were smaller on accelerating trials (24 ms between slow and medium and 44 ms between medium and fast) than on nonaccelerating trials (56 ms between slow and medium and 73 between medium and fast).

Finally, this analysis revealed a significant three-way interaction of equipment, acceleration, and velocity, F(2, 68) = 7.31, p < .01 (Figure 5). Decomposition of the three-way interaction revealed that the participants were the most late on targets that were both accelerating and moving at the fastest speed. This pattern was observed on both equipment types with the participants averaging responses that are were 68 ms ± 5.49 ms late on fast accelerating targets on the laboratory version and were 123 ms ± 5.93 ms late on the tablet version. Thus, while the overall pattern of responding is similar across equipment types with regard to target velocity and acceleration, the effect of the fastest velocity and the presence of acceleration seemed heightened on the tablet.

We conducted a correlation analysis to determine the degree to which performance on the tablet accounted for performance on the lab version, an analysis made possible by the within-subject design employed here. The Pearson correlation coefficients are shown in Table 2 and indicate that, especially when acceleration is present and the task is most demanding, the participants’ performance on the tablet is significantly correlated to their performance on the lab version.

Table 2.

Correlations between performance on the tablet and lab versions.

	Tablet–Lab
	No acceleration	With acceleration	Overall
Hit rates	r = .298, p = .041	r = .444, p = .004	r = .427, p = .005
Time error	r = .402, p = .008	r = .410, p = .006	r = .409, p = .007

Task 3: The stop-signal task

The stop-signal task measures participants’ ability to interrupt simple, planned movements in response to an unpredictable stop signal.²⁴ Previous research suggests a possible correlation between head injuries and deficits in inhibitory control.^37–40 Indeed, people who suffer from deficits in inhibitory control may be in danger of further injury if they are unable to adjust their responses quickly and reliably to unpredictable stimuli. To measure this capacity, we developed a tablet application of the stop-signal task for use in our battery of visuomotor tasks.

The participant’s primary task was to make a fast response to the left or right in response to the visual presentation of a blue or green target circle, respectively. On 50% of trials, however, the participants heard an auditory tone (the stop signal) that indicated the need to cancel their response. The relative timing of the stop-signal tone was varied such that the duration between the go signal (visual target) and the stop signal (stop signal), the SOA, was set at 50, 100, 150, and 200 ms. It was predicted that the number of response errors (failures to inhibit the response) would increase with SOA because movements become more difficult to interrupt as execution becomes more likely.²⁴ We compared the pattern of response errors observed on the tablet to the pattern observed on a lab computer.

Methods

Participants

Forty-five healthy undergraduate students (20.8 ± 7.8 years (range: 17–59 years), 76% female) were recruited from Trent University. The Trent University Research Ethics Board approved all procedures, and each participant gave written informed consent before participation. All were compensated monetarily or by receiving bonus grades. The Dutch Handedness Questionnaire³¹ revealed that all but two participants were right-hand dominant. All participants reported having either normal or corrected-to-normal vision and hearing. Two participants who had suffered from head trauma within the past year were excluded from the study. All other participants who had received an injury previously reported no lingering complaints and a recovery from all associated symptoms.

Apparatus

The participants completed the tablet version of the task on the same tablet (iPad 2, Apple Inc., Cupertino, CA) described previously. The tablet task was programmed with Xcode (https://developer.apple.com/xcode/). The laboratory version was conducted using a PC computer (Bolen Systems, London, ON, Canada) and was programmed in MATLAB using the Psychophysics Toolbox^32,33 version 3.0.8. The displays appeared on an upright LCD screen (45.7 cm, 1280 × 1024 pixels), and participants provided their responses by pressing keys on a standard USB keyboard.

Task

Each participant was tested on both the tablet and the laboratory version of the task. Task order was counterbalanced across participants with 23 participants beginning with the laboratory task and 22 beginning with the tablet task. At the beginning of each trial on the laboratory version, a fixation cross was presented in the center of the computer screen and was replaced by either a blue or a green circular target after a variable foreperiod (1.5–2.5 s). The target remained on screen for 2 s. On go trials (50% of trials), the participants responded to the presentation of a blue or green target by pressing the “1” key or the “3” key with their second or fourth digit, respectively. On stop trials, a 900-Hz tone, 50 ms in duration, was emitted from the portable computer speakers after one of four delays (50, 100, 150, or 200 ms) relative to the initial presentation of the target. On stop trials, the participants’ task was to withhold their response (do nothing).

On the tablet, the participants were first presented with a red circle at the bottom of the screen that acted as the starting point. The participants touched the start circle, and after a variable foreperiod, a blue or green circle would appear at the top center of the screen. On go trials (50% of trials), the participants responded to the presentation of a blue or green target by tapping the left or right side of the screen, respectively. On stop trials, a tone was emitted from the tablet speakers at one of four delays (50, 100, 150, or 200 ms) relative to the initial presentation of the visual target. On stop trials, the participants task was to withhold their response (do nothing) by keeping their finger in contact with the red dot.

Design

We used a two-equipment (laboratory version, tablet version) by two-response side (left, right) by two-trial type (go trial, stop trial) design. Target side, trial type, and equipment were all within-subject independent variables. The fourth factor of the design, stop-signal delay, was nested within the stop-trial type. Stop-signal delay had four levels consisting of 50, 100, 150, or 200 ms.

Both versions of the task had six experimental blocks with 48 trials in each block. The dependent variable for go trials was RT (in milliseconds), and the dependent variable for stop trials was the percentage of trials on which participants successfully withheld their response.

Procedures

The participants entered the testing room and were positioned at the apparatus to which they had been first assigned. Each participant was instructed both verbally and in writing on how to complete each task, with identical instructions for both apparati. The participants were instructed to respond to each trial as quickly and as accurately as possible but to interrupt and prevent the button-press response when they heard the stop tone. Feedback regarding speed and accuracy was given after every block of trials. The participants completed a practice block of 18 trials on each equipment type to acquaint them with the task before continuing to the experimental blocks.

Upon completion of both equipment types, the participants immediately moved to complete the task on the other equipment type. The participants were debriefed at the end of the study.

Data analysis

Data processing for this task used both MATLAB software (MathWorks, Natick, MA), the output from the tablet, and SPSS (Version 20, SPSS Inc.). RT (measured in milliseconds) was trimmed to exclude values less than 100 ms and greater than 1500 ms (0.7% of the data removed). RT from go trials was collapsed across target side and submitted to a two-equipment (lab, tablet) repeated measures ANOVA. The percentage of correct withholding on stop trials was collapsed across response choice and submitted to a two-order (lab first, tablet first) by two-equipment (lab, tablet) by four-SOA (50, 100, 150, and 200 ms) repeated measures ANOVA. The source of main effects was determined using planned comparisons (pairwise comparisons), and interactions were decomposed using simple main effects analyses.

Results

On go trials, the participants were faster to respond on the lab equipment (M = 635 ms; SD = 2 ms) than on the tablet (M = 745 ms; SD = 2 ms), F(1, 43) = 139.87, p < .001. We also analyzed RT when failing to inhibit responses on stop trials and found no significant effect of equipment type, F(1, 94) = 1.24, p > .05. The timing of the participants’ failure to inhibit did not depend on equipment type: computer (M = 614 ms; SE = 20 ms) and tablet (M = 676 ms; SE = 16 ms).

We calculated the percentage of stop trials on which the participants were able to successfully interrupt their responses as a function of equipment type and SOA. Our analysis revealed no effect of equipment type, F(1, 43) = 2.25, p > .05. The participant’s percentage of successful inhibition on the computer (M = 95%; SE = 1%) was not different from the tablet (M = 97%; SE = 1%).

We did find, however, the expected relationship between successful inhibition and SOA on stop trials: inhibition was significantly less successful as SOA increased, F(3, 129) = 36.77, p < .05. The participants successfully inhibited more often when the SOA was 50 ms (M = 97%; SE = 1%) compared to when the SOA was 150 ms (M = 96%; SE = 1%) and 200 ms (M = 92%; SE = 1%). The participants were less able to correctly inhibit their actions as SOA increased, which agrees with the findings of others.⁴¹ This effect also interacted with order, F(3, 126) = 3.28, p = .023, such that the effect was stronger when participants performed the lab version first.

We also found a significant interaction between equipment type and SOA, F(3, 129) = 2.89, p < .05 (see Figure 6). The general pattern of performance on the two pieces of equipment was the same across SOAs with the exception of SOA of 200 ms. At an SOA of 200 ms, the participants successfully inhibited their actions more often on the tablet (M = 93%; SE = 1%) than on the computer (M = 90%; SE = 1%), p = .042. Importantly, this effect did not interact with the order in which participants experienced the task (p = .544).

Figure 6.

Mean percent correct and standard error on stop trials as function of equipment type and SOA. Performance declined significantly as the delay between the go signal and stop signal increased on both equipment types. Differences between equipment types were found at the 200-ms delay (*p < .05).

We conducted a correlation analysis to determine the degree to which performance on the tablet accounted for performance on the lab version. The overall Pearson correlation coefficient between the participants’ error rates on stop trials on the tablet and on the lab version was significant, r = .360, p = .001. As with the interception task, correlation coefficients shown in Table 3 indicate that in the condition in which the task is most demanding (SOA = 200 ms), the participants’ performance on the tablet is significantly correlated to their performance on the lab version.

Table 3.

Correlations between performance on the tablet and lab at each level of stimulus-onset asynchrony.

Tablet–Lab	Tablet–Lab	Tablet–Lab	Tablet–Lab
SOA = 50 ms	SOA = 100 ms	SOA = 150 ms	SOA = 200 ms
r = .025, p = .436	r = .013, p = .467	r = .139, p = .184	r = .462, p = .001

General discussion

The assessment of visuomotor function can provide important information about neurological status, but the availability of tests that combine accessibility and portability with spatial and temporal accuracy and precision has been limited. The success of the PoMLab,²⁸ a tablet-based visuomotor rotation learning application, demonstrated that at least some of the potential issues associated with shifting performance measurement to a tablet (like the smaller work-surface area and lack of control over the participants view of their limb) are reasonably surmountable problems. The goal for this project was to develop an assessment tool that can effectively measure aspects of visual processing and movement essential to healthy, active visuomotor performance.^16,23,42 The double-step task^11,12 measures the ability of the visuomotor system to track hand location relative to a visual target during reaching and to make movement adjustments in response to unexpected changes in target location. Interception of a moving target depends on the participants’ ability to predict target motion so that they can time their movement to hit the target.¹⁸ The stop-signal task measures how much time people need to interrupt preplanned movements in response to an unpredictable need to stop.²⁴ We reasoned that if the pattern of performance observed on lab versions of these tasks was also observed on the portable tablet platform, this would constitute a first step at validating the use of this portable battery in the assessment of visuomotor performance. On all three tasks, the pattern of responding observed in the lab (and in others’ labs) was preserved on the tablet versions of the tasks, suggesting that the tablet version assesses visuomotor processing in the expected manner.

The double-step task

The double-step task has been used in the past to assess how well one can update their movements toward a target that has been displaced shortly after movement onset (during the saccade to the target).^11,12,43 Studies of healthy participants show that people continuously assess the trajectory of their hand while it is moving toward a target and adjust their hand movement midflight in a smooth and reliable manner to reach for a target that has unexpectedly jumped from its initial location. The ability to update hand movements in response to visual perturbations is observed even when participants are not able to view their hand during the movement.¹²

When measuring movement accuracy, it was observed that target perturbations significantly increased movement error on both versions of the task and significantly increased movement time, relative to jump-absent trials, on the tablet. Although the original studies indicated no cost associated with adjusting to the target perturbation without vision of the limb is absent,^11,12 other studies are consistent with our findings of a small but significant increase in either error or movement time (see Gaveau et al.,⁴⁴ for a review). As was observed here, the performance time cost is especially associated with perturbations that demand movements be extended over greater distances. Interestingly, perturbations that demand movements of shorter distance, or even movement rehearsals, as in the case of perturbations that see the target jump back toward the start location, were associated with relatively shorter movement times, a good indication of early adjustment to the change.

The additional time needed to adjust to perturbations on the tablet may be because the participants could see their hand on the tablet version but could not see their hand on the lab version. When the participants have full vision of their hands during reaching, they make corrective saccades from the hand to the target to ensure an accuracy.^45–49 It is possible that the participants were engaging in more of these behaviors on the tablet version than on the laboratory version, where visual feedback of the hand was restricted. It is also possible that the differences we observed between the lab and tablet versions simply reflect differences between groups of participants. Importantly, the adjustments made on the lab version were not significantly different from the adjustments made on the tablet, suggesting good agreement between the two equipment types. The patterns of performance reported here demonstrate that the tablet version of the double-step task does a similar job at assessing adjustments to visuomotor perturbations as the laboratory-based version.

The interception task

The interception task measured how well participants were able to intercept targets that moved at different velocities and with or without acceleration. Previous research suggests that healthy participants can intercept targets moving horizontally at constant velocity well, even when targets are moving quickly. However, performance drops notably when the target accelerates (even when viewing time is controlled), as participants continue to predict constant velocity of target movement and consequently arrive too late to intercept the target.^17,18,34,50

Our analysis revealed that our participants’ interception performance was highly influenced by target velocity and acceleration, with the participants intercepting fewer targets and arriving at the interception zone increasingly later than the target as velocity increased and with acceleration present. As expected, hit rates decreased and timing error shifted substantially at all target velocities when acceleration was added to target movement.^17,18,34,36 Importantly, this pattern of results was observed across equipment types supporting the notion that both versions measure similar aspects of motion perception and visuomotor performance.

Although this general pattern was the same across equipment types, overall, performance was not as good on the tablet as on the lab version: the participants intercepted fewer targets and were consistently later on the tablet than in the lab. This was observed even though target viewing time was preserved across equipment types (see Table 1). These findings suggest that adjusting to quickly moving targets was more challenging on the tablet than in the lab. This may be due to the obvious differences in screen size. In the lab version, the target moves from peripheral vision into central vision, but because the tablet is smaller, the target spends less time in peripheral vision. Peripheral vision is coded primarily by rods that innervate the M-ganglion cells that synapse onto the magnocellular layers of the lateral geniculate nucleus. M-ganglion cells and their magnocellular targets are more sensitive to low-frequency spatial information, high-frequency temporal information and to visual motion stimuli.^51,52 In other words, the neural apparatus designed to detect early motion signals fires preferentially to motion stimuli in the periphery, which might lend some advantage to the lab version of the interception task. Perhaps because the tablet is smaller than the laboratory apparatus, the participants had a harder time responding to target motion because there is less of the display in their peripheral visual field, equating to later responses and less targets successfully intercepted.

Despite these differences, the overall pattern of response in the face of varying speed and acceleration was similar on both equipment types, suggesting that our tablet application is assessing the participants’ visuomotor function in the same way as the laboratory task is. Our correlational analysis corroborated this view, as the participants’ performance on the lab version was significantly positively correlated with their own performance on the tablet version. Interestingly, these correlations were stronger in the more difficult acceleration condition.

The stop-signal task

The stop-signal task measured the participants’ ability to interrupt a previously cued movement.²⁴ Previous research shows that as the delay between the signal to initiate the movement and the signal to stop the movement increases, the participants find it increasingly difficult to interrupt their responses.^24,41 We found this same pattern on both the tablet and the lab versions of the task. There were small differences between the two equipment types, as the participants were significantly less successful at interrupting an action at the longest delay on lab version of the task than on the tablet version. This difference may be due to differences in the method of responding on each equipment type. On the tablet, the participants responded by lifting their finger from a virtual start button on screen and making a small movement to touch the left or right side of the screen. By contrast, on the lab version, the participants responded by pressing a keyboard button on which their fingers rested. It is easier to press a button with one finger than it is to lift the entire hand to a new location, and it may be more difficult to interrupt the easier, shorter movement.

Despite this difference, the overall pattern of performance over the differing delay durations was similar on both equipment types, suggesting that our tablet and lab applications are assessing the participants’ inhibitory function in the same way. Our correlational analysis showed that the participants’ performance on the lab version was significantly positively correlated with their own performance on the tablet version in the most difficult (SOA = 200 ms) condition.

Conclusion

The general findings from these experiments indicate that our tablet versions measure visuomotor function in the same way as do typical laboratory versions of the same tasks. The differences between equipment types observed in our study do not obscure the expected pattern of performance, indicating the potential usefulness of tablet versions of the tasks. One area in which this application may be useful is in the field of concussion assessment, as individuals who have sustained a concussion show impairments to visuomotor performance.^9,10 Indeed, research shows that visuomotor performance is more sensitive to disruption than other cognitive tasks, showing significant performance deficits even after neuropsychological testing returns to normal. The availability of this tool may mean that this crucial assessment can be done without bringing the patient into the laboratory. This is important, as healthy visuomotor function is fundamental not only to sport performance but also to everyday function. To explore this further, future testing with this application will include a comparison between a healthy population and a clinical population who have sustained concussions. Other future studies will also determine the utility of the battery for sideline assessment (in the sporting arena) and for use in a rehabilitation training capacity by patients who are recovering at home.²⁹

Footnotes

Authors’ Note

The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.

Acknowledgments

The authors would like to acknowledge the programming contributions of David Hallin.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by a Trent University Internal NSRC award to LEB and HL and an NSERC Discovery Grant to LEB.

References

Gordon

Dooley

Wood

EP.

Descriptive epidemiology of concussion. Pediatr Neurol 2006; 34: 376–378.

Laker

Epidemiology of concussion and mild traumatic brain injury. Pm R 2011; 3(Suppl 2): S354–S358.

Signoretti

Lazzarino

Tavazzi

et al . The pathophysiology of concussion. PM R 2011; 3(Suppl 2): S359–S368.

Alla

Sullivan

Hale

et al . Self-report scales/checklists for the measurement of concussion symptoms: a systematic review. Br J Sports Med 2009; 43(Suppl 1): i3–i12.

Echemendia

Putukian

Mackin

et al . Neuropsychological test performance prior to and following sports-related mild traumatic brain injury. Clin J Sport Med 2001; 11: 23–31.

Broglio

Macciocchi

Ferrara

MS.

Neurocognitive performance of concussed athletes when symptom free. J Athl Train 2007; 42: 504–508.

McCrory

Meeuwisse

Aubry

et al . Consensus statement on concussion in sport: the 4th International Conference on Concussion in Sport held in Zurich, November 2012. Br J Sports Med 2013; 47: 250–258.

Killam

Cautin

Santucci

AC.

Assessing the enduring residual neuropsychological effects of head trauma in college athletes who participate in contact sports. Arch Clin Neuropsychol 2005; 20: 599–611.

Heitger

Anderson

Jones

et al . Eye movement and visuomotor arm movement deficits following mild closed head injury. Brain 2004; 127: 575–590.

10.

Heitger

Jones

Dalrymple-Alford

et al . Motor deficits and recovery during the first year following mild closed head injury. Brain Inj 2006; 20: 807–824.

11.

Prablanc

Pélisson

Goodale

MA.

Visual control of reaching movements without vision of the limb. I. Role of retinal feedback of target position in guiding the hand. Exp Brain Res 1986; 62: 293–302.

12.

Pélisson

Prablanc

Goodale

et al . Visual control of reaching movements without vision of the limb. II. Evidence of fast unconscious processes correcting the trajectory of the hand to the final position of a double-step stimulus. Exp Brain Res 1986; 62: 303–311.

13.

Ellermann

Siegal

Strupp

et al . Activation of visuomotor systems during visually guided movements: a functional MRI study. J Magn Reson 1998; 131: 272–285.

14.

Gréa

Pisella

Rossetti

et al . A lesion of the posterior parietal cortex disrupts on-line adjustments during aiming movements. Neuropsychologia 2002; 40: 2471–2480.

15.

Vindras

Blangero

Ota

et al . The pointing errors in optic ataxia reveal the role of “peripheral magnification” of the PPC. Front Integr Neurosci 2016; 10: 27.

16.

Desmurget

Epstein

Turner

et al . Role of the posterior parietal cortex in updating reaching movements to a visual target. Nat Neurosci 1999; 2: 563–567.

17.

Brouwer

Brenner

Smeets

JB.

Perception of acceleration with short presentation times: can acceleration be used in interception?

Percept Psychophys 2002; 64: 1160–1168.

18.

Brown

Wilson

Goodale

et al . Motor force field learning influences visual processing of target motion. J Neurosci 2007; 27: 9975–9983.

19.

Lisberger

Movshon

Visual motion analysis for pursuit eye movements in area MT of macaque monkeys. J Neurosci 1999; 19: 2224–2246.

20.

Maunsell

Newsome

WT.

Visual processing in monkey extrastriate cortex. Annu Rev Neurosci 1987; 10: 363–401.

21.

Barton

Sharpe

Raymond

JE.

Directional defects in pursuit and motion perception in humans with unilateral cerebral lesions. Brain 1996; 119: 1535–1550.

22.

Schenk

Mai

Ditterich

et al . Can a motion-blind patient reach for moving objects? Eur J Neurosci 2000; 12: 3351–3360.

23.

Bosco

Carrozzo

Lacquaniti

Contributions of the human temporoparietal junction and MT/V5+ to the timing of interception revealed by transcranial magnetic stimulation. J Neurosci 2008; 28: 12071–12084.

24.

Logan

Cowan

Davis

KA.

On the ability to inhibit simple and choice reaction time responses: a model and a method. J Exp Psychol Hum Percept Perform 1984; 10: 276–291.

25.

Hughes

Johnston

Fulham

et al . Stop-signal task difficulty and the right inferior frontal gyrus. Behav Brain Res 2013; 256: 205–213.

26.

Matthews

Simmons

Arce

et al . Dissociation of inhibition from error processing using a parametric inhibitory task during functional magnetic resonance imaging. Neuroreport 2005; 16: 755–760.

27.

Rieger

Gauggel

Burmeister

Inhibition of ongoing responses following frontal, nonfrontal, and basal ganglia lesions. Neuropsychology 2003; 17: 272–282.

28.

Takiyama

Shinya

Development of a Portable Motor Learning Laboratory (PoMLab). PLoS One 2016; 11: e0157588.

29.

Gonzalez-Franco

Gilroy

Moore

JO.

Empowering patients to perform physical therapy at home. Conf Proc IEEE Eng Med Biol Soc 2014; 2014: 6308–6311.

30.

Sharika

Ramakrishnan

Murthy

Control of predictive error correction during a saccadic double-step task. J Neurophysiol 2008; 100: 2757–2770.

31.

Van Strien

JW.

The Dutch Handedness Questionnaire, http://repub.eur.nl/res/pub/956/ (2003, accessed 22 August 2018).

32.

Brainard

DH.

The psychophysics toolbox. Spat Vis 1997; 10: 433–436.

33.

Pelli

DG.

The VideoToolbox software for visual psychophysics: transforming numbers into movies. Spat Vis 1997; 10: 437–442.

34.

Port

Lee

Dassonville

et al . Manual interception of moving targets. 1. Performance and movement initiation. Exp Brain Res 1997; 116: 406–420.

35.

Schenk

Ellison

Rice

et al . The role of V5/MT+ in the control of catching movements: an rTMS study. Neuropsychologia 2005; 43: 189–198.

36.

Zago

Bosco

Maffei

et al . Internal models of target motion: expected dynamics overrides measured kinematics in timing manual interceptions. J Neurophysiol 2004; 91: 1620–1634.

37.

Bonnelle

Ham

Leech

et al . Salience network integrity predicts default mode network function after traumatic brain injury. Proc Natl Acad Sci U S A 2012; 109: 4690–4695.

38.

DeHann

Halterman

Langan

et al . Cancelling planned actions following mild traumatic brain injury. Neuropsychologia 2007; 45: 406–411.

39.

Rochat

Beni

Annoni

et al . How inhibition relates to impulsivity after moderate to severe traumatic brain injury. J Int Neuropsychol Soc 2013; 19: 890–898.

40.

Stewart

Tannock

Inhibitory control differences following mild head injury. Brain Cogn 1999; 41: 411–416.

41.

Logan

Van Zandt

Verbruggen

et al . On the ability to inhibit thought and action: general and specific theories of an act of control. Psychol Rev 2014; 121: 66–95.

42.

Rieger

Gauggel

Inhibition of ongoing responses in patients with traumatic brain injury. Neuropsychologia 2002; 40: 76–85.

43.

Goodale

Pélisson

Prablanc

Large adjustments in visually guided reaching do not depend on vision of the hand or perception of target displacement. Nature 1986; 320: 748–750.

44.

Gaveau

Pisella

Priot

et al . Automatic online control of motor adjustments in reaching and grasping. Neuropsychologia 2014; 55: 25–40.

45.

Apker

Dyson

Frantz

et al . Efficiency of visual feedback integration differs between dominant and non-dominant arms during a reaching task. Exp Brain Res 2015; 233: 317–327.

46.

Gopal

Murthy

Eye-hand coordination during a double-step task: evidence for a common stochastic accumulator. J Neurophysiol 2015; 114: 1438–1454.

47.

Gopal

Murthy

A common control signal and a ballistic stage can explain the control of coordinated eye-hand movements. J Neurophysiol 2016; 115: 2470–2484.

48.

Lünenburger

Hoffmann

KP.

Arm movement and gap as factors influencing the reaction time of the second saccade in a double-step task. Eur J Neurosci 2003; 17: 2481–2491.

49.

Saunders

Knill

DC.

Humans use continuous visual feedback from the hand to control fast reaching movements. Exp Brain Res 2003; 152: 341–352.

50.

Tresilian

Oliver

Carroll

Temporal precision of interceptive action: differential effects of target size and speed. Exp Brain Res 2003; 148: 425–438.

51.

Connolly

Van

ED.

The representation of the visual field in parvicellular and magnocellular layers of the lateral geniculate nucleus in the macaque monkey. J Comp Neurol 1984; 226: 544–564.

52.

Van Essen

DeYoe

Concurrent processing in the primate visual cortex. In: Gazzaniga

(ed) The cognitive neurosciences. Cambridge, MA: MIT Press, 1995, pp. 383–400.