The staircase method was used to calculate the just noticeable difference (JND) in the visual discrimination task. The staircase is a psychophysical procedure for estimating perceptual thresholds, which makes the judgments harder or easier as a function of the correctness of the last trials. In this study, the magnitude of distracting patterns (the difference between the 2 displayed coordination patterns) was determined by a 1-up or 2-down rule defined in the staircase method. Considering the time constraint for a screening task and based on a previous study on the perception of coordination patterns [38], we applied a step size “up” of 1° and a stop rule of 3 reversals (errors) to the in-phase, and a step size “up” of 5° and a stop rule of 2 reversals (errors) to the anti-phase. Step size “down” was fixed based on Table 5.1 of Kingdom and Prins [39]. Thus, it was 0.55° for in-phase and 2.74° for anti-phase. The initial difference was set to 5° and 40° for in-phase and anti-phase, respectively. Each time the trial stepped down until the first error occurred, then it repeated until meeting the stop rule. In addition, the coefficient of variation (CV) was calculated for JNDs collected from the 6 assessments, respectively, for in-phase and anti-phase patterns to examine how consistent the participant was able to discriminate intrinsic coordination patterns.

In unimanual and bimanual coordination tasks, the proportion of time-on-task (PTT) was automatically calculated by the system and used to quantify the coordination performance. Specifically, the proportion of each continuous relative phase time series (trial) that fell within the range of the target phase ±20° tolerance was computed. PTT has been proven as a valid measure of performance at the required relative phase that can capture both accuracy and stability of coordination performance in numerous studies [32,33,40].

To examine the group difference in each of the eye-hand coordination tasks across the 6 testing sessions, the mixed-design ANOVA was performed on JNDs in the visual discrimination task and mean PTTs in both unimanual and bimanual coordination tasks, respectively, treating group (group with a history of concussion vs healthy controls) as a between-subject variable and testing sessions (1 through 6) as a within-subject variable. Prior to conducting ANOVAs, the normality of outcome variables was assessed using the Shapiro-Wilk test and all variables met the normality assumption. For post hoc pairwise comparisons following the detection of significant ANOVA effect, we applied Tukey HSD (Honestly Significant Difference) corrections to adjust for multiple comparisons. Since the study purpose was to examine the use of eye-hand coordination tasks to detect the postconcussion cognitive-motor deficits, we were more interested in the main effect of group, and the main effect of testing sessions would indicate only the learning effect associated with participants being familiar with the assessment tasks. In addition, the 2-tailed independent samples t test was performed on CV of JND at in-phase and anti-phase patterns, respectively.

Finally, due to the small sample size that might have limited the power of the study to detect the effects of interest, Cohen d effect size was calculated for each investigated factor. According to Cohen [41], d=0.20, d=0.50, and d=0.80 were considered as “small,” “medium,” and “large” effect sizes, respectively. A large effect size would indicate a strong effect to be detected in the future by a larger sample size.

The 2-way mixed-design ANOVA on mean PTTs yielded a main effect for target coordination pattern (F_2,36=61.35, P<.001, and Cohen d=3.52). As seen in Figure 5, in general, performance of in-phase coordination (mean PTT 0.68, SD 0.12) was significantly better than that of anti-phase coordination (mean PTT 0.47, SD 0.17) and then that of 90° coordination (mean PTT 0.30, SD 0.14). Significant 2-way interaction between group and target coordination pattern was detected (F_2,36=12.60, P<.001, and Cohen d=1.54). Simple main effect analysis showed that healthy controls outperformed participants with a history of concussion in performing in-phase (F_1,54=4.20, P=.045, and Cohen d=0.48; mean PTT_Concussed=0.62, SD 0.13; and mean PTT_Healthy=0.74, SD 0.07) and anti-phase coordination patterns (F_1,54=10.26, P=.002, and Cohen d=0.81; mean PTT_Concussed=0.37, SD 0.15; and mean PTT_Healthy=0.56, SD 0.14), while the participants with a history of concussion outperformed healthy controls only in performing 90° coordination (F_1,54=6.36, P=.015, and Cohen d=0.62; mean PTT_Concussed=0.37, SD 0.13; and mean PTT_Healthy=0.23, SD 0.12). To be noted, the 2 groups demonstrated different patterns in performing the 3 coordination patterns. As revealed by simple main effect analysis followed by post hoc of Tukey HSD test, healthy controls (F_2,54=41.20, P<.001, and Cohen d=2.38) performed in-phase coordination (mean PTT 0.74, SD 0.07) significantly better than anti-phase coordination (mean PTT 0.56, SD 0.14), while their performance of 90° coordination (mean PTT 0.23, SD 0.12) was significantly worse than that of both in-phase and anti-phase patterns. However, this differentiated performance between the 3 coordination patterns (as healthy controls showed in Figure 5) was absent for participants with a history of concussion (F_2,54=12.57, P<.001, and Cohen d=1.27) and their performance of in-phase coordination (mean PTT 0.62, SD 0.13) was better than that of anti-phase (mean PTT 0.37, SD 0.15) and 90° coordination (mean PTT 0.37, SD 0.13), with no difference detected between the latter 2, suggesting that their cognitive-motor system could not differentiate the 2 coordination patterns involving displacement of dots.

In experiment 1, three coordination tasks have been implemented to examine the comprehensive cognitive-motor functions of people with and with no history of concussion, and among the 10 task measures, 4 measures have been identified with a great effect size to capture the group difference (Table 2). The results suggest that participants with a history of concussion demonstrated greater variability in perceptual discrimination of the in-phase coordination pattern and worse motor performance in both unimanual and bimanual coordination tasks than healthy controls. These deficits were consistently seen in the following bimonthly sessions, suggesting persistent cognitive-motor impairments in individuals with prior concussion.

Experiment 2 was conducted using an iterated protocol based on the results of experiment 1 and examined the possibility of using the unimanual coordination task alone to detect the cognitive-motor deficits associated with concussion. Compared with the protocol used in experiment 1, the current testing protocol could be completed within 10 minutes (instead of an hour) with the group difference being captured in just 1 testing session. The results showed that unimanual coordination tasks were sensitive to detecting performance deficits in participants with a history of concussion in that individuals with a history of concussion performed worse on both in-phase and anti-phase tasks and failed to differentiate between anti-phase and 90° coordination patterns.

In experiment 1, the visual discrimination task was designed to assess cognitive functions of coordination, which include attending to and tracking the 2 moving dots on the screen, memorizing the spatial-temporal pattern of their movements, and then identifying the target coordination pattern from those distracting patterns. Although these cognitive functions have been reported to be impaired by concussion in previous studies [42-44], participants in this study did not show group differences in discriminating the intrinsic coordination patterns, suggesting that our participants with a history of concussion have recovered the cognitive functions required for perceiving the fundamental coordination tasks such as walking and jumping. This is conforming to the literature that most cognitive dysfunctions associated with concussion spontaneously recover to baseline within 2 weeks [45]. Nevertheless, the analysis on the CV of JNDs revealed that our participants with a history of concussion were more inconsistent in discriminating the in-phase pattern than their healthy counterparts, suggesting that the ability to consistently detect the minor decoupling between 2 oscillators (eg, minor asynchronization of foot takeoff in jumping) was impaired following concussion despite the spontaneous recovery of the ability to perceive intrinsic coordination patterns.

The unimanual coordination task was designed to examine cognitive-motor functions of coordination, specifically, the ability to produce and maintain a rhythmic movement to interact with a computer based on the established perception of coordination pattern. Such an ability not only requires the aforementioned cognitive functions to detect and retain the coordination pattern but also demands motor functions to move one hand in coordination with the moving dot on the screen and correct any perceived movement error to maintain the coordination pattern. The group difference was captured when motor functions were demanded together with cognitive functions. Our healthy controls outperformed participants with a history of concussion in performing both intrinsic coordination patterns. Such a finding conforms to a previous study [46] suggesting that the deficits of dynamic motor functions (eg, visuomotor tracking and walking under various conditions of attention) persist following the concussion despite the spontaneous recovery of cognitive functions. As for the performance of the novel coordination pattern, both participants with a history of concussion and healthy participants improved over the testing sessions without showing any group difference, suggesting that the unimanual coordination task was quite challenging in nature, and the ability to learn new coordination pattern remained intact postconcussion.

Finally, the bimanual coordination task was used to examine the motor functions of coordination. Compared with the unimanual task, the bimanual task demands the motor production of the coordination pattern entirely by requiring participants to move 2 hands rhythmically; therefore, the independent visuomotor control and hemisphere cross talk are both challenged. The results showed the group difference only in performing the in-phase pattern with healthy controls outperforming the participants with a history of concussion. Such a finding not only supported previous studies suggesting that bimanual coordination deficits are associated with concussion [47] but also indicated that the concussion-induced malfunction of CC might have caused the problem for spatial and temporal coupling of 2 hands, which is typically seen in patients with split brain or callosotomy [48]. Compared with the anti-phase and novel coordination patterns, the in-phase pattern demands more spatial and temporal coupling of 2 hands. Therefore, people with a history of concussion are more likely to be screened out by the in-phase bimanual coordination task.

In sum, all 3 coordination tasks have been proven effective for discriminating people with a history of concussion with at least 1 task measure that could be used. Interestingly, most effective task measures are related to the ability to perceive or perform the in-phase pattern, suggesting that the cognitive-motor deficits associated with concussions may be specifically tied to the inability to detect and maintain spatiotemporal coupling in performing eye-hand coordination tasks. The unimanual task emerged to be a good candidate for screening people with a history of concussion because (1) half of effective task measures were found in the unimanual task with the largest effect size to detect group difference, (2) it examines the cognitive-motor functions required for both perceiving and performing coordination patterns, and (3) the performance can be quickly assessed and used for evaluation. Consequently, experiment 2 was designed to examine whether the unimanual task alone can serve as a quick and efficient tool for detecting the chronic cognitive-motor deficits following concussion.

In experiment 2, the unimanual coordination task entails the integration of cognitive-motor functions including attention, memory, and visuomotor tracking, for which one needs to successfully detect the spatiotemporal relationship between the 2 moving dots on the screen and then adjust the manual movement to maintain that pattern. According to the perception-action theory of coordination [29,49], the relative direction of 2 oscillators is used in perceiving and performing both in-phase and anti-phase coordination patterns, while the relative position of 2 oscillators is used for the 90° coordination. Since in-phase and anti-phase patterns represent the coordination movements with the relative direction always being identical (ie, in-phase) or opposite (ie, anti-phase), they are distinctive and easy to be perceived and produced. In contrast, the relative position is ambiguous in 90° coordination due to the fluctuations in the movement direction, which makes it difficult for both perception and production [50]. Both perceptual and motor studies on rhythmic coordination using healthy participants [31,35,38,51] have shown that the performance of 90° coordination is worse than that of in-phase and anti-phase, with the former outperforming the latter (the differentiated performance between 3 coordination patterns as seen in Figure 5 for healthy controls).

In both experiments when unimanual coordination performance was assessed, we found consistently that healthy controls outperformed participants with a history of concussion in producing both in-phase and anti-phase patterns while maintaining the differentiated performance between all 3 coordination patterns, suggesting that they were sensitive to the information of relative direction that is required for perceiving and performing the intrinsic coordination patterns, although having difficulty in finding the information of relative position for successful perception and production of the novel coordination pattern due to lack of practice. As for the participants with a history of concussion, they showed a significantly degraded performance at both in-phase and anti-phase while losing the performance difference between anti-phase and 90°, suggesting that the previous concussion might have limited their ability to discriminate and use the appropriate spatiotemporal information for successful production of both intrinsic and novel coordination patterns. As shown in experiment 1, the participants with a history of concussion did not have a problem with visual discrimination of in-phase and anti-phase patterns; therefore, their degraded performance in performing unimanual in-phase and anti-phase patterns had to do with the deficit in perception-action coupling. Although their performance at 90° was seemingly better than the healthy controls, such a performance was not different from their performance in producing the anti-phase pattern, suggesting that anti-phase and 90° coordination patterns may have been treated as the same by the individuals with a history of concussion since they both involve the displacement and opposite relative direction of moving oscillators. Hence, their seemingly better performance at 90° was a false positive, stemming from their confusion between 90° coordination and anti-phase coordination.

Although more studies are warranted to prove the impaired ability to discriminate and use spatiotemporal information for coordination by the participants with a history of concussion, experiment 2 overall showed that participants with a history of concussion demonstrated deteriorated cognitive-motor functions in performing the unimanual in-phase and anti-phase patterns, which could be attributed to the dissociation between perception and action. In addition, the conventional differentiated performance between the 3 coordination patterns (in-phase > anti-phase > 90°) was not seen for participants with a history of concussion, which suggests that the differentiated performance between the intrinsic and novel coordination patterns in the unimanual coordination task could potentially serve as a useful indicator for screening concussion.

This study has several limitations that should be acknowledged. First, although 2 experiments were conducted, the sample size in each was relatively small, which may limit the generalization of both study findings. While several outcomes demonstrated large effect sizes for group main effect (Cohen d>0.80), which suggest a possibility of using a combined approach for assessment and screening, a larger sample size is warranted in future studies to prove the effectiveness and reliability of using eye-hand coordination tasks to detect the chronic or subacute cognitive-motor deficits associated with concussion.

Second, there was a wide range of the reported time from the last concussion. The reported last concussion occurrence ranged from 2 to 96 months in experiment 1 and from 4 to 54 months in experiment 2. As a result, participants could be at different stages of spontaneous recovery from concussion. Also, during that time, participants could be reinjured by concussion that was not medically examined. Future studies could improve on this by focusing on participants within narrower postinjury time frames or by using recovery stage as a covariate in analyses.

Finally, poor eye-hand coordination could result from neurocognitive disorders other than concussion (eg, posttraumatic stress disorder and attention-deficit/hyperactivity disorder). Although all participants in this study were college students with ongoing study programs, and they self-reported to have no history of substance abuse, peripheral neuropathy, and neurological disorders, there is still a possibility for undiagnosed neurocognitive disorders other than concussion to impact the eye-hand coordination performance. In future work, standardized screening tools or neuropsychological assessments could be included to help isolate the impact of concussion from other potential influences on cognitive-motor performance.

This study provides preliminary evidence for using the computerized eye-hand coordination tasks to detect the chronic cognitive-motor dysfunctions associated with concussion. In developing neurobehavioral tools to capture the postconcussion chronic cognitive-motor deficits, the unimanual coordination task should be considered, and the differentiated performance between intrinsic and novel coordination patterns should be examined.