Traumatic brain injury (TBI) is one of the most challenging problems faced by the medical community. It is estimated that 5.3 million Americans currently live with disabilities resulting from brain injury. The cost of TBI in the United States is estimated to be $56 billion annually. Sports and physical activity provide significant exposures to TBI with some 300,000 sports-related brain injuries estimated to occur each year in the United States (1). More than 90% of these injuries fall into the category of cerebral concussion or mild traumatic brain injury (mTBI). Although many sports-related injuries are considered to be mTBI, an estimated 900 brain injury deaths per year occur in sports and recreational activities. Mild traumatic brain injury is not synonymous with insignificant brain injury as there is considerable individual variation in rate of recovery, a relatively high rate of postconcussion syndrome after mTBI, and potentially serious consequences of repeated mTBIs (11).
Despite the general understanding of the effects of mTBI from a clinical point of view, the basic mechanisms underlying this injury and how these mechanisms interact to lead to the deficits observed at a behavioral level are poorly understood. Over the past several years, our research group has undertaken a series of studies that have directly addressed this issue. In particular, we have examined how sensorimotor deficits associated with dynamic stability during gait are related to attentional and executive dysfunction after mTBI. Our basic observation is that both attentional and biomechanical performance deficits are exacerbated as task complexity increases and that even with a very mild TBI (in which the participants are asymptomatic 1-2 wk after the injury and did not lose consciousness at the time of the injury) certain deficits remain at least 1 month after the injury. Our working model (Fig. 1) postulates that mTBI reduces the capacity of executive function making it difficult to easily switch between two simultaneous tasks. In addition, processing speed is also slowed resulting in longer times spent in the act of switching. Finally, attention itself is also deficient so that the demands of each task when combined exceed the maximum capacity of attention. Together, these deficits result in difficulties appropriately performing one or both tasks under dual task conditions. Such deficits are important from a clinical point of view because they tap into situations that are relevant to everyday activities. In particular, an individual with mTBI may appear normal when attempting activities in isolation yet display noticeable deficits when performing two or more tasks simultaneously. Such hidden deficits should be taken into account when making decisions about when to begin normal activities after the injury. This is especially important in return to play decisions for athletes who participate in sports with a significant risk of reinjury such as football or ice hockey.
ATTENTIONAL DEFICITS IN MTBI
People suffering an mTBI typically display attentional deficits. In particular, previous studies have revealed difficulties associated with maintaining and distributing attention within and between tasks in patients with mTBI (2). Attention itself can be broken down into several different components that are dissociated and can be mapped onto specific circuits within the brain. Posner originally suggested that the spatial orienting of attention is comprised of disengagement, movement, and reengagement processes (9). Clinical studies in patients with brain damage and functional imaging studies in healthy subjects have demonstrated that these different attentional components engage portions of the parietal, frontal, temporal, and cingulate cortices and midbrain in various combinations (10). For example, lesions to the parietal lobe, but not to the frontal or temporal lobes or midbrain, result in deficits in the disengagement from locations at which subjects are paying attention. By contrast, the movement and reengagement of attention are thought to be mediated by activity in the superior parietal lobule and intraparietal sulcus of the posterior parietal cortex, the frontal eye fields, and cingulate gyrus.
Given the specificity of certain brain regions to the different aspects of attentional processing, it appears to be possible to gain insight into the relationship between functional deficits after mTBI and the areas of the brain that may be most susceptible to the injury process. We have addressed this relationship by using the attentional network test (ANT) recently developed by Fan et al. (6) to probe the effects of mTBI on the alerting, orienting, and executive components of attention within 2 d of the injury (12). The ANT combines in a single session a number of task types that have been used to validly test these aspects of attention in isolation. In this task, the subject visually fixates on a central target that automatically engages attention and responds to the appearance of an arrow pointing to the left or right by pressing on the corresponding button of a mouse (Fig. 2A). The arrow can be preceded by a center or double precue that provides temporal information about when the arrow will appear (i.e., a set period after the precue) or a spatial precue that provides both temporal and spatial information about when and where the arrow will appear (i.e., the precue and arrow appear at the same location in space) (Fig. 2B). In addition, the arrow itself can appear in isolation or surrounded by other arrows that point either in the same, congruent direction or the opposite, incongruent direction (Fig. 2C).
The alerting component of attention is associated with the ability to maintain vigilance or arousal during continuous task performance. The orienting component of attention contributes to the ability to covertly direct visual or other sensory processing resources to a particular region of space so that targets subsequently presented there are detected more quickly and/or more accurately. Finally, the executive component of attention allows us to switch between different task demands easily and resolve contextual conflict appropriately. It is typically probed in experimental settings using the Stroop, flanker, or set-switching tasks to examine the ability of the subject to make use of or ignore extraneous information.
Using the ANT we found that mTBI affects the orienting and executive, but not the alerting, components of attention. In general, participants with mTBI (n = 20) were slower overall than an age-, gender-, education-, and activity-matched group of controls (n = 20)-a finding that is consistent with previous research suggesting slowed information processing speed in this population. This slowing was apparent in the alerting effect for trials both with and without temporally informative precues (Fig. 3A). However, even though providing these double precues diffused attention across the two target locations, they also shortened the median reaction time by a similar amount in both subject groups by alerting them to the time at which the target arrow would appear. Thus, suffering an mTBI does not appear to differentially alter the ability to make use of an alerting precue to reduce reaction time. It may seem counterintuitive to be slower overall but not deficient in alerting. However, it is important to remember that alerting is one of several aspects that form the overall reaction time. These other aspects include sensory/perceptual processing, sensorimotor integration or mapping, and motor output. The implication is that mTBI influences some of these aspects to cause slower reaction times, while leaving alerting unscathed. By contrast, data from the orienting component of attention demonstrated that participants with mTBI were disproportionately slower than controls when central precues were presented providing only temporal information and focusing attention at the central fixation point; compared to when peripheral precues were presented providing both temporal and spatial information (Fig. 3B). This was confirmed statistically by a significant group x precue condition interaction. Thus, without the information provided by the spatially informative precues, participants with mTBI took a markedly longer time to disengage attention from the central fixation point, search alternative spatial locations, and reengage attention at the appropriate location. Finally, for the executive component of attention, we found that the increase in median reaction time for trials with the incongruent target configuration compared to congruent target configuration was similar in both subject groups (Fig. 3C). However, further probing of these data revealed that group differences were apparent when median reaction times of accurate versus inaccurate responses were examined. In particular, generating accurate responses took disproportionately longer for participants with mTBI than controls. By contrast, the median reaction times for inaccurate responses were similar for both groups (Fig. 3D). These differences were not due to any speed-accuracy trade-offs: both participants with mTBI and controls had similar levels of accuracy across trials. Thus, a closer analysis of the response characteristics in trials probing the executive component of attention suggests that mTBI produces a subtle yet systematic effect on this aspect of attention. We have subsequently collected data over the course of 1 month postinjury, which demonstrates that the deficit in the orienting component of attention resolves within 1 wk, whereas that for the executive component is still present after 1 month (13). Thus, by using the ANT, we were able to demonstrate that the general attentional deficits typically noted in mTBI are due to dysfunction specifically to the orienting and executive components of attention, whereas the alerting component remained unaffected.
To further clarify the influence of mTBI on the orienting component of attention, additional preliminary data were collected for a gap saccade task. In this task, a central fixation target is presented for a variable period after which it disappears and is replaced by a peripheral target 10 degrees to the left or right of center. The subject is required simply to make a saccadic eye movement to this new target as quickly and accurately as possible. On some trials, a short delay (the gap) is inserted between the disappearance of the central fixation target and the presentation of the peripheral target. This delay was varied between 50 and 250 ms on separate trials. On the remaining trials no gap was present. Previous research using this task in normal, healthy young subjects has demonstrated that the gap leads to a significant reduction in the time required to initiate a saccadic eye movement. This is thought to occur because the central fixation target automatically engages attention. When no gap occurs between the disappearance of the fixation target and the peripheral target, the subject must then disengage attention from the fixation target, move it to the peripheral location, and reengage it at this location in a similar manner to that which is required during orienting trials in the ANT. However, when a temporal gap is present, the disengagement of attention can occur before the appearance of the peripheral target and saccadic reaction time is reduced accordingly. Thus, this task provides a means to examine one aspect of the orienting effect-namely the disengagement of attention-in isolation from the others (movement of attention to the peripheral location, and reengagement of attention at that location). We found that within 2 d of the injury, participants with mTBI (n = 20) showed much slower saccadic reaction times on the no-gap trial relative to controls (n = 20), whereas their reaction times progressively approached that of controls as gap duration increased (Fig. 4). Given that the contribution of the disengagement process decreases as gap duration increases, this implies that the performance of participants with mTBI normalized as the disengagement process was removed. This, in turn suggests that the difficulty participants have orienting attention (Fig. 3B) are due a deficit, or slowing of, disengagement, while moving and reengaging attention appears largely intact.
These deficits in attentional tasks were reflected in alterations in the pattern of brain activation in symptomatic participants with mTBI as assessed by other research groups with functional magnetic resonance imaging. In particular, comparison of a spatial working memory task to rest revealed a smaller activation in the right prefrontal cortex in patients with mTBI relative to controls despite equivalent behavioral performance, whereas some individual patients displayed additional activation outside the regions normally associated with working memory tasks (3). However, as working memory load was increased by adding more items to be remembered, patients with mTBI displayed disproportionately larger increases in the magnitude of activation in the right parietal and dorsolateral frontal areas and additionally in the cerebellum when the task was a movement sequence produced from memory. These data suggest that participants with mTBI must activate a larger swath of cortex to achieve equivalent levels of task performance.
BIOMECHANICAL ASSESSMENT OF GAIT STABILITY IN MTBI
Biomechanical studies of individuals with TBI have, for the most part, been limited to postural sway during quiet standing or during standing with altered sensory inputs. A number of studies have reported static postural stability deficits lasting several days after sport-related mTBI (7). We have extended this approach by examining dynamic stability in participants with mTBI during gait under a variety of conditions. Maintaining whole body balance in the frontal plane during gait is a difficult task due to the limited base of support during the single support phase. An active hip abduction moment (torque) about the supporting leg plays a crucial role in maintaining balance of the trunk and swing leg during normal walking. Interaction between the whole body center of mass (COM-measured using a motion analysis system) and center of pressure (COP, i.e., a point on the base of support at which ground reaction forces can be considered to be concentrated-measured using force platforms embedded in the walking surface) is tightly regulated to control total body balance during gait initiation and termination. The instantaneous velocity of the COM and its location with respect to the COP are important factors in maintaining balance during standing. The horizontal distance between the COM and COP of the stance foot may be one of the factors that determine the external moments acting at the joints of the supporting limb.
The inclusion of an obstacle to crossover makes this task even more challenging. When stepping over an obstacle, the longer swing time required for the swing limb implies a longer duration of single stance for the supporting limb (4). Imbalance of the whole body during obstacle crossing may cause inappropriate movement of the lower extremities or striking an obstacle with the swing foot, and result in a fall. Greater and faster motion of body segments while negotiating an obstacle will result in greater and faster movement of the COM and perturb balance maintenance. Therefore, proper control of the COM motion and its coordination with the COP of the stance foot are important for the maintenance of the dynamic stability of the whole body when stepping over obstacles.
It was demonstrated that participants with moderate TBI (loss of consciousness, posttraumatic amnesia, Glasgow Coma Scores: 9-15) were unable to perform this obstacle-crossing task normally within 2 yrs of their injury (5). In particular, these participants (n = 10) demonstrated significantly greater COM sway and reduced COM range of motion during the crossing stride over the obstacle compared to control subjects (n = 10). The peak COM velocities during the weight-shifting period for participants with TBI were significantly faster than that of control subjects. Similar to walking speed, the peak forward COM velocities in the participants with mTBI were significantly slower. Finally, the maximum distances between the whole body COM and the corresponding COP during single stance periods were significantly greater in participants with TBI compared to the controls (Fig. 5).
Further data analysis was performed to examine whether the greater COM sway was systematically related to the slower gait speed. This was accomplished by including gait speed during unobstructed walking as a covariate. The results indicated that there was no significant association between the greater COM sway and slower gait speed. Therefore, the greater COM motion in participants with TBI indicates difficulty in maintaining dynamic stability in the frontal plane, which may directly reflect their sensation of instability. Smaller whole body COM displacement and peak instantaneous velocity in participants with mTBI are primarily due to their adoption of a significantly slower walking speed and a shorter stride length compared to controls. It has been demonstrated previously that both the distance between the COM and the COP and the instantaneous velocity of COM are critical for the ability to successfully arrest forward momentum and maintain dynamic stability in the sagittal plane. Thus, in participants with mTBI the feasible range of COM movement during which dynamic balance can be maintained has been reduced, that is, only a slower COM forward velocity is permissible at a given separation distance between the COM and the COP.
Taken together, these findings demonstrate that examining the COM motion provides an objective measurement that reflects complaints of instability not observable with a clinical examination for individuals who have suffered an mTBI. Thus, given sufficient resources inclusion of an obstacle-crossing task in clinical gait analysis may enhance the assessment of dynamic instability in this population.
INTERACTION OF DEFICITS IN MTBI
Given that participants with mTBI have deficits in the ability to orient attention across space and between tasks and are generally slower in processing information, we asked whether the gait stability difficulties this population has could be partially due to a breakdown in the interaction between attentional and gait control processes (8). This was accomplished by examining gait stability during unobstructed level walking either in isolation or while concurrently performing an attentionally demanding cognitive task, such as "naming the months of the year backward" or "continuously subtracting 7 backward from 100."
Relative to the controls (n = 10) the participants with mTBI (n = 10) adopted a more conservative gait strategy within 2 d of the injury-walking more slowly and with less sway. Performing the secondary task altered the gait of both participants with mTBI and controls causing significantly slower gait velocity, shorter stride length, longer stride time, smaller range of motion of the COM, and peak anterior velocity of the COM than during unobstructed level walking performed in isolation. Performance on the secondary tasks in terms of speed and accuracy of the responses was equivalent in both groups. Importantly, the participants with mTBI displayed a significant increase in the range of motion of the COM in the mediolateral direction when performing the secondary task, whereas the controls did not (Fig. 6). Moreover, preliminary data collected to examine recovery from the injury demonstrated that this effect was still present 1 month later.
Taken together, these results revealed that participants with mTBI were able to conservatively adjust their whole body COM motion to maintain gait stability without divided attention. They accomplished this with a significantly slower instantaneous forward velocity and significantly smaller mediolateral sway. Walking with a concurrent cognitive task resulted in significant changes in gait and COM measurements for both groups, but led to proportionally larger increases in mediolateral COM sway in the participants with mTBI. In other words, they were unable to maintain appropriate gait stability when confronted with the more cognitively demanding dual task situation. We believe that this is due to the general slowing in processing speed and the reduced ability to orient and switch attention from one task to another as measured directly with the ANT.
Many individuals with mTBI complain of problems with balance and instability that may not be apparent on clinical examination. In addition, attentional problems are also common in this patient population and our research demonstrates that these problems can be restricted to a subset of attentional processes. Moreover, the deficits in this population tend to be exacerbated when two or more tasks are performed at the same time. Such covert deficits are especially important to consider when making decisions about when it is safe for a player in a contact sport such as ice hockey or football to return to playing. Like athletes in many sports, hockey and football players must make quick decisions based on a multitude of information to be successful and to avoid injury. If this decision-making process is disrupted by an mTBI from which the athlete has not fully recovered, then athletes may increase their chances of exposing themselves to a subsequent reinjury. Assessments of recovery from mTBI should include dual-task protocols in addition to neuropsychological and balance testing and close monitoring of symptomatology so that the best decision is made regarding return to play or normal activities.
The authors wish to thank Robert Catena, Anthony Drew, Charlene Halterman, Jeanne Langan, Tonya Parker, and Erika Rodriguez for their help with this project. This study is supported by CDC grant R49-CCR021735.
1. Centers for Disease Control. Traumatic Brain Injury in the United States. A Report to Congress, 1999.
2. Chan, R., R. Hoosain, M. Lee, W. Fan, and D. Fong. Are there sub-types of attentional deficits in patients with persisting post-concussive symptoms? A cluster analytical study. Brain Inj.
3. Chen, J. K., K. M. Johnston, S. Frey, M. Petrides, K. Worsley, and A. Ptito. Functional abnormalities in symptomatic concussed athletes: An fMRI study. Neuroimage
4. Chou, L. S., and L. F. Draganich. Stepping over an obstacle increases the motions and moments of the joints of the trailing limb in young adults. J. Biomech.
5. Chou, L. S., K. R. Kaufman, A. E. Walker, R. H. Brey, and J. R. Basford. Dynamic instability during obstructed gait following traumatic brain injury. Gait Posture
6. Fan, J., B. D. McCandliss, T. Sommer, A. Raz, and M. I. Posner. Testing the efficiency and independence of attentional networks. J. Cog. Neurosci.
7. Guskiewicz, K. M. Postural stability assessment following concussion: one piece of the puzzle. Clin. J. Sports. Med.
8. Parker, T. M., L. R. Osternig, H.-J. Lee, P. van Donkelaar, and L.-S. Chou. The effect of divided attention on gait stability following concussion. Clin. Biomech.
9. Posner, M. I. Orienting of attention. Q. J. Exp. Psychol.
10. Posner, M. I. Attention: the mechanisms of consciousness. Proc. Natl. Acad. Sci. USA
11. Thurman, D.J., C.M. Brache, and J.E. Sniezek. The epidemiology of sports-related traumatic brain injuries in the United States: recent developments. J. Head Trauma Rehabil.
12. van Donkelaar, P., J. Langan, E. Rodriguez, A. S. Drew, C. Halterman, L. R. Osternig, and L.-S. Chou. Attentional deficits in concussion. Brain Inj.
13. Halterman, C., J. Langan, A. S. Drew, E. Rodriguez, L. R. Osternig, L.-S. Chou, and P. van Donkelaar. Tracking the recovery
of visuospatial attentional deficits in mild traumatic brain injury. Brain