Evidence for the Specificity of Intracortical Inhibitory Dysfunction in Asymptomatic Concussed Athletes

Introduction

T he Centers for Disease Control and Prevention (CDC, 1997) estimate that every 21 seconds someone sustains traumatic brain injury (TBI) in the United States, which represents about 1.2 million Americans per year. The CDC also estimate that of all traumatic brain injuries occurring in a year, 75% of them are mild TBI (mTBI). When mTBI takes place in a sports context, it is called a sports concussion, which is defined as a complex pathophysiological process induced by a near instant transfer of kinetic energy that affects the brain (McCrory et al., 2005). In the United States alone, between 50,000 and 300,000 contact sports athletes will sustain a concussion within the course of a single season. This high prevalence makes sports-related concussion a major public health concern (CDC, 1997).

Post-concussion symptoms usually consist of headache, dizziness, visual difficulties, memory disturbance, and concentration problems (Cantu, 1996). It is generally agreed that post-concussive symptoms typically disappear between 2 to 10 days after the incident (McCrory et al., 2005), a time window that closely coincides with the resolution of the neurometabolic cascade of concussion (Giza and Hovda, 2001). In addition to the transient effects on cognition, motor function alterations in the form of gait stability and balance control have been documented in the acute post-concussion phase (Cavanaugh et al., 2005; Guskiewicz et al., 2001; Parker et al., 2006). In the 48 hours following brain injury, athletes display changes in postural control that appear to be linked to deficits in sensory interactions between the visual, somatosensory, and vestibular systems (Guskiewicz et al., 2001). Furthermore, Catena and collaborators (2007a, 2007b) have shown that concussed individuals show increased medial/lateral motion in gait stability tasks.

Despite the apparent ephemeral nature of post-concussive symptomatology and the uncommon presence of tissue damage using routine imaging techniques such as CT scan (Kibby and Long, 1996), recent studies have suggested detrimental long-term effects of concussions, which might eventually evolve into devastating neurological conditions with aging. Neuropathological analysis of brain tissue has revealed chronic traumatic encephalopathy, a progressive tauopathy, in athletes who suffered multiple concussive injuries (McKee et al., 2009), and diffusion tensor imaging (DTI) has shown numerous white matter alterations after mild TBI (Maller et al., 2010). Studies have also shown that approximately 17% of retired professional boxers will eventually develop dementia pugilistica, a disorder characterized by motor and cognitive symptoms resembling those of Parkinson's disease (Rabadi and Jordan, 2001). The exact mechanism by which concussions contribute to the development of severe neurological conditions remains largely unknown; only a handful of studies have investigated the neurophysiological impact of brain injury in otherwise healthy individuals beyond the acute phase.

Despite the fact that concussed athletes typically stop complaining about motor symptoms 10 days after concussion, a recent study suggests long-lasting gait stability abnormalities whereby athletes showed altered postural stability in a dual task more than 28 days following concussion (Parker et al., 2006). Supporting these findings, neurophysiological studies of mTBI using transcranial magnetic stimulation (TMS) have recently shown primary motor cortex (M1) dysfunctions in the acute phase (Chistyakov et al., 2001) that seem to be long-lasting (De Beaumont et al., 2007). In a series of studies, Chistyakov and collaborators (1998, 1999, 2001) used TMS to investigate the presence of M1 neurophysiological alterations after traumatic brain injuries of different severity. These investigators reported altered motor cortex excitability in minor-to-severe TBI (Chistyakov et al., 1998, 1999, 2001), as well as prolonged cortical silent period (CSP) duration in mild-to-moderate TBI (Chistyakov et al., 2001), which is indicative of dysfunctional intracortical inhibitory systems. While these results documented immediate neurophysiological dysfunctions, the possible persistence of those alterations was unknown. In fact, it is only recently that researchers have investigated the possible long-term effects of sport concussions on M1 neurophysiology. Using TMS, De Beaumont and collaborators (2007a) have reported the presence of long-lasting intracortical inhibitory system abnormalities within the primary motor cortex of university football athletes who sustained multiple concussions. In line with the results of Chistyakov and colleagues (2001), they observed a lengthening of the CSP, which was linked to concussion severity but was independent of the amount of time elapsed since the last concussion. This suggests that sports concussions can produce neurophysiological alterations that persist well beyond the acute phase. Although the neurophysiological underpinnings of the CSP are still debated, the majority of pharmacological studies have attributed CSP lengthening to alterations of GABA _B inhibitory receptors. Indeed, CSP is prolonged by GABA re-uptake inhibitor tiagabine (Werhahn et al., 1999) and intrathecal GABA _B agonist baclofen (Siebner et al., 1998). Interestingly, M1 GABA alterations in concussed athletes appear to be limited to GABA_B receptors, as reflected by preserved short latency intracortical inhibition (SICI; De Beaumont et al., 2007a, 2009), a measure of GABA_A-related intracortical inhibition (Ilic et al., 2002; Ziemann et al., 1996).

In addition to specific alterations of GABA_B receptor activity, there is evidence for an alteration of cholinergic systems in severe traumatic brain injury resulting in diffuse axonal injury (Fujiki et al., 2006), where short latency afferent inhibition (SAI), a marker of cholinergic activity involved in sensorimotor integration, was reduced (Tokimura et al., 2000). Alterations in cholinergic activity and motor inhibitory circuits have been found in diverse pathologies such as Gilles de La Tourette syndrome (reduced SAI, Orth et al., 2005), asymptomatic Parkin mutation carriers (reduced SAI, Bäumer et al., 2007), Alzheimer's disease (reduced SAI, Di Lazarro et al., 2002, 2004, 2005), and Parkinson's disease (reduced long latency afferent inhibition [LAI], Sailer et al., 2003). Knowing that the aforementioned pathologies all share motor/memory dysfunctions similar to those associated with post-concussion syndrome (Jotwani and Harmon, 2010) and that SAI/LAI interacts with concussion-vulnerable inhibitory circuits of the primary motor cortex for sensorimotor integration (Chen, 2004), assessing the integrity of this system in concussed athletes is of particular clinical interest.

In contrast with the growing body of evidence for M1 inhibitory mechanism alterations (De Beaumont, 2007a, 2009), few concussion studies have investigated the presence of neurophysiological abnormalities in non-motor areas. Somatosensory evoked potential (SEP) testing represents a useful technique to assess the integrity of somatosensory cortex, as well as afferent conduction, and it is routinely used as a clinical tool in severe TBI to predict functional recovery (Chistyakov et al., 1999; Lew et al., 2003). Abnormalities in the SEP N20 component were found in comatose, diffuse axonal injury patients and in moderate brain injury victims (Chistyakov et al., 1999), and altered N20 latency seems to be related to clinical disability in severe TBI (Rappaport et al., 1990). Abnormal N60 latencies lasting up to three months were also found with concussion in a sample of consecutive patients presenting to the emergency unit (Zumsteg et al., 2006). Prolonged central sensory conduction time (CSCT) has also been reported in the acute coma phase after brain injury (Chistyakov et al., 1999) and abnormalities in central motor conduction time (CMCT) were found after brain injury, mostly in patients who sustained axonal damage (Chistyakov et al., 1999).

The principal objective of this study was to investigate the specificity of the previously reported long-lasting inhibitory dysfunction in primary motor cortex of concussed athletes. To this end, a comprehensive neurophysiological evaluation of sensorimotor function was performed in concussed athletes to better circumscribe areas of dysfunction in the aim of developing objective markers of concussion to facilitate diagnosis, gather prognostic insights, and facilitate return-to-play decisions. This is especially relevant in the case of asymptomatic athletes, in whom neuropsychological testing, neurological examination, and symptom checklist often fail to reveal any lingering dysfunction. A sample of symptom-free concussed athletes who sustained their last concussion on average two years prior to testing were assessed on the following measures: short latency afferent inhibition, long latency afferent inhibition, and somatosensory evoked potentials. Furthermore, the integrity of ascending and descending pathways was evaluated with sensorimotor conduction times. Finally, M1 GABA-mediated intracortical inhibition was evaluated with measures of short interval intracortical inhibition, long interval intracortical inhibition, and cortical silent period.

Methods

Results

Short and long latency afferent inhibition

SAI

A one-way ANOVA was first performed on raw MEP amplitude to verify that the CS modulated TS response. There was a main effect of ISI (F_{3, 23}=13.11; p=0.0001), which was caused by inhibition of the conditioned TS compared to TS alone at all intervals. Using the ratio of the conditioned stimulus over the test stimulus alone, a mixed ANOVA (Group X ISI) revealed no main effect of ISI (F_{2, 23}=2.88; p=0.08), no main effect of group (F_{2, 23}=0.02; p=0.88), and no interaction (F_{2, 23}=0.15; p=0.80; Fig. 1A).

OPEN IN VIEWER

FIG. 1.

(A) SAI ratios for every time condition. There is no significant difference between groups. Error bars represent standard error of the mean. (B) LAI ratios for every time condition. There is a significant difference between the two time conditions, where LAI 100 msec produces stronger inhibition than LAI 200 msec. Error bars represent standard error of the mean.

LAI

A one-way ANOVA was first performed on raw MEP amplitude to verify that the CS modulated TS response. There was a main effect of ISI (F_{2, 23}=16.78; p=0.0001), which was caused by inhibition of the conditioned TS compared to TS alone at all intervals. Using the ratio of the conditioned stimulus over the test stimulus, mixed ANOVA (Group X ISI) revealed a significant main effect of ISI (F_{2, 23}=9.80; p=0.005), no main effect of group (F_{2, 23}=0.02; p=0.88), and no interaction (F_{2, 23}=0.15; p=0.80). The main effect of ISI was explained by increased inhibition at the 100 msec interval (Fig. 1B).

Discussion

The major finding of the present study is the specificity of long-term intracortical inhibitory dysfunction in primary motor cortex of concussed athletes that is presumably GABA_B mediated. Compared to unconcussed athletes, CSP and LICI measures were abnormal, suggesting increased GABA_B inhibition. By contrast, GABA_A-mediated inhibition (SICI), sensorimotor integration assessed with cholinergic-dependent SAI, and basic somatosensory processing (SEPs) were similar in the concussed and unconcussed groups. Finally, both afferent and efferent sensorimotor conduction times were of similar duration in the two groups.

Cholinergic abnormalities have been reported in patients with severe TBI associated with diffuse axonal injury, where short latency afferent inhibition was significantly reduced (Fujiki et al., 2006). Administration of a single dose of an acetylcholinesterase inhibitor was sufficient to restore normal SAI in this population, in line with previous studies showing similar effects in Alzheimer's disease (e.g., Di Lazzaro et al., 2002, 2004). The present findings suggest that cholinergic circuits in sensorimotor areas are unaffected by sports concussions two years after the last concussive event. The absence of significant differences between the two groups suggests that sensory signals originating from the median nerve interact normally with primary motor cortex circuits in the concussed brain. SAI appears to be partly mediated by GABA_A receptors at the motor cortex level (Di Lazzaro et al., 2007). Data from the present and other (De Beaumont et al., 2007, 2009) studies have shown spared GABA_A-mediated SICI in athletes tested 1–2 years after their last concussion. Although SAI and SICI appear to involve different subtypes of GABA_A receptors (Di Lazzaro et al., 2007), it has been shown that they are reciprocally connected (Alle et al., 2009). Normal LAI was also found in the present population of formerly concussed athletes, and it is believed that LAI and SAI are mediated through different sensory-motor circuits. Although M1-S1 cortico-cortical interactions appear to underlie LAI, the exact nature of this inhibitory phenomenon is unclear (Pirio et al., 2009). Nevertheless, the present data point to the absence of long-term effects of sports concussions on the interaction between sensory input and corticospinal excitability. Whether this represents a state of recovery from short-term alterations is an open issue that warrants further investigation.

It is important to note that in contrast to previous studies from our group (De Beaumont et al., 2007b, 2009), MEPs were recorded from the APB rather than from the first dorsal interosseus muscle (FDI). It has been shown that excitatory and inhibitory patterns induced with TMS differ between muscle groups. For example, MEP area for muscles at rest is greater and onset latency longer for small hand muscles compared to forearm muscles, while CSP durations are longer for small hand muscles (Wu et al., 2002). Differential effects have also been reported between proximal and distal muscles, where increased intracortical inhibition and decreased intracortical facilitation are present in proximal compared to distal muscles (Abbruzzese et al., 1999). More relevant to the present study, it has been suggested that function may explain differences in inhibition between muscles, with intrinsic hand muscles playing an active role in fine motor acts requiring enhanced inhibitory control (Abbruzzese et al., 1999). As such, it is unlikely that the selection of one hand muscle over the other significantly affected the present data, but it is an open question whether similar group effects would be observed in proximal muscle groups.

The absence of sensorimotor integration dysfunction was matched by a lack of short/middle latency SEP abnormalities in concussed athletes. Concussions have been shown to increase middle-latency SEPs (N60) in the acute phase, an effect that tended to normalize to pre-injury levels three months after the concussive event (Zumsteg et al., 2006). The same study reported no significant latency differences in the early SEP components between concussed individuals and healthy controls (Zumsteg et al., 2006). The present data are in agreement with these findings, revealing no differences between concussed and control athletes on measures of SEP amplitude and latency. As such, if an increase in N60 latency in the acute and post-acute (3 months) phases was present in our group of concussed athletes, a post-injury period of more than one year is sufficient to restore basic somatosensory processing to normal levels. It should be noted, however, that participants in the study by Zumsteg and colleagues (2006) had lost consciousness following concussion, were older (35.4 years), spanned a much wider range of ages (22 to 62 years), and included patients with varied mTBI causes. This contrasts with the current sample, which was younger and much more homogeneous. As such, it is difficult to ascertain whether the lack of SEP abnormalities in the present sample of concussed athletes reflects recuperation from a dysfunctional state or merely the absence of SEP effects in the acute phase. This is an important issue in light of the fact that short-latency SEPs have been repeatedly shown to be good predictors of outcome in patients with severe TBI and are one of the best predictors of coma outcome (Carter and Butt, 2005). If SEP abnormalities are present acutely following sports concussions, it would be of great clinical interest to determine whether they can predict the severity of long-lasting impairments that have recently been discovered in the cognitive and motor domains.

In keeping with the lack of SEP abnormalities and sensorimotor integration dysfunction, the present data suggest that ascending and descending peripheral and central pathways are not affected by the presence of multiple sports concussions. In addition to normal SEP latencies, central somatosensory and motor conduction times were similar between groups. Numerous studies have used diffusion tensor imaging (DTI) to investigate the integrity of white matter fibers in mTBI (for review, see Maller et al., 2010). Of particular relevance to the present findings, reduced fractional anisotropy (FA) has been reported in the corticospinal tract and internal capsule of mTBI individuals (e.g., Bendlin et al., 2008). Indeed, reduced FA in the internal capsule of mTBI patients has been shown numerous times (Bendlin et al., 2008; Lipton et al., 2008; Miles et al., 2008) and has been observed up to 6 years post-injury (Inglese et al., 2005). Similarly, reduced white matter integrity of the corticospinal tract has been reported both in the short-term (2 months post-injury) (Bendlin et al., 2008) and long-term setting (107 months post-injury) (Kraus et al., 2007). Our physiological data show that if structural damage is present in white matter fibers of the corticospinal tract, it is not sufficient to modify ascending and descending conduction times. It should be noted that DTI studies of mTBI sometimes included patients with visible structural brain damage, prolonged unconsciousness, variable age groups, and presence of cognitive impairments, which contrasts with the homogeneous nature of the present sample. Indeed, every participant in the current study reported no overt symptom, was young and active, and had identical educational profiles. In this specific population, integrity of the corticospinal tract may be spared altogether.

The lack of significant somatosensory, sensorimotor, and conduction abnormalities found in the present study thus points to the specificity of long-term intracortical inhibitory dysfunction in concussed athletes. Previous findings of increased CSP duration in concussed athletes were replicated and additional evidence for the involvement of GABA_B receptor activity was provided in the form of increased LICI. In healthy subjects, administration of selective GABA_B agonist Baclofen increases LICI, possibly through facilitation of inhibitory post-synaptic potentials (McDonnell et al., 2006). As such, increased CSP and LICI measures suggest that both magnitude and duration of GABA_B-mediated intracortical inhibition (McDonnell et al., 2006) is affected in sports concussion. It is important, however, to consider that TMS is an indirect measure of GABAergic activity. Evidence for an association between activity of specific receptor subtypes and TMS measures of intracortical inhibition comes primarily from pharmacological studies where GABA_A or GABA_B agonists or antagonists modulate M1 responses to TMS (Ziemann, 2004). Direct confirmation of the involvement of specific GABA receptors in the pathophysiology of sports concussions are needed to establish this fact with certainty.

A possible mechanism explaining presumed increased GABA transmission in concussion may be related to protective effects against glutamate excitotoxicity. Excessive glutamate stimulation is a core feature of brain response to TBI (Baker et al., 1993; Faden et al., 1989) leading to “glutamatergic excitotoxicity” and neurodegeneration (Rothman and Olney, 1986). Increased levels of glutamate are associated with NMDA receptors, and beneficial effects of NMDA receptor antagonists in TBI patients have been reported (e.g., Yurkewicz et al., 2005). In addition to modulation of glutamate levels, there is evidence for the involvement of GABA in the response to TBI, which could be secondary to the increase in glutamate and represent an attempt at minimizing glutamatergic excitotoxicity. Although the effects of TBI on GABA transmission in humans have been poorly studied, it has been reported that GABA concentration in ventricular CSF is greatly elevated in patients with severe brain injury (Palmer et al., 1994). There is significant data suggesting increased GABA levels following TBI in rat models of injury. For example, impact injury of rats has been shown to induce long-lasting working memory (WM) deficits that are associated with increased GABA levels for as long as 1 month post-TBI (Kobori and Dash, 2006). Reduction of GABA levels in those rats by administration of GABA antagonists restored memory function, suggesting that TBI is associated with “excess GABA-mediated inhibition” (Kobori and Dash, 2006). Magnetic resonance spectroscopy has shown a similar pattern of response following TBI in rats, where GABA concentration is elevated after injury (Pascual et al., 2007). The idea that long-term GABA increases in concusssed athletes is a response to glutamatergic excitotoxicity is obviously highly speculative. Indeed, it has recently been shown with MR spectroscopy that glutamate levels are decreased in the motor cortex of concussed athletes 3–4 days post-injury (Henry et al., 2010). However, since animal models have shown an immediate release of glutamate following mTBI (Katayama et al., 1990), complex interactions between excitatory and inhibitory neurotransmitters may significantly modulate the initial response in the days following concussive injury.

It is important to take into account the fact that the neurophysiological profile that is reported here comes from an homogeneously small sample made up of young, otherwise healthy high-level athletes. This has some important advantages, such as the presence of a comparable control group and reduction in the prevalence of co-morbid conditions. It should therefore be emphasized that the present results may not easily generalize to a broader population of individuals with TBI, where etiology, age, general health, and co-morbidity are important variables. For example, it remains to be seen whether individuals with more severe TBI also display abnormal M1 intracortical inhibition. With respect to concussed athletes, however, we can safely assume that our results are generalizeable despite the small sample size since similar dysfunction of inhibition in M1 have been found in older former athletes (DeBeaumont et al., 2009) and in another sample of young, active athletes (DeBeaumont et al., 2007).

In conclusion, the current data show that intracortical GABA dysfunction in the motor cortex of concussed athletes shows a surprising degree of specificity. The reported alteration appears to be long lasting and stable, as abnormal CSP durations have also been found more than 30 years after the last injury (De Beaumont et al., 2009). The specificity and duration of M1 dysfunctions make it plausible that TMS measures of cortical excitability may in the long run provide diagnostic and prognostic cues in TBI. In light of the fact that depression and post-traumatic stress disorder (PTSD) are often present in TBI patients (Kim et al., 2007; Rogers and Read, 2007), objective measures of M1 neurophysiology may find useful clinical utility. This is particularly relevant since both depression and PTSD have specific cortical excitability dysfunction profiles that are different from that of TBI. In PTSD, hemispheric-specific reductions in SICI and SAI (Rossi et al., 2009) have been reported, whereas depressive patients show a consistent pattern of right-left hemisphere motor threshold differences (Maeda et al., 2000) and reduced CSP durations (Bajbouj et al., 2006). Further studies will be necessary to determine the value of TMS measures in the differential diagnosis of these pathologies. Additionally, at present, return-to-play decisions following sports concussion are highly dependent on the athlete being physically and cognitively asymptomatic, as assessed by medical and neuropsychological evaluations (Reddy and Collins, 2009). The present data show that when these symptoms subside, highly specific dysfunctions in primary motor cortex may linger for years in some athletes. The development of objective, neurophysiological measures of brain dysfunction following concussion may therefore provide valuable return-to-play information in the future. However, whether the presence of intracortical inhibitory dysfunction can be a predictor of future concussive events, for example, is an open issue that will need to be addressed directly in future studies.