Higher Dependency on Visual Inputs During Posture Control: Another Long-Term Effect of mTBIs

This study investigated the hypothesis that individuals living with long-term effects of mild traumatic brain injury (mTBI) develop increased dependency on their visual inputs to control their vertical posture. To test this hypothesis we quantied visuo-postural dependency indices (VPDIs) calculated independently for multiple postural behavioral markers extracted from the body’s center of pressure coordinates signals recorded during the execution of a quiet bipedal stance. One hundred and twenty-nine volunteers participated in this study. An mTBI (n = 50) and a neurotypical Control group were formed (n = 79). VPDIs were calculated as the normalized pair-wise subtraction of recordings obtained under Vision and No-Vision experimental conditions. Consistent with our hypothesis the results of this study show that balance behavior of mTBI participants deteriorate more abruptly in the absence of visual inputs when compared to neurotypical controls. These impairments may increase the likelihood of recurrent traumas when fast reactions are needed in daily activities, sports practice, or military operations. Additionally, the methodology used in this study showed to be potentially useful to aid future investigations of neural circuitry impaired by mTBI and provide indices of recovery in future clinical trials testing mTBI-related clinical interventions. loss of self-esteem, anxiety, anger, frequent depression, feeling of despair, or emotional detachment; (d) no previous history of peripheral neuropathy (including loss/decreased sensory or motor function in the upper extremity), acute upper extremity injury, recurring and/or unexplained headaches, cardiac pacemaker, pulmonary disease, or metallic implants in head, spine, or upper extremity; (e) no history of any type of neurosurgery; and (f) no abnormalities of cranial nerve functions. (p = 0.008), and VPDI COPml_mean_jerk_ml (p < 0.001). These results were conrmed by the Mann-Whitney U tests ran with factor Group (Control vs mTBI) and each VPDI.


Introduction
Impairment to body vertical control is a common nding in patients suffering from the long-term effects of mild traumatic brain injuries (mTBIs) [1][2][3] . Its effects are captured several months after an injury (> 6 months) and they indicate deterioration to one's ability to control its vertical posture. Signi cant increases in indices of body sway amplitudes accompanied by increases on body sway velocities corroborates with this view 4 . Despite the fact these postural behavior abnormalities are subtle and may go unnoticed for years, they are a key factor for the development of re-occurring injuries during young adult age as well as higher risks of falling on later stages of life. Multiple efforts are currently underway to clarify the neural mechanisms behind these behavioral modi cations but current evidence suggests they stem from suboptimal multi-muscle control by the central nervous system (CNS) [2][3][4] .
This rationale nds support on the notion that multi-muscle control is a complex neural process depending on the optimal functioning of its conduction system [5][6][7] . This view is supported by the large number of neural substrates involved in the process of integrating sensory inputs from the visual, somatosensory, and vestibular organs into coherent motor outputs sent to the axial skeletal muscles. As a result, sensory neural potentials must travel relative long distances to be integrated. To cope with such complexity and allow for a fast integration of sensory information into motor commands the CNS counts with a fast and reliable conduction system. In fact, severe reduction in velocity conduction will make impossible for someone to respond to sudden environmental changes (e.g., changing direction to avoid collision with a tree while skiing) increasing the risks of severe injuries.
Conduction functions have shown to be compromised by mTBIs. For example, a recent study performed on patients suffering from concussive syndrome showed signi cant increases in the response latencies invoked by both simple visual and auditory stimuli forty-eight months post-injury 8 . These ndings are considered of importance since they suggest a wide-spread reduction in conduction of information across CNS substrates. This point of view has gained support from multiple observations showing signi cant reduction in the volume of important tracts associated with intra-and inter-hemispherical neural transmission [9][10][11][12] . For example, Hulkover et al (2013) 11 compiled results from a large cohort of studies that utilized diffusion tension imaging and suggested the existence of white-matter abnormalities in important areas for the integration of neural information within and between distinct brain areas including the superior longitudinal fascicule (SLF). Interestingly, the SLF is an important intra-hemisphere association pathway establishing connections between postrolandic regions and the frontal lobe 13-15 by conveying important information about the perception of the visual space (SFL subdivision II, SFL-II).
Therefore, the construction of a reliable neural representation of body con guration needed for optimal vertical posture control is dependent on the functional integrity of these tracts 16-20 .
An important characteristic of the abnormal postural behavior observed in mTBI patients regards to their subtle progression during the initial years after an mTBI episode. During this period, despite the absence of reported symptoms a series of underlying modi cations to the mechanisms of posture control may be under way to cope with the delays found in the neural processing of sensory-motor integration. Currently, the extent of the sensory functional re-weighting among major sensory systems after an mTBI is unclear however, we hypothesized that under such constraints a reorganization on the level of reliance on sensory modalities will occur. This hypothesis is based on previous reports supporting the notion the CNS is able to re-weight the role of each sensory modality on generating motor commands [21][22][23][24][25] . Direct evidence for a re-weighting process comes from previous studies using manipulations of characteristics of sensory stimuli and their resulting effect to postural behavior indices (e.g. body sway amplitude and frequency) 26-28 . Collectively, these studies suggest that a healthy central nervous system can reduce its dependency on a less reliable sensory modality while increasing its dependency on the remaining channels. However, mTBI patients suffer from delays in information transmission of multiple sensory modalities therefore increasing the complexity of such adaptations. In this study we tested the hypothesis that individuals living with long-term mTBI effects develop a larger dependency on their visual inputs to control their vertical posture. This expectancy is based on the rationale that under the scenario of multiple delays in information processing, the CNS will increase its reliance on the visual system due to its capacity to produce a more complete description of its environment.

Participants.
One hundred and twenty-nine volunteers participated in this study (x̄ = 25.3 years of age; s = 7.2 years). All provided their informed and written consent based on procedures approved by the local Institutional Review Boards (University of Montana and Western Michigan University) and conforming to The Declaration of Helsinki. Informed consent was also provided to publish any images reported here.
An mTBI group was formed by fty individuals (twenty-one females; twenty-nine males; x̄ = 28.8 years of age; s = 9.4 years) with history of single or multiple mild traumatic brain events. This cohort included occurrences from sports related activities, motor vehicle accidents, and military operations. The average number of previous concussive events was 2.15 (s = 2 events) and the average time elapsed from the last traumatic episode was 89.9 months (s = 90 months). Inclusion criteria for this group followed the standard classi cation proposed by the American Psychiatric Association as published in the Diagnostic and Statistical Manual of Mental Disorder 29 and included (a) mTBI caused by either direct impacts or blast-induced closed head trauma; and (b) presence of one or more of the following manifestations after the mTBI episode: confusion or disorientation, loss of consciousness for thirty minutes or less, posttraumatic amnesia for less than twenty-four hours, Glasgow Coma Scale score of 13-15 after thirty minutes post-injury or later on presentation for health care, and/or other transient neurological abnormalities such as focal signs, seizure, and intracranial lesion not requiring surgery. Exclusion criteria included the following: (a) manifestations of mTBI caused by drugs, alcohol, medications, and other injuries or treatment for other injuries (e.g. systemic injuries, facial injuries, or intubation); (b) other potentially confounding issues (e.g. psychological trauma, language barrier, or coexisting medical conditions); (c) history of penetrating cranial injury; (d) history of any type of neurosurgery; and (e) abnormalities of cranial nerve functions.
A control group (Control) was formed by seventy-nine healthy individuals ( fty-seven females; twenty-two males; x̄ = 23.4 years of age; s = 4.5 years). The exclusion criteria for the control group included the following: (a) no previous history of head trauma, traumatic brain injury (including cerebral vascular accident), epilepsy, or seizure disorders; (b) no history of substance abuse (drugs, alcohol, or controlled medication); (c) no symptoms of psychological trauma such as panic attacks, frequent nightmares, insomnia, loss of self-esteem, anxiety, anger, frequent depression, feeling of despair, or emotional detachment; (d) no previous history of peripheral neuropathy (including loss/decreased sensory or motor function in the upper extremity), acute upper extremity injury, recurring and/or unexplained headaches, cardiac pacemaker, pulmonary disease, or metallic implants in head, spine, or upper extremity; (e) no history of any type of neurosurgery; and (f) no abnormalities of cranial nerve functions.

Materials and Procedures.
A force platform (AMTI BP400600, AMTI Inc.) was used to acquire the vertical and horizontal components of the ground reaction force (GRF) as well as moments of force around the frontal and sagittal axes. These signals were used for computations of the body's center of pressure coordinates in anterior-posterior and medial-lateral directions (COPap and COPml, respectively). Consistent with previous studies COPap and COPml were de ned by: where h is the height of the base of support above the force plate, F x is anterior-posterior component of the GRF, F y is medial-lateral component of the GRF, F z is vertical component of the GRF, M x is the moment of force around the sagittal axis and M y is the moment of force around the frontal axis. All signals were sampled at 50 Hz with a 12-bit resolution. Prior to data recording, each participant went through a general neurological screening then provided self-reports utilizing Neurobehavioral Symptom Inventory (NSI) to record their prevailing current symptoms. Participants who met the established inclusion and exclusion criteria were tested in an area free of noise and distractions.
Protocol and Variables of Interest.
All participants were asked to perform the simple task of standing still for one-hundred and twenty seconds on a force platform with their either eyes opened or closed (Vision and No-Vision conditions, respectively). Feet were placed in parallel and fteen centimeters apart and consistency of foot placement across participants was achieved by markings drawn on the surface of the force plate (Fig.  1a). Since participants did not move their feet during the entire experimental session this position was consistent along the entire experiment. Participants were barefoot during the entire experiment and instructed to cross their upper limbs against the chest. During execution of the Vision condition participants focused their sight to a physical static point placed at eyes level and at approximately one meter. Once this initial position was taken, participants were asked to remain steady, as vertical as possible, and to distribute their body weight evenly between the two feet. To avoid recording of any transient effects during the initial moments of a trial, data recording started after ve seconds from the moment the subject took position. During execution of the No-Vision condition participants these same procedures were followed however participants kept their eyes fully closed during the entire time of data recording. The average duration of the entire experimental session was approximately forty minutes and none of the participants reported fatigue or dizziness as an issue.
Signals from the force plate (Fig. 1b) were analyzed off-line with customized routines (BalanceLab vs 3.0, Synergy Applied Medical and Research Inc, USA). COPap and COPml coordinate signals were down sampled to 10 Hz and detrended by the mean of each time series. The following postural indices were computed for both Vision and No-Vision conditions: area of COP displacement (COP_area), total length of COP displacement (COP_total_length), maximum amplitude of COP displacement in each direction (COPap_range and COPml_range), root mean square estimations of COP signals (COPap_rms and COPml_rms), mean velocities of COP displacement (COP_total_mean_vel, COPap_mean_vel, and COPml_mean_vel), mean jerkiness of COP displacement (COP_total_mean_jerk, COPap_mean_jerk, and COPml_mean_jerk), frequency containing 80% of signal spectral power (COPap_F80 and COPml_F80), amount of regularity and predictability of COP displacement in time quanti ed by sample entropy of COP signals (COP_sent_ap and COP_sent_ml), and degree of asynchrony or dissimilarity between COPap and COPml signals in time quanti ed by cross sample entropy (COP_crosssent).
COP_area was de ned similarly to procedures employing the sector formula of Leibniz previously described and used in the literature 30,31 . COP_total_length was computed as the total length of the COP displacement during the whole stance trial. The maximum amplitudes of the COP displacement in each direction (COPap_range and COPml_range) were computed by the difference between their maximum and minimum coordinates recorded. COP_total_mean_vel, COPap_mean_vel, and COPml_mean_vel were computed as the length of the COP trajectory divided by the duration of the trial. COP_total_mean_jerk, COPap_mean_jerk, and COPml_mean_jerk represented the rate of change of COP acceleration. They were computed as the third derivative of the COP position with respect to time. COPap_sent, COPml_sent, and COP_crosssent were computed through estimations of the correlation, persistence, and regularity of the COP signal in time. Smaller sample entropy estimates indicate many repetitive patterns of COP uctuation in time, whereas larger estimates indicate a more irregular, random, and unpredictable pattern. For cross-sample entropy, higher estimates indicate larger levels of asynchrony of postural sway between the two directions, whereas lower estimates indicate more co-dependence 1,32,33 .

Results
Recordings of current symptoms con rmed the presence of long-term effects on our mTBI cohort. Table 1 shows averages and standard deviations of NSI scores obtained across participants. They represent symptoms felt by the participants along the two weeks preceding the experimental session. Multiple symptoms affected our mTBI cohort including di culties to fall or stay asleep, headaches, irritability, poor concentration, di culty remembering information and making decisions, slowed thinking, poor coordination, changes in appetite, and hearing and vision di culties. These ndings were con rmed by a series of two-tailed student's T tests. A signi cant increase in the NSI's total score recorded for the mTBI participants was found (p < 0.0001).  Figure 2 shows stabilograms recorded from representative participants from the Control (panels A and C) and mTBI (panels B and D) groups under the Vision and No-Vision conditions. Note that for both participants the transition between these two conditions had a signi cant impact to their sway magnitude. However, the impact of visual disruption to mTBI participants was disproportionate larger and impacted several other indices of postural sway. Such observations were held true for both cohorts and are presented on Tables 2 and 3.   Tables 2 and 3 show the medians and quartiles (25th and 75th ) across participants for all seventeen postural indices of interest and VPDIs recorded under Vision and No-Vision conditions. Signi cant higher modulations for the mTBI group were found for VPDI COP_area (p = 0.013), VPDI COPap_range (p = 0.005), VPDI COPml_rms (p = 0.012), VPDI COP_total_mean_vel (p < 0.001), VPDI COPap_mean_vel (p = 0.013), VPDI COPml_mean_vel (p = 0.002), VPDI COP_total_mean_jerk (p = 0.008), and VPDI COPml_mean_jerk_ml (p < 0.001). These results were con rmed by the Mann-Whitney U tests ran with factor Group (Control vs mTBI) and each VPDI.

Discussion
In this study we tested the hypothesis that individuals living with long-term effects of mild traumatic brain injury (mTBI) develop a larger dependency on their visual inputs to control their vertical posture. Consistent with this hypothesis our results show balance behavior of mTBI participants to deteriorate more abruptly in absence of visual inputs when compared to neurotypical controls. This deterioration was quanti ed by visuo-postural dependency indices (VPDIs) calculated independently for seventeen postural behavioral markers extracted from the body's center of pressure coordinates signals recorded during the execution of quiet bipedal stance.
As expected, our control group showed signi cant changes in body sway dynamics when visual inputs were disrupted. We observed signi cant increases of body sway amplitude and body sway velocity in both directions as well as increases to the total area and total length of sway. These ndings are well aligned with previous reports demonstrating similar effects on both young and older healthy adults 21,25,34−37 . These results support the rationale that vison is a sensory modality able to produce a reliable source of information for the control of the vertical posture. Under this rationale, visual inputs are constantly fed into the CNS resulting in the construction of an internal representation of one's relation to its surroundings (exteroception). This information is then utilized for the elaboration of motor outputs sent to postural muscles responsible for stabilization of major joints along the axial skeleton (e.g., ankle, knee, hip, intervertebral joints). Therefore, once vision is disrupted the quality of exteroception information is reduced and production of motor outputs is performed under reduced levels of certainty regarding the current state of body dynamics. As a result, the vertical position becomes less stable resulting in larger amplitudes and faster body sway. This rationale is particularly supported by evidence showing that shortterm disruptions of visual inputs are indeed related to modi cations observed to the usual mechanisms of muscle activation utilized to maintain the vertical posture 38 as well as the presence of phase-locked body sway to motion of the visual eld [39][40][41][42] .
In addition to changes in the indices representing the temporal domain of analysis, we also observed signi cant increases to the distribution of the power spectrum density content of the COPap signal as well as increases in its temporal regularity. This nding is well aligned with several previous Observations presented in Table 2 show that both groups responded with signi cant increases in most markers included in this study. However, analyzes of VPDIs indicated our mTBI cohort showed a disproportionate larger effect of visual disruption to body sway indices recorded in medial-lateral direction (Table 3). Speci cally, we found signi cant increases in VPDIs indicating larger medial-lateral amplitude of sway accompanied by larger sway velocities, larger variations along the time (COPml_rms), and a larger presence of corrective actions during sway (COPml_mean_jeark). Even though we expected a signi cant deterioration of postural stability on our mTBI cohort, we did not expect a directional selectivity effect.
Disproportionate deterioration of body postural control in medial-lateral direction has been described in postural studies conducted in other populations 47,48 . However, its driving mechanisms remain unclear.
We speculate this effect may be driven by a combination of anatomical aspects of the axial skeleton and the segmental distribution of its postural muscles. Speci cally, the presence of two lower limbs positioned side-by-side and the reduced range-of motion of the knee joints in the frontal plane allow for relatively better medial-lateral stability. The parallel positioning of the two lower limbs increases the size of our base of support in the medial-lateral direction while allowing an effective transmission of reactive forces to partially counter-act the tendency of the body to lean on its side. As a result of this mechanical stability the number of muscles dedicated to control medial-lateral balance can be reduced therefore, making the muscular system dedicated to this type of control less redundant. Under our interpretation, even though this less redundant system functions well in health, it may reduce the ability of the central nervous system to adapt its neuromechanisms of control when the controller itself is compromised by pathologies. Instability will then follow.
Based on this rationale, a more redundant system, such as the one formed by the muscles controlling anterior-posterior sway will be able to respond better to impairments while delaying the progression of instability. This idea of a redundant system of control is well-known and has been demonstrated in multiple levels of analysis and in several types of motor actions [49][50][51] . According to its principles, the CNS functions by exploring the motor redundancy offered by groups of muscles by co-vary their level of activity in a way that the nal mechanical constraints of the motor task are respected. Despite its speculative status, the results presented here support this rationale. Note, that despite the fact the majority of VPDIs reaching signi cant threshold for inference, we also uncovered sizeable increases in VPDIs computed for anterior-posterior sway.
It is also important to note that mTBI affects a younger population who's professional and leisure activities involve faster reactions to environmental changes and failures to comply with their temporal constraints will increase the risk of segmental injuries and mTBI recurrences. This rationale is supported in part by the long-term effects mTBIs have to the CNS' white matter. Modern imaging studies have revealed signi cant reductions in the volume of important neural tracts. For example, Hulkover et al (2013) 11 compiled results from a large cohort of one hundred studies that utilized diffusion tension imaging indices (e.g., fractional anisotropy) and suggested the existence of white-matter volume reduction in important areas for the integration of neural information within and between distinct brain areas including the superior longitudinal fascicule (SLF). SLF is responsible for intra-hemispherical exchange of neural information exchanged among the occipital, parietal, and frontal cortices, including their important areas of association located in the parietal and frontal cortices. Such reductions in volume occur due demyelination of these tracts resulting in extended delays in information exchange and integration of multiple sensory modalities intended to the creation of an internal representation of the body's current status. Delays of this nature have been recorded in mTBIs for visual and auditory stimuli 8 .
More recent studies have documented subtle but signi cant reductions in infratentorial areas of the CNS (ex. brainstem and cerebellum) in more severe cases of traumatic brain injury [52][53][54][55] . These observations suggest that a potential wide-spread effect of TBI to the CNS' nuclear and conduction system is likely. We interpret our results as corroborating with this rationale.