Inertial Sensor-Based Assessment of Central Sensory Integration for Balance After Mild Traumatic Brain Injury

Abstract

Introduction

Optimal balance control requires a complex integration of sensory information from the visual, vestibular, and proprioceptive systems. The goal of this study is to determine if the instrumented modified Clinical Test of Sensory Integration and Balance (mCTSIB) was impaired acutely after mild traumatic brain injury (mTBI) when postural sway under varying sensory conditions was measured with a wearable inertial sensor.

Materials and Methods

Postural sway was assessed in athletes who had sustained a mTBI within the past 2–3 d (n = 38) and control athletes (n = 81). Postural sway was quantified with a wearable inertial sensor (Opal; APDM, Inc.) during four varying sensory conditions of quiet stance: (1) eyes open (EO) firm surface, (2) eyes closed (EC) firm surface, (3) eyes open (EO) foam surface, and (4) eyes closed (EC) foam surface. Sensory reweighting deficits were computed by comparing the postural sway area in eyes closed versus eyes open conditions for firm and foam condition.

Results

Postural sway was higher for mTBI compared with the control group during three of the four conditions of instrumented mCTSIB (EO firm, EC firm, and EC foam; p < 0.05). Sensory reweighting deficits were evident for mTBI individuals compared with control group on foam surface (EC firm vs EO firm; p < 0.05) and not on firm surface (EC firm vs EO firm; p = 0.63).

Conclusions

The results from this study highlight the importance of detecting postural sway deficits during sensorimotor integration in mild TBI individuals.

Introduction

Every year, between 1.6 and 3.8 million people sustain a traumatic brain injury (TBI) in the United States and of these, 75% are considered mild TBI (mTBI).^1–4 Many symptoms after mTBI are subtle and may be overlooked by the examiner, so sensitive and objective measures of deficits are essential. Particularly, some studies have reported decreased balance performance when using objective measures such as force plates, accelerometers, or testing postural sway in more challenging conditions,^5–7 even when clinical measures appear normal (i.e., stopwatch times or error counts). Potential symptoms in the acute and subacute phase include confusion, loss of memory, loss of consciousness, headaches, dizziness, nausea, balance disturbances, impaired oculomotor control, gait unsteadiness, and poor coordination. Balance impairments are one of the frequently reported complaints.^8,9 It has been reported that anywhere from 23% to 81% of people report imbalance in the acute phase after mTBI and that up to 31% of people with mTBI have prolonged symptoms of imbalance.^8,10–12

Optimal balance control requires a complex integration of sensory information from the visual, vestibular, and proprioceptive systems.¹³ Proper sensory integration relies on the central nervous system to (1) depend on the optimal combination of sensory sources for balance and (2) reweight the sensory contributions as sensory conditions change. Abnormalities of peripheral vestibular function, such as benign paroxysmal positional vertigo or otolith dysfunction, are well documented in patients with mTBI.^14–17 However, even when peripheral vestibular, visual, and somatosensory systems are intact, difficulty performing static and dynamic balance tasks under challenging surface and/or visual conditions could occur due to deficits in weighting/reweighting use of vestibular, visual, and somatosensory information for balance when the sensory conditions change.^13,18 For example, standing or walking on compliant foam or irregular surfaces requires the nervous system to rely more on vestibular and visual inputs while down-weighting the normal dependence upon somatosensory inputs.¹³ When healthy subjects stand on foam with eyes closed, they away 50% more because they must rely on vestibular inputs, which have higher sensory noise than somatosensory inputs.¹⁹ Thus, a person who cannot reweight reliance on sensory inputs properly when the sensory conditions change will be less stable than someone who can reweight sensory information for balance control.

Such central sensorimotor integration, the ability to process and integrate information coming in from the senses and to transform the information into a motor output, is a critical aspect of balance control that may underlie dysfunction in static and dynamic balance. The clinical gold standard for assessment of sensorimotor integration for balance control is the Sensory Organization Test (SOT) performed on an Equitest (NeuroCom, Clackamas, Oregon) balance device, which uses a sway-referencing force plate and visual surround. A recent study found measurable sensorimotor integration deficits in service members after mTBI (58% of people in the acute phase and 47% in the subacute phase) using SOT.²⁰ Similarly, Lin et al. used the Biodex Stability System to investigate deficits in sensorimotor integration deficits in mTBI.²¹ However, such equipment assessing sensory organization is expensive, large, and non-portable, which limits its use by clinicians such as those practicing in small rural and outpatient settings.

The Clinical Test of Sensory Integration and Balance (CTSIB) was developed as a simple clinical substitute for the SOT.²² The original CTSIB attempted to simulate the six conditions performed on the SOT by using foam and a visual surround. However, the version currently used by clinicians is modified to include only firm and foam condition with eyes open (EO) and eyes closed (EC) – referred to as the modified CTSIB (mCTSIB). Although the mCTSIB provides a useful screening tool for sensory integration, it is based on subjective observation by the examiner to observe quality of sway and relies on a stopwatch measure for time. Like many subjective rating scales, the mCTSIB may be insensitive to subtle change or to mild deficits.^5,23,24

Recently, advances in wearable inertial sensors have made it possible to quantify existing clinical balance test results.^6,25,26 Specifically, postural sway measured with wearable inertial sensors have recently been validated with gold-standard force plate measures of postural sway, making it possible to instrument common clinical tests of balance, including the mCTSIB.^25–27 We recently reported that people with acute mTBI had increased postural sway during quiet stance, but it is unclear if varying sensory conditions such as visual and surface would provide further information.⁶

The purpose of this study is to explore the usefulness of the instrumented mCTSIB in the acute post-mTBI population. We hypothesized that by instrumenting mCTSIB with a wearable inertial sensor, we would see differences between those with acute mTBI and controls, particularly in those conditions that require upweighting vestibular information to maintain balance.

Methods

Study Design

This was a cross-sectional study of balance in college athletes who had sustained a mTBI within the past 2–3 d (case group) or who had not sustained a mTBI (control group) within the past 6 mo. This study was part of a larger study to monitor recovery of balance after mTBI (NCT01377454).

Participants

College-aged student athletes were recruited during their academic sports season from six universities: Portland State University, Lewis & Clark College, Linfield College, Pacific University, Concordia University, and George Fox University (Table I). In order to participate, interested athletes were required to be above 18 yr of age, have sustained a mTBI confirmed by an Oregon Health & Science University (OHSU) team Physician or Certified Athletic Trainer in past 2–3 d (case group), or have a history of no mTBI in past 6 mo (control group). Athletes were excluded if they (1) met DSM-IV criteria for substance abuse or self-reported drug use within the past 24 h, (2) had a history of a medical condition such as stroke or multiple sclerosis, which could impair balance or cognition, or (3) had an injury or surgery that would affect balance within the past 6 mo. The OHSU Institutional Review Board (IRB) approved this study. All subjects signed an informed consent forms approved by OHSU IRB. All work was conducted in accordance with the Declaration of Helsinki (1964).

Table I.

Participant Characteristics

Characteristic	mTBI	Control
	n = 38	n = 81
Age, mean ±SD	20.6 ± 1.3	21.0 ± 1.4
Gender, n	25 M/13 F	44 M/37 F
Height, mean ±SD, cm	178.0 ± 9.7	176.0 ± 9.8
Weight, mean ±SD, kg	86.2 ± 21.1	80.4 ± 20.7
BMI, mean ±SD	26.9 ± 4.7	25.7 ± 4.6
DHI	32.4 ± 15.4	0.67 ± 4.2
Sport, n
Football	22	20
Soccer	7	14
Basketball	3	9
Lacrosse	2	—
Volleyball	1	4
Softball	1	6
Baseball	1	9
Track and field	1	17
Tennis	—	1
Swimming	—	1

BMI, body mass index; DHI, Dizziness Handicap Inventory.

Demographics

Consenting and questionnaires were performed in a private office and the balance testing was completed indoors in an open area on a hard flat surface, each at the participating university. Subject demographics are presented in Table I, which includes description of subjects characteristics, including the sports they played and self-reported Dizziness Handicap Inventory (DHI). The DHI is a 25-item assessment tool to evaluate the self-perceived handicapping effects imposed by vestibular system disease.²⁸ The scores on DHI range from 0 to 100, higher scores representing worse self-reported symptoms of dizziness.

Balance Assessment

mCTSIB

The test involves four discrete conditions of quiet stance for 30 s with the feet close together and arms across the chest: (A) EO firm surface, (B) EC firm surface, (C) EO foam surface, and (D) EC foam surface (Fig. 1). Postural sway was quantified with the instrumented mCTSIB using a wearable inertial sensor (Opal; APDM, Inc.), which was placed at the L-5 (Lumbar 5th) vertebrae with an elastic belt. This inertial sensor contains a tri-axial accelerometer, a tri-axial gyroscope, and a tri-axial magnetometer. Accelerometers measure linear accelerations, gyroscopes measure angular velocity, and magnetometers measure heading with respect to the Earth’s magnetic field. We used the total sway area, automatically calculated with Mobility Lab V1 software (APDM, Inc.) as the primary metric.²⁶ Sway area was computed as the area spanned from the acceleration signals from the sensor, per unit of time (m²/s⁵).

Figure 1.

Photo of subject standing in the four sensory conditions of the instrumented mCTSIB. The traces display the postural sway during the 30 -s quiet stance as detected by the wearable inertial sensor placed on the lumbar fifth vertebra. The black trace displays the postural sway for a healthy control individual and the green sway displays the postural sway trace for a representative subject with mTBI. (Note: EO, eyes open; EC, eyes closed).

Statistics

To determine the differences in the postural sway area between the groups on the four conditions for the instrumented mCTSIB, we performed a repeated-measures analysis of variance between groups (mTBI and control) and within the different sway area conditions (EO firm, EC firm, EO foam, and EC foam). We performed post hoc tests to determine the sources of interaction between group and conditions, if any. We objectively defined sensory reweighting as the extent of change in postural sway in eyes closed versus eyes open on firm and foam surface. We then compared these two ratio measures using univariate analysis of variance. We also performed correlation analysis between the DHI scores and the postural sway area for the four conditions of instrumented mCTSIB.

Results

All but one subject (an mTBI individual) were able to complete 30 s of quiet standing in all four conditions of the mCTSIB. Postural sway area differed between the groups depending on the condition performed on instrumented mCTSIB (group effect, F_3,351 = 6.8, p < 0.001). The mTBI group had larger postural sway compared with the control group for three of the four conditions: EO firm, EC firm, EC foam, except EO foam (Fig. 2, Table II).

Figure 2.

Comparison of total sway area between mTBI and control groups for all conditions of the modified clinical test of sensory integration and balance: eyes open (EO) firm surface, eyes closed (EC) firm surface, EO foam surface, and EC foam surface. Asterisk (*) indicates a significant difference between groups (p < 0.05).

Table II.

Postural Sway Comparison Between mTBI and Control Groups for the Four Conditions of mCTSIB

	Group Effect (mTBI vs Control)
EO firm	F1,117 = 7.64; p = 0.007
EC firm	F1,117 = 10.79; p = 0.001
EO foam	F1,117 = 2.57; p = 0.11
EC foam	F1,117 = 10.14; p = 0.002

We approximated sensory reweighting by comparing the postural sway area in EC versus EO conditions. The extent of increase in postural sway with EC, compared with EO, on the firm surface was similar between control and mTBI individuals (p = 0.63). However, when standing on foam, the mTBI group increased their postural sway more than the control group under EC condition (F_1,117 = 6.05, p < 0.05).

We found a significant positive correlation between the EC foam condition and DHI index (EC foam: r = 0.4, p < 0.05). Correlation analysis between the other three conditions of mCTSIB and DHI index were not significant (EO firm: r = −0.02, p = 0.9; EC firm: r = −0.06, p = 0.7; EO foam: r = −0.14, p = 0.4).

Discussion

This study has three primary findings: (1) At approximately 2–3 d post injury, people with mTBI had increased postural sway under altered sensory conditions of the instrumented mCTSIB test. (2) Sensory reweighting (EC compared with EO) was more impaired under foam conditions compared with firm surface. (3) Self-reported dizziness severity was related to the postural sway deficits observed in the EC foam condition, suggesting the contribution of vestibular deficits in the postural sway impairments in mTBI.

Deficits in central integration of sensory systems for balance control became more evident when multiple sensory systems were challenged simultaneously. In line with the Haran paper,²⁰ we found that the condition that required the most use of vestibular information (EC foam) was significantly impaired in the mTBI group. Haran et al. concluded that postural instability in the acutely injured service members was primarily a result of vestibular and visual integration dysfunction and abnormalities determined on the SOT during the acute phase in service members were higher than previously reported (i.e., 50–58%). Our results support this notion as 45% of the mTBI group had abnormal balance (at least 1 standard deviation outside the normal range) based on postural sway area in the instrumented EC foam condition.

We found that the mTBI group was more dependent on vision compared with the control group, particularly when proprioceptive information was reduced. Specifically, the extent of increase in postural sway was similar for mTBI and control individuals when vision was removed while standing on the firm surface (when proprioceptive information was available to help compensate for the lack of vision). Interestingly, postural sway for EO foam was similar for controls and mTBI, but sway was higher for mTBI than controls for EC foam. Thus, both controls and mTBIs were relying predominantly on vision in the EO foam condition and visual orientation seems to be equally accurate and reliable in the mTBI and control groups. However, the eye closure caused a larger increase in sway in mTBIs than in controls in the EC foam condition as both groups shift toward high reliance on vestibular information in the EC foam condition. This suggests that accuracy/reliability of vestibular information is less in mTBIs than in controls.

Interestingly, self-reported dizziness severity was related to the postural sway deficits observed in the EC foam condition, suggesting that those people who had more difficulty reweighting to use vestibular information had more subjective complaints recorded on the DHI. Lin et al. reported significant differences for the DHI index between mTBI and the control group.²¹ However, they did not observe significant deficits in the EC foam condition. Surprisingly, the observed deficits in postural sway were observed only for the EO firm condition. It is conceivable that the non-significant differences in postural sway measures for Lin et al. were because of mildly impaired individuals. It is also possible that inertial sensor-based evaluation is more sensitive or reliable than the Biodex Stability System. This supposition requires sensitivity and reliability testing against the gold-standard equipment.

The larger postural sway area for mTBI than control individuals standing on foam with EC could be due to either deficits in peripheral vestibular system or due to deficits in central integration.^14,29 We are currently addressing these potential contributors in ongoing studies in our laboratory (NCT02748109). Regardless of underlying pathology, the current findings suggest the usefulness of performing instrumented mCTSIB after mTBI to determine acute deficits in postural control. Thus, the instrumented mCTSIB may prove to be useful in quantifying postural sway deficits in sensorimotor integration in mTBI.²⁷

One important limitation to our study was that we did not have information on the athlete’s balance performance before the mTBI. It is possible that the people who sustained an mTBI in our study had worse balance to begin with and therefore we cannot attribute all of their balance deficits to the immediate effects of an mTBI. It is difficult to differentiate between central and peripheral mechanisms of sensorimotor integration based on our data. We did not test the three sensory systems to investigate the deficits in the individual systems.

The results from this study highlight the potential usefulness of using a wearable inertial sensor to measure balance impairments after mTBI under varying sensory conditions. When returning to play or duty, common functional scenarios include momentary loss of sensory orientation information from the surface or the visual surround. In these moments, it is critical that people can adequately reweight sensory information to maintain stability or they could fall. People who cannot successfully and quickly reweight sensory information may be more susceptible to another injury.

Presentations

Presented as a poster at the 2016 Military Health System Research Symposium, Kissimmee, Florida (Abstract number: MHSRS-16-0783).

Funding

This project was supported by the Oregon Clinical and Translational Research Institute (KL2TR000152) from the National Center for Advancing Translational Sciences at the National Institutes of Health (NIH); (UL1TR000128) and Eunice Kennedy Shriver National Institute of Child Health & Human Development of the NIH under Award Number (R21HD080398).