The role of sensory augmentation for people with vestibular deficits: Real-time balance aid and/or rehabilitation device?

1 Vestibular disorders

Vestibular and balance disorders are burdensome to society, limiting participation in work [72], exercise [64], driving [23] and social contexts, as well as resulting in increased health care costs [2, 73]. Vestibular disorders are primarily characterized by vertigo, dizziness, visual complaints, or unsteadiness [16], which often lead to secondary physical and psychological impairments [69]. A recent initiative by the Classification Committee of the Barany Society (CCBS) recommended that the International Classification of Vestibular Disorders (ICVD) be organized into four layers: (1) symptoms and signs, (2) syndromes, (3) disorders and diseases, and (4) mechanisms, in order to accommodate the breadth of clinical and research applications [16]. A few examples of specific types of vestibular disorders include: vestibular neuritis, bilateral vestibulopathy – distinguished by lack of vertigo but presence of notable imbalance, and central vestibular dysfunction resulting in gait ataxia [30].

Data collected from the National Health Interview Survey, USA found that 19.6% of 37 million older adults surveyed reported an issue with balance or dizziness during the past year [64]. The National Health and Nutrition Examination Survey (NHANEA) concluded that 35% of US adults age 40 years and older demonstrated evidence of balance dysfunction and patients who were clinically symptomatic had a 12-fold increase in the odds of falling [1]. Approximately 8 million American adults report chronic problems (lasting three months or longer) with balance, and the cost of medical care for patients with balance disorders exceeds $1 billion per year [74].

People with vestibular disorders often experience balance and gait deficits [1, 100]. Persons with bilateral vestibular loss report much higher fall rates than control subjects and those with unilateral peripheral loss [39]. Walking with head movements appears to be particularly difficult for persons with vestibular deficits. Persons with mixed vestibular loss appear to report more falls while walking with head movements in the pitch plane versus yaw plane [99], yet others have reported no difference in gait speed during walking with pitch and yaw head movements [81]. Gait speed has been reported to be slower in persons with either peripheral or central vestibular loss compared to control subjects [67, 81]. Walking backwards was slower than walking while performing a cognitive task consisting of counting backwards by 7 s from 100 in persons with vestibular disorders [81].

Falling [1] and fear of falling [89] can both result from having a balance/vestibular disorder. Wrist and hip fractures have been related to vestibular hypofunction in a subset of persons presenting to the emergency room [28, 49], suggesting that vestibular disorders and fall events may be related. Staab has suggested that persons with fear of falling may become more unstable during gait because of diverting their gaze as a result of their anxiety [89]. Heights and surroundings that are visually provocative such as grocery stores and train stations can cause an increase in dizziness and provoke imbalance in a certain subset of persons with vestibular disorders [17, 80]. Persons with vestibular disorders become more rigid compared to control subjects when exposed to heights, suggesting that they may have different attentional demands and be threatened by heights [109]. Others have suggested that persons who are fearful of heights when viewing a visual cliff move their head less spontaneously and suppress eye movements [50].

Vestibular disorders have been shown to impact mood by increasing rates of anxiety and depression and reducing overall health related quality of life (HRQOL) [107]. People with vestibular disorders tend to restrict movement and exercise thereby limiting vestibular compensation and prolonging overall recovery [108]. Symptoms related to vestibular disorders are often debilitating, can lead to decreased independence in activities of daily living (ADL) [70], and increase the likelihood of falls [1].

With the increased incidence of falls and fear of falling in a subset of persons with vestibular disorders [39], physical assistance through the use of a balance aid may provide adequate support to enable people to continue to participate in their daily life roles. Balance aids include assistive devices such as canes, walking poles and walkers which help with postural stability [71] but also provide a physical sign to the community that the person has a disability. For the persons with bilateral loss, an assistive device is less effective in preventing falls in low-lit conditions due to decreased visual inputs and oscillopsia.

A substitute for vestibular information in persons with bilateral vestibular loss is needed [61] and sensory augmentation has potential to fulfill this need.

2 Sensory augmentation/sensory substitution/biofeedback/feedback

The vestibular system of the inner ear provides the central nervous system (CNS) with information about angular velocity in three dimensions. Similarly, it provides the CNS with information about linear acceleration and gravity. The CNS, in turn, processes this information to estimate the orientation of a person in space and to estimate how far he/she is tilted away from gravity vertical. The latter estimate helps people to remain upright while standing and walking. Falls may occur if the person cannot sense how far he/she is tilted away from the vertical, or if he/she lacks muscle control that contributes to balance reactions necessary to maintain stable posture and locomotion. Thus, a sufficiently sensitive and accurate estimate of body tilt is necessary for walking and standing. When the vestibular system is compromised for example by injury, disease, or ototoxic drugs the CNS may not be able to give reliable tilt estimates. An alternative is to use an electronic tilt sensor attached to the body and connected to a microprocessor to provide a real-time body tilt estimate, then to display the estimate using another sensory channel.

Sensory substitution can be defined as the delivery of sensory information from one sense through an alternate sensory channel using a translation device. The concept is attributed to Paul Bach-y-Rita who used it first in the 1960’s [7]. A classic example is the display of visual information (e.g., a letter of the alphabet) with an array of vibrotactile elements placed on the skin of a blind subject [7]. Thus, somato-sensation serves as a substitute for vision. In the example above, the translation device could be a video camera trained on a letter of the alphabet. The video output is then sampled at a spatial frequency that corresponds to the number of elements in the vibrotactile elements in the array. Samples that are dark would activate the corresponding element in the array, while samples that are light would not activate their corresponding elements. The term sensory augmentation is somewhat similar to sensory substitution except it implies that the alternate pathway does not necessarily completely replace the missing sense, since that sense may not be completely absent.

A sensory augmentation device typically uses inertial measurement units or pressure sensors and an appropriate feedback display to provide cues of body motion that supplement an individual’s intact sensory inputs. Visual feedback displays are the most common means of conveying knowledge of performance; however, there are practical considerations that must be taken into account for certain populations, such as persons with vestibular deficits and older adults who rely heavily on the visual system. Visual motion can produce postural responses that may supersede proprioceptive and vestibular inputs, thereby disrupting postural stability especially in persons with visually induced dizziness [19]. Persons with phobic postural vertigo (a form of visually induced dizziness) when walking with eyes closed had worsening of their gait compared to controls, suggesting that persons with phobic postural vertigo may have greater reliance on vision while ambulating [82].

Unimodal and multimodal feedback displays including vibrotactile [86, 93], electric currents applied to the tongue [5, 6], auditory [26, 38], and combinations of visual, vibrotactile, and auditory cues [24, 41] offer varying degrees of non-invasive self-motion cues. Depending on the application (e.g., continuous use as a real-time balance aid, periodic rehabilitative use), some display modalities present practical difficulties. While visual feedback devices offer some of the best outcomes during certain tasks, they may not suitable for activities that involve head movements or closed eyes. Auditory displays can be problematic for the many people with vestibular loss who also have hearing problems. Technologies using head-based motion sensors may result in the “locking” of head motion to trunk motion rather than allowing the two to be decoupled as occurs during normal postural control; a primary goal of balance therapy, however, is to increase the ability of a patient to make asymptomatic head movements and increase their degrees of freedom. Torso-based vibrotactile displays, which offer less spatial resolution than visual, auditory, and electric lingual displays and require longer reaction times than head-mounted displays [4, 11] don’t directly compete with tasks that involve speaking, hearing, or seeing.

3 Stance

Results from the literature have shown that vibrotactile feedback of body motion reduces the number of falls in people with severe vestibular hypofunction during standard clinical tests when there were no applied external perturbations. Using a body mounted vibrotactile display, Wall and Kentala demonstrated a significant reduction of falls during sensory organization tests (SOT) conditions 5 and 6 from 50% to 15% in eight persons with unilateral or bilateral vestibular diagnoses who had severe deficits as defined by their computerized dynamic posturography (CDP) scores [93]. The SOT 5 is designed to make somatosensory information unreliable while subjects stand with their eyes closed. The SOT 6 is designed to make both somatosensory and visual information unreliable. A second group of subjects with mild to moderate vestibular deficits did not fall as frequently as the severe group, and the number of their falls did not change significantly when they used feedback. Using a head mounted vibrotactile display Goebel et al. showed a similar significant reduction in the number of falls during SOT conditions 5 and 6 in five persons with severe bilateral vestibular loss [32].

The effect of vibrotactile feedback on postural control has also been investigated using externally applied perturbations. Six subjects with vestibulopathies were tested during conditions that induced a mild two-axis random platform motion. Only anterior-posterior (A/P) tilt feedback was provided. Diffusion analysis showed that the vibrotactile feedback of body tilt allowed the subjects to control posture more quickly than without feedback. There was also a significant decrease for all subjects for A/P sway with A/P tilt displayed. The change in medial-lateral (M/L) sway was not significant. This is evidence of direction-specific control [96]. Further studies using pseudorandom sequences showed that vibrotactile feedback is most effective in the 0.02 Hz–0.3 Hz range [33]. Loughlin et al. have used feedback control theory to suggest that a “one size fits all” approach to providing the vibrotactile stimulus to the subject needs to be replaced with an approach that is customized to the individual [65]. This study also suggests that this approach is likely to increase the effective bandwidth of the response.

Vibrotactile feedback also promotes faster recovery from discrete surface perturbations. In one study, Wall and Kentala applied perturbations in the A/P direction using the CDP motor control tests [93]. The results were mixed, and depended upon whether the response pattern was monotonic or oscillatory. There was a tendency for the peak tilt and the time to recover to decrease with vibrotactile feedback. Sienko et al. used multidirectional discrete perturbations to characterize recovery trajectories of subjects with vestibulopathies [88]. The initial trajectory with and without feedback was ballistic and did not vary based on the presence of feedback. Vibrotactile feedback significantly decreased the recovery response times and decreased postural sway following recovery from the perturbations. The spatial resolution of the vibrotactile display was systematically varied from 90° to 22.5°. There was no significant effect on the responses due to the change of spatial resolution. Asseman et al. (2007) investigated the use of vibrotactile feedback during large backward support surface perturbations that elicit step responses in young and older adults and people with either bilateral vestibular deficits or peripheral neuropathies [4]. Only the older adults who exhibited slower stepping times during baseline trials showed significantly shorter stepping reaction times with versus without the vibrotactile cue. Lee et al. (2013) performed a similar study to examine the effects of vibrotactile feedback on the stepping responses of people with Parkinson’s disease (PD) and age-matched controls [59]. The presence of vibrotactile cues did not affect the timing or the length of the steps, but it reduced trunk displacements prior to step initiation. These collective findings suggest that feedback is effective in reducing sway during normal stance and during recovery from perturbations, but not during the ballistic phase of a perturbation.

The visual system provides the CNS with information about the direction of vertical and its inputs are used in combination with inputs from the vestibular and somatosensory systems to maintain balance. Persons with vestibular loss can be expected to have increased reliance on the visual and somatosensory systems for spatial orientation; in an eyes closed condition, they have been found to heavily rely on somatosensory inputs [76]. Persons with bilateral peripheral vestibular loss have demonstrated the ability to significantly reduce pelvis roll and pitch angle sway while receiving vibrotactile and auditory feedback when standing on foam with eyes closed, however they did not demonstrate significant decreases in pelvis pitch angle sway when performing the same task with their eyes open [40]. Persons with peripheral vestibular involvement were able to use visual, vibrotactile and multimodal feedback (all conditions performed with eyes open) to improve postural sway metrics over baseline values during tandem Romberg stance; the best performance was achieved with continuous visual feedback, but some subjects reported dizziness when using it [14]. Subjects with vestibular loss used a smart phone balance trainer providing vibrotactile cues to improve several postural sway metrics during semi-tandem Romberg stance trials regardless of whether their eyes were open or closed; performance was slightly better during the feedback trials with eyes open [54]. Use of multimodal feedback (vibrotactile and auditory) during tandem stance trials with either eyes open or closed resulted in significant reduction in trunk sway in older adults [24].

4 Gait

While fall-related injuries can occur from loss of balance during either standing or walking, the latter accounts for the majority of cases and has proven significantly more difficult to address due to its greater complexity. A main challenge of treating locomotor imbalance is that its negative effects manifest themselves in a number of different gait characteristics, including gait initiation, gait velocity, step length, step width, toe clearance, continuity, symmetry, trunk sway, path adherence, and ability to turn. Although significant progress has been achieved in developing sensory augmentation methods to improve standing balance, attempts to extend this research to locomotion have been quite limited in scope due to a lack of mechanistic knowledge and supporting science regarding which biomechanical signals to measure, how to process them, and how to provide meaningful feedback signals to the user. While the sensory systems providing input to the CNS presumably operate by the same principles during standing and walking, little correlation exists between the metrics and tasks used to assess stability during standing (which has greatest instability in the sagittal plane) and those used to assess stability during walking (which has greatest instability in the frontal plane). For example, a relatively simple measure such as trunk tilt angle might be used as a feedback signal to improve balance during standing, yet effective measures during walking (which likely involve the dynamics of a number of body segments) require additional investigation. Because of this, the majority of gait-based studies have simply asked users to walk in step with an auditory or visual cadence, or have provided vibrotactile stimulation of a single body segment or joint to warn of extension beyond a desired angle.

Limited studies have been conducted to assess the capability of sensory augmentation to improve stability during locomotor tasks for people with vestibular deficits. Among the studies conducted to date, the types of locomotor tasks performed have varied considerably. Horak et al. showed that vibrotactile feedback of M/L trunk tilt reduces M/L trunk tilt in people with unilateral vestibular loss during paced heel-to-toe walking [27]. Hegeman et al. tested six compensated bilateral vestibular loss subjects and found that auditory feedback was not effective in reducing sway during various gait tasks [38]. Sienko et al. demonstrated that use of roll tilt vibrotactile feedback by subjects with compensated vestibular loss resulted in a significant decrease in roll sway solely in challenging locomotor tasks (i.e., narrow stance-walking) [87]. Janssen et al. studied the effects of multimodal head-mounted feedback on trunk motion in young adults during various locomotor activities and found that feedback significantly reduced trunk sway velocities regardless of whether feedback was provided in the A/P or M/L directions [43]. They observed greater reductions in pitch angle with A/P vs. M/L feedback when subjects climbed up and down stairs and walked over barriers, and similar reductions in pitch angle during normal and tandem walking for both feedback directions. Several vibrotactile augmented locomotor-based studies have been performed among the general aging population demonstrating improvements in clinical metrics or gait parameters. For example, Wall et al. showed that older adults could significantly increase their Dynamic Gait Index scores while using vibrotactile feedback [95]. Verhoeff et al. provided multimodal feedback to young and older adults that reduced A/P and M/L tilt and A/P tilt velocity during normal walking [92]. Shull et al. provided vibrotactile cues during walking to reduce knee adduction moments in people with knee osteoarthritis [84].

Two potentially negative side effects have emerged when subjects use sensory augmentation cues following limited training; subjects decrease their gait velocity and move in more of an “en bloc” manner. Janssen et al. (2009) displayed multimodal cues of trunk sway on the heads of young healthy subjects during several gait tasks. The subjects’ trunk velocities decreased in both the A/P and M/L directions for all but one task (tandem steps). Trial duration, i.e., gait velocity, significantly increased when cues were provided [43]. The authors note that in addition to the short training period, subjects were only asked to focus on reducing their sway, not maintaining their gait velocity. Sienko et al. (2012) reported a slight increase in gait velocity in approximately half of the trials when people with vestibular deficits used vibrotactile sensory augmentation during various gait tasks [88]. In this study, subjects were observed to walk less naturally, altering their segmental control strategy likely to prevent themselves from moving beyond the feedback activation thresholds.

5 Dual tasking

One of the main concerns about providing feedback during gait is whether a person can process the external input (visual, auditory or proprioceptive) provided by a sensory augmentation device and concurrently ambulate safely while potentially performing an additional task (e.g., talking on the phone, using a map). Providing feedback during gait in young and older adults while performing a secondary task has been studied [92]. Subjects were asked to walk, walk while verbally counting backwards, and walk while carrying cups of water on a tray. The younger subjects reduced their trunk sway during the feedback trials. Older adults had less sway during the tray-carrying task but not during the counting and walking task, however they did improve on the cognitive task while receiving the feedback [92].

Greater attentional resources were required when performing a dual task during standing in both older and younger persons although it appears that older adults utilized greater attentional resources than the younger subjects [62]. Similar findings were observed in a study involving people with unilateral vestibular loss; the use of vibrotactile feedback significantly reduced reaction time performance in people with unilateral loss and age-matched controls, but people with unilateral vestibular loss had slower reaction times compared to controls [63]. Others have also demonstrated increased auditory choice reaction times in older adults during trials with vibrotactile feedback compared to the trials without feedback [35]. Even though their choice reaction times increased (worsened), their RMS trunk tilt decreased (improved). In another study involving older adults, sub-threshold vibrotactile stimulation applied to the sole of the foot during a dual task did not adversely affect the older adults’ postural control [25].

6 Training/retention/carry-over effects

Only a limited number of mostly uncontrolled studies or laboratory based studies with limited number of training sessions have examined the use of sensory augmentation devices for improving rehabilitation outcomes; sensory augmentation devices have rarely been evaluated outside of the laboratory setting. Decreases in body sway over short periods of time (hours to days) have been observed in people with vestibular deficits after using a vibrotactile sensory augmentation device during a small number of sessions (two sessions, three hours each) [86]. Barros et al. demonstrated significant improvement in SOT CDP testing after two 15-min training sessions with electrotactile tongue feedback every other day for six days (total of 12 sessions); however, the study was uncontrolled and balance improvement was lost soon after training ended [12]. Polat et al. reported improved composite SOT scores for subjects undergoing a regimen combining static and dynamic training positions with electrotactile tongue feedback during ten 20-minute sessions over five days, compared to a control group which participated in an eight-week course of staged traditional vestibular rehabilitation and a loosely controlled home exercise program [77]. However, the measured improvements were not retained for more than a few days. Allum et al. used multimodal feedback to significantly reduce trunk angular displacement during real-time use and over short periods of time (∼1 week) following training (three times per week for two weeks) in older adults and people with unilateral vestibular loss [3]. Basta et al. (2011) significantly reduced trunk and ankle sway as well as subjective symptom scores following short training (daily for two weeks) sessions involving a wide range of people with balance disorders including people with vestibular impairments [13]. Brugnera et al. (2015) showed significant improvements in SOT 5 and 6, the Activities-specific Balance Confidence (ABC) Scale, and the functional aspect of the Dizziness Handicap Inventory (DHI) following balance training (5 times per week for two weeks) with vibrotactile feedback in seven participants with vestibular deficits (with substandard responses to conventional vestibular rehabilitation); these changes were not observed for the six control group participants who did not receive feedback during training [20]. Only one published case study has examined usage over a large number of sessions; this study however, involved a single subject who performed 40 sessions with electrotactile tongue feedback and demonstrated balance improvements that persisted for eight weeks after the final session [5].

Technologies such as the Wii Fit and Kinect systems, which provide visual cues, have been used as an adjunct to vestibular rehabilitation and have led to decreased postural sway after six weeks of participation in healthy older adults [91]. In a pilot study by Wang et al., persons with vestibular dysfunction demonstrated decreased postural sway and improved balance function after completing 20 sessions of virtual reality enhanced vestibular rehabilitation within a 4-week period [97]. A preliminary torso-based vibrotactile sensory augmentation 6-week training study involving people with vestibular deficits suggests maintenance of improved balance performance as indicated by SOT composite scores, Mini-BESTest scores, and gait speed one month following training [10]. Exercise interventions for fall prevention require more than 10 weeks of progressively challenging balance exercises with an ongoing home program to maintain the benefits [21, 95]. For example, a recent review of balance training literature with community dwelling older adults concluded that a training period of 11–12 weeks with 36–40 sessions at 31–45 min/session is most effective for improving balance performance outcomes [60]. Furthermore, two studies have examined the negative effects of “detraining” occurring 6 to 12 weeks following documented improvement in static/dynamic balance measures and recommend ongoing balance programs, especially to negate age-related declines [79, 90].

A major barrier to performing long-term training studies is subjects’ unwillingness and/or inability to travel to a clinical or research setting for a large number of sessions. Lee et al. (2012) recently described a smart phone balance trainer that can provide vibrotactile sensory augmentation along a single axis (Lee, Kim et al. 2012). An updated version of this technology provides vibrotactile feedback in both the A/P and M/L directions and includes multimodal instructions for balance exercises including icons depicting the conditions, minimal written text and videos [22, 58].

7 Mechanism

The mechanism by which information is integrated and used by the CNS is not well understood. The dominant hypothesis, which has not been supported by rigorous experimental evidence, holds that observed balance improvements are due to sensory reweighting: feedback of body motion provides the CNS with a correlate to the inputs from its intact sensory channels (e.g., vision, proprioception), so subjects receiving sensory augmentation learn to increasingly depend on these intact systems. Other possible mechanisms for observed improvement that merit further exploration include, but are not limited to: cognition (processing of sensory augmentation information is solely cognitive with no selective adjustment of sensory weights by the CNS), “sixth” sense (CNS interprets sensory augmentation information as a new and distinct sensory channel), context-specific adaptation (new sensorimotor program is developed through repeated interaction with the device and is accessible only when the device is used), and combined volitional and non-volitional response [53, 68].

Honegger et al. investigated movement strategies and muscle synergies when subjects with bilateral vestibular loss were provided with mulitmodal feedback regarding pelvic sway angle during static balance activities. Feedback reduced amplitudes of EMG activity ratios from pairs of antagonistic muscles on the lower leg, trunk and upper arm. However, subjects with bilateral vestibular loss used the same movement strategies as healthy controls [40].

Multiple mathematical models have been developed to describe how sensory augmentation may affect postural control. For example, Goodworth et al. (2009) created a single-inverted pendulum model that incorporated the information provided by a torso-based vibrotactile sensory augmentation device with native intact sensory inputs, relatively upstream in the perceiving, processing, and responding pathway [33]. Ersal and Sienko (2013) developed a multibody model that incorporated information provided by a torso-based vibrotactile sensory augmentation device further downstream, preceding the generation of joint torques [29]. Both of these models were effective in capturing the kinematics of people with vestibular deficits using sensory augmentation. Goodworth et al. (2011) also described a time-delayed sensory feedback control model that effectively represented the motion of people with severe vestibular loss using sensory augmentation during perturbed stance and predicted that postural performance could be further improved by altering the feedback parameters [34].

Researchers have debated whether the reductions in postural sway observed when sensory augmentation is provided are the results of the use of the information conveyed or the effects of stimulation alone. Some hypothesize that stimulation alone – devoid of meaningful information – can serve as an alert mechanism to shift one’s attention to the balance task or possibly elicit a general stiffening behavior. In a study involving a group of people with vestibular loss, Sienko et al. (2008) provided erroneous information during continuous multidirectional surface perturbations and demonstrated that people performed worse than when meaningful cues were provided [86]. During discrete perturbations, people also performed worse when erroneous feedback was provided compared to the feedback off condition suggesting that people are actively using the information to make postural corrections [88]. It should be noted however, that in the presence of erroneous information, people will not continue to use the information if it results in poorer performance (i.e., causes them to take a step or fall) or if the information is noticeably conflicting with information from intact sensory systems (e.g., somatosensory). Therefore it is important to introduce sham conditions at the end of data collection sessions, because subjects will not trust the information from the device after they’ve concluded that it isn’t representative of their motion. Janssen et al. (2009) points out that training in and of itself has an effect on performance and not all gains in performance should be attributable to sensory augmentation [44]. It is also not unreasonable to think that people receiving cues interpret the cues as instructions to limit their sway and therefore limit their sway in the presence of any cues (meaningful or not). Wildenberg et al. (2010) showed that stimulation that doesn’t convey information about body posture or gravity can improve balance performance in people with balance dysfunction [102]. They’ve also investigated the effects of electrotactile tongue stimulation on balance performance and have used imaging techniques such as MRI and fMRI to show that stimulation upregulates visual sensitivity to optic flow in people with balance impairments [102–105].

8 Sensory augmentation technology design considerations

Numerous studies have investigated the roles of various feedback display design-related factors on balance performance. Proportional plus derivative feedback has been shown to be significantly better than proportional or derivative feedback alone in a vestibular population during CDP [94]. In a study investigating the effect of feedback modality on balance performance during standing tasks, subjects with vestibular loss performed better while using visual feedback, which provided information about current and future positions, compared to vibrotactile feedback based on a proportional control signal presented on the torso [14]. Continuous vibrotactile feedback produced slightly better results than periodic feedback during locomotor tasks in people with vestibular loss [87]. Displays positioned on the torso [54, 96] or head [3, 32] have yielded positive results during real-time use in people with vestibular deficits. However, faster reaction times have been observed when vibrotactile stimuli are applied to the head versus the torso [11] or sternum [4]. Both attractive (“move toward the vibration”) and repulsive (“move away from the vibration”) cuing strategies are effective, but repulsive cues yield slightly better results during short-term usage, while the rate of improvement is greater for attractive versus repulsive cuing [46].

9 Potential applications

10 Limitations

This narrative review highlights findings from the literature, but does not critically examine efficacy or the quality of the research methods used in the reviewed studies.

11 Summary

Sensory augmentation has been shown to reduce postural sway in people with vestibular deficits during static, dynamic, and mild-moderate perturbed stance conditions. Its usefulness during locomotor activities has been limited; however, increased training time and different feedback strategies may improve efficacy. The lack of significant findings during step-inducing perturbations and gait may limit the applications as an all-encompassing real-time balance aid. There have been limited systematic studies to determine the relationship between the number of training sessions with sensory augmentation and the retentive and/or carry-over effects, but a handful of studies that have involved more than 10 sessions seem to indicate benefits over training alone (i.e., without the addition of sensory augmentation). If shown to be effective following use during balance rehabilitation training, sensory augmentation may further facilitate functional recovery in patients who have reached a plateau with their traditional vestibular/balance therapy programs. Sensory augmentation devices designed for use in the home may provide an important transitional link between clinic and home with potential to provide an ongoing balance intervention to high fall risk populations.

Competing interests

The authors state that C. Wall is an inventor on an issued patent and has equity interest in BalanceTek, Inc.