Auditory contributions to maintaining balance

1 Introduction

The brain possesses a remarkable ability to link together multiple sensory inputs into seamless perceptions of environmental objects and conditions, providing much more robust perceptions than any single sense alone [2, 11]. Maintaining balance has classically been considered to require the successful integration of visual, proprioceptive and vestibular inputs. The auditory system, which operates with remarkable speed and generates a fine-grained representation of the three-dimensional space surrounding an individual, would be an obvious additional source for balance-related information. However, thus far only congenitally blind people have been shown to improve postural sway based on external auditory information [9]. As this represents a select population subject to remapping of the brain’s sensory cortex and adoption of unique auditory techniques assisting their navigation through the environment [13, 15], the question arises whether the balance of other individuals can also benefit from auditory inputs. Here, we tested the influence of hearing on balance by systematically manipulating the amount of spatial auditory, visual, and proprioceptive cues available to human subjects and measuring their ability to maintain an upright stance.

2 Methods

All experiments were conducted with the approval of the Washington University School of Medicine Human Studies Committee. We recruited a sample of 18 subjects (9 to 78 years old, mean (±standard deviation) age 47±20 years; 10 female). Twelve subjects had no subjective balance problems and six were recruited from a neurotologic clinic with imbalance. Two patients had idiopathic unilateral vestibular loss as demonstrated on caloric testing, one had Pendred’s syndrome, one had a history of meningitis, one had a history of gentamicin toxicity, and one had positional bouts of vertigo typical of BPPV with ongoing imbalance between episodes. All patients had hearing better than a pure-tone average of 30 dB HL (0.5–2 kHz) bilaterally, with the assistance of hearing amplification as needed. All reported normal or corrected to normal vision.

Each subject stood on a balance platform (Equitest, Neurocom, Clackamas, OR) instrumented to allow measurement of center of pressure. Subjects were tested in the dark and in the light, with and without spatial sound cues, and on a stable or an unstable support surface. Visual cues were provided by a high-contrast image of an abstract distant landscape, illuminated with ambient indoor lighting, located 72 cm in front of the subject and occupying the entire visual field. In the lights-off condition, these cues were eliminated by blindfolding subjects with their eyes closed. In the sound-off condition, omnidirectional background room noise was quantified as 35 dB SPL (A-weighted scale, Model 831 sound pressure meter, PCB Piezotronics Inc., Depew, NY). In the sound-on condition, spatial auditory cues were provided by noise generated by the wgn (white Gaussian noise) function in MatLab (version 7.5, The Mathworks Inc., Natick, MA) and were presented through four small (2 cm diameter) speakers (SPKR-R1-BK-L02, GrandMax, Piscataway, NJ) placed at ear level 62 cm in front of the center of the interaural line, 50 cm behind, and 43 cm on both the left and the right, delivering sound at 55 dB SPL (Fig. 1). The noise was uncorrelated among the speakers, so that the same exact noise was not broadcast by all four speakers simultaneously. The auditory stimulus was present for five seconds before recording of force plate measurements was begun.

We also evaluated the contribution of external, spatial “earth-fixed” sound sources as described above to the effect of a “head-fixed” sound source. For that condition, subjects wore ear muffs (-26 dB noise reduction ratio, Leightning L3, Howard Leight/Honeywell, San Diego, Calif. USA) over standard ear buds (Apple Inc., Cupertino, Calif., USA). The ear buds were adjusted by the subject to match the external sound source’s amplitude subjectively.

The position of the center of pressure was recorded at 100 samples/s and was rectified, differentiated, and smoothed using a cubic spline to calculate speed. “Sway” was defined as the root-mean-square of this value. This measure has been shown to represent a useful marker for balance and correlates to the risk of falling [17]. Individual testing conditions consisted of three twenty-second blocks. The blocks were interleaved among the different conditions to prevent learning effects.

3 Results

Raw data tracings from one subject are shown in Fig. 2. In the unfavorable sound-off, lights-off condition, sway (measured as the root-mean-square of velocities collected over each trial) ranged from 0.9–7.0 cm/s across all participants. With the addition of spatial sound, sway decreased to a range of 1.0–4.7 cm/s. There was a close linear relationship between sway in the sound-on condition as a function of sway in the sound-off condition, where subjects with greater baseline sway (i.e. more imbalance) showed more improvement than those with less sway (Fig. 3). The slope of the line relating these conditions was 0.59 (Pearson’s product-moment, 95% CI: 0.49–0.68; r = 0.96, 95% CI: 0.89–0.98) indicating that adding sound reduced sway to less than two-thirds of what it was in silence.

We also analyzed the six imbalanced and the normal subjects separately and found that both groups demonstrated an improvement in balance with the addition of sound. For the imbalanced group, the slope of the line relating the sound-off to sound-on condition was was 0.60 (Pearson’s product-moment, 95% CI: 0.36–0.85, r = 0.96, 95% CI: 0.67–0.99). For the normal group, the slope of the line relating the two conditions was 0.61 (Pearson’s product-moment, 95% CI: 0.46–0.76), r = 0.94, 95% CI: 0.80–0.98).

We next sought to determine how the improvement in balance (i.e., decreased sway values) due to auditory input compared to the benefit provided by visual cues alone. In the lights-on, sound-off condition, sway ranged from 0.7 to 2.7 cm/s. The decrease in sway due to visual cues showed a close linear relationship to the decrease due to auditory cues, with the subjects improving most in response to one modality being the same as those who improved the most with the other. The slope of the line relating the two conditions was 0.54 (Pearson’s product-moment, 95% CI: 0.38–0.69; r = 0.88, 95% CI: 0.70–0.95), indicating that auditory cues alone can provide more than half of the benefit of visual cues for improving balance over the sound-off, lights-off condition (Fig. 4). We were unable to detect a decrease in sway reliably, however, when sound was added to an already illuminated environment, implying that audition can play a significant role in some conditions but its contribution may remain relatively limited when strong visual cues are present (two-tailed Wilcoxon signed-rank test, p = 0.5093) (Pearson’s product-moment, slope = 0.89, 95% CI: 0.54–1.23; r = 0.81, 95% CI: 0.55–0.93). In the imbalanced group with their eyes open, Pearson’s product-moment, slope = 1.39, 95% CI: 0.57–2.21, r = 0.92, 95% CI: 0.43–0.99. In the normal group with their eyes open, Pearson’s product-moment, slope = 0.27, 95% CI: 0.01–0.54, r = 0.59, 95% CI: 0.03–0.87

In addition to vision, proprioception provides another important input for maintaining balance. We tested 15 of the subjects in the dark standing on a “sway-referenced” surface, which limits proprioceptive sensations of posture by moving in tandem with subject sway. Proprioceptive input proved to be such a strong contributor to balance that only ten of these subjects were able to stand upright in the sound-off condition. Sound did not improve performance for these ten subjects as a group (p = 0.61, two-tailed paired t-test) but a large improvement of 3.0 cm/s was seen in one patient from the normal group and 3.7 cm/s in one patient from the imbalanced group. Remarkably, two subjects (from the imbalanced group) who were incapable of standing upright under sway-referenced conditions with no sound were able to maintain their balance with the addition of auditory input.

Finally, we considered the possibility that the improvement in balance seen with sound was due to increased attention or an alerting effect of the sound rather than providing spatial information itself. There was no difference in sway velocity between subjects standing without additional earth-fixed sound sources and subjects standing with head-fixed sound (p = 0.25, two-tailed, paired t-test).

4 Discussion

The data presented here offer compelling evidence that external sound stimuli can provide significant improvement in postural stability, thus helping demonstrate that the auditory sense can combine with other cues to provide useful information about body orientation. The benefit each subject received from sound inputs was inversely related to the degree of imbalance without sound and was minimal when visual cues were also present. This is consistent with the general concept in sensory physiology that the contribution of any particular sensory cue will be in proportion to its relative precision when combined with other inputs [10]. The lack of effect we saw in a well-lit environment, for example, could be due to the angular resolution of vision being only one minute of arc, while the minimal audible angle in the horizontal plane is only about 1 degree [16, 19].

Our findings, while limited by the relatively small group of participants, suggest that auditory information appears likely to be clinically important, at least in situations where vision is limited. Although the extreme visual condition used here simulates complete blindness, other visual deficits, such as glaucoma, presbyopia, cataracts, and macular degeneration are common among the elderly and can be severe enough to degrade visual input substantially. Individuals with these intermediate levels of vision loss will likely experience a significant balance benefit from augmenting their visual inputs with spatial auditory information. Auditory input might have similar beneficial effects on posture in patients with loss of proprioceptive or vestibular information. Under certain circumstances it may be that auditory input is more likely to reduce sway in particular directions, for example in the anteroposterior versus the lateral direction. We were unable to evaluate that possibility, as we measured only overall sway rather than decomposing it into orthogonal directions.

Hearing loss is associated with an increased risk of falling among the elderly [14, 20]. This could be due to either a global loss of function shared between the cochlea and vestibular system, as has been demonstrated in subjects with noise-induced hearing loss [1], or a loss of spatial auditory cues contributing to balance. The results here support the second possibility, although they do not specifically rule out the alternative explanation that global labyrinthine loss is at least partially responsible. This insight suggests that optimizing spatial auditory cues may be a novel method for reducing the risk of falling, particularly in subjects with compromises of vision or even other potentially offering a significant public-health benefit through environmental sound modifications and/or the use of hearing amplification or cochlear implantation to improve balance as well as hearing. Indeed, a recent study has documented that, in experienced hearing aid users over the age of 65, stability is improved when they wear their aids versus without amplification [18].

The results presented here, where static external auditory cues served as stable spatial landmarks, must be contrasted with previous work examining the effect of external auditory sources programmed to provide feedback in an effort to improve balance. In one set of experiments, the frequency and amplitude of four speakers surrounding the subject were systematically varied in response to the sway of subjects standing on a pressure plate [4, 6–8]. In one set of these feedback experiments, speakers in front of and behind a subject changed their frequencies in correlation with the amount and direction of front-back sway and speakers placed at the sides of the subject changed their amplitude in response to lateral sway. Overall, the positive effect of feedback from varying auditory sources is analogous to that provided from other feedback techniques including the use of vibratory and electrotactile devices [5, 23].

A few previous studies have examined the effect of fixed environmental auditory stimuli. In one, sway among congenitally blind and normal-sighted people was measured when sound was provided from a pair of speakers placed 5 cm lateral to each ear [9]. This study did show some improvement in subjects when collapsed across groups. However, our study expands on those previous results in several ways. First, our data include only people without visual losses. This may be an important confounding variable, as remapping of the sensory cortex among blind people could affect the results [13, 15]. Second, our speakers were significantly farther away from the ear than the 5 cm standoff used previously. This short distance might accentuate the effect of sound and might limit its generalizability to other more typical environmental sound sources. Third, we found a clear dose-response relationship so that subjects with greater baseline sway had greater improvement from the addition of auditory information. Finally, and perhaps most importantly, our work shows that the importance of auditory input is not necessarily overwhelmed by the very challenging situation when the substrate is unstable and proprioception is limited. Finally, our data here provide convincing proof that it is the presence of earth-fixed sound sources rather than attentional mechanisms that improve balance.

At least two other studies deserve mention. First, a related work placed subjects standing on a Nintendo Wii platform into a sound booth with and without auditory input. The authors reported an improvement in postural sway in the presence of sound, but the results were inconsistent and did not reach statistical significance when multiple comparisons were taken into account [12]. Another study reported the ability of an external sound source to stabilize the head in space while standing in place [22]. This reported a reduction in head movement in the presence of an external auditory source, but the effect of sound on postural stability was not evaluated.

Our finding that auditory cues contribute meaningfully to balance may have implications for questions of comparative anatomy and physiology. Animals that cannot always rely on visual cues, such as nocturnal or cave-dwelling animals, or some bats and marine mammals, may have developed neural pathways that are particularly facile at providing spatial auditory information. In the case of cetaceans, which have remarkably small vestibular systems and exceptional auditory ability [3], auditory inputs may be particularly important contributors of spatial information for inferring body orientation.