Medical, surgical, and rehabilitative approaches are all used to treat balance problems. Even though all of these approaches have greatly improved over time, they are still not 100% effective in all cases, and there is still room for improvement or for other approaches. Vestibular prostheses—although not yet available—could provide another option for those with balance problems as well as for those who are elderly and prone to falling (Wall, 2002).
Vestibular prostheses include a range of approaches to vestibular impairment. All approaches rely on motion detectors that measure head movement in all three directions and send the signal back to the central nervous system (CNS) indirectly. A sensory substitution device sends the information to the CNS through sound, vision, tactile vibration, or electro-stimulation of the tongue, while an implantable vestibular prosthesis conveys information through selective stimulation of the vestibular nerve.
These prostheses could be used for permanent replacement of vestibular function; as a temporary replacement following surgery; as part of vestibular/balance rehabilitation or fall prevention in the elderly; and for basic research on sensory interactions using animal models. However, not all candidates may want or need an invasive, surgically implanted device. Even if an implant proved to be more effective than using sensory substitution in some applications, the surgical risks may not outweigh the benefit for all patients. For some patients, electrical stimulation of the end organ will provide the greatest benefit, but for others, sensory substitution may be more appropriate.
Three factors can be used to predict when vestibular prostheses are likely to be effective. One predictor of success for vestibular implants would be the presence of stimulable neural material. Temporal bone studies have already determined that there is still enough neural material available for electrical stimulation in individuals with common otologic diseases that can damage hair cells as well as in those who experience a loss of hair cells due to natural aging.
A second factor would be a lack of response to other forms of treatment. Implants might be useful for individuals with bilateral vestibular hypofunction who cannot rely adequately on other senses during everyday activities. Another use would be for those with unilateral vestibular hypofunction whose central nervous systems have a tonic imbalance to which they cannot adjust.
A third, critical factor for success is the ability to place the electrodes of an implantable device near the targeted neural matter in order to stimulate the target without stimulating other undesirable targets, such as the facial nerve. There is some preliminary agreement that the most desirable targets would be the nerve fiber bundles that separately innervate the three individual semicircular canals.
In Figure 1 [PDF, 1.4MB] (a highly schematic diagram), the vestibular input to the central nervous system is illustrated in the upper left corner and some of the responses to this input are on the right. In each inner ear there are five motion sensors: three semicircular canals (horizontal, anterior, and posterior) that sense angular motion, and two otolith organs (saccule and utricle) that sense linear motion and gravity. Each of these five sensors works on the same principle shown in the upper left corner. Acceleration of the head in space causes a motion-detecting seismic mass (M) to be displaced away from its equilibrium position. This relative displacement within the inner ear is detected by sensory hair cells that are mechanically coupled to the seismic mass. In the diagram, motion to the right (shown by the large green arrow) has caused the seismic mass to be displaced to the left and toward the kinocilia of the hair cells, thus depolarizing them. This process modulates ongoing, spontaneous neural firing in the vestibular portion of the eighth cranial nerve.
The central nervous system processes motion inputs; this process can affect autonomic function, provide spatial orientation cues, help stabilize vision, and help maintain equilibrium while standing and walking. An increase in the overall neural activity above the spontaneous background activity drives the eyes in one direction, while a decrease to below the background activity drives the eyes in the opposite direction. This activity is the basis of the vestibuloocular reflex (VOR). Individuals without a functional VOR report that they cannot recognize a person's face or a street sign while walking, and must stop to see these objects clearly. Thus, the ability to engage in everyday activities is strongly affected by the presence or absence of this important reflex.
A lesion or break in the three-link transduction chain from the inner ear to the brain, shown in the figure by Xs in the eighth nerve, compromises the ability to interact with one's motion environment in many ways. One way to replace this lost sensory input—at least partially—is to use a prosthesis, as shown in the bottom left of the figure. A motion sensor package that uses linear accelerometers and rate-of-turn gyroscopes sends its signals to a processor that can code the motion in physiologically useful ways, and then to a stimulator that feeds the signal back to the central nervous system via one or more paths. The implant version uses electrodes in contact with parts of the vestibular sense organ or its connecting nerves, while the sensory substitution version sends the information to the central nervous system by other means, which could include sound, vision, tactile vibration, or electro-stimulation of the tongue.
Motion Sensor Technology
Motion sensors are approaching acceptable accuracy for the patient and size for the surgeon to allow for implantation, but must achieve a lower power consumption and longer battery life. Which motion detectors are best suited to different sensory substitution applications is still under investigation. Processors for implant applications must be both fast and power-efficient.
Another approach would be to borrow from cochlear implant technology, using electrode designs that are based on cochlear implants and silicon-array technology. Contemporary cochlear implant electrodes can be configured with closely spaced multiple-electrode linear arrays. After implantation, the most effective sites of stimulation can be determined by the audiologist by trial and error. One approach for a vestibular implant is to create an opening in the semicircular canal surgically, quickly insert a similar multi-site electrode array, and seal the opening. Post-operatively, the one or two most effective sites for stimulation could be determined. Another possibility that also mimics current auditory implant technology would be to use multi-electrode arrays to stimulate the vestibular brainstem nuclei, or possibly at Scarpa's ganglion where the first-order afferent cell bodies are located. This method is similar to the cochlear implant practice in which multiple-electrode arrays stimulate the tonotopic auditory nerve fibers.
Sensory Substitution Approaches
Sensory substitution devices present a potentially attractive alternative to implants. They do not carry the risk of surgery, and they will likely be available for clinical use much sooner than implants will become available. There are two important questions regarding sensory substitution devices: how should body motion be displayed to get information into the central nervous system, and what motion sensors best characterize that information? Displays now under investigation include lingual electro-stimulation, auditory biofeedback, and vibrotactile tilt feedback (VTTF) (Bach-y-Rita, 1998; Dozza, Chiari, & Horak, 2005; Wall & Weinberg, 2003). Each type of display has advantages and disadvantages. Our group has chosen VTTF because it uses a sensory channel that is not normally used for communication. Our group has tried all three types to determine how they work in different contexts/environments.
There are three engineering options for motion sensors that use:
- Accelerometers, which cannot disambiguate tilt (gravity) from translation (linear motion)
- Rate gyroscopes, which provide inaccurate tilt estimate due to integration of the random bias signal
- A combination of accelerometers and gyros,which provides an accurate tilt estimate, but is more expensive
Our approach has been to use the combination because it yields an accurate estimate of body tilt that has acceptable error rates over relatively long time periods (e.g., less than 0.2 degrees in 10 hours; Wall & Weinberg, 2003).
We have used the best-available microelectromechanical systems (MEMS) motion sensors, knowing that the cost would decrease as the technology was commercialized. Although the cost of the first-generation devices was quite high, the cost for eachsuccessive generation technology fell dramatically—from $300,000 to $35,000 to $300 in less than a decade, a trend that is likely to continue.
Vibrotactile Tilt Feedback
Our VTTF device senses body motion with gyroscopes and accelerometers mounted on the small of the back, processes the motion-sensor information to make an estimate of tilt from the vertical, then feeds this information back to the patient using noninvasive, vibrating elements (tactors) that circle the trunk. One display scheme, called position-based coding, is shown in Figure 2 [PDF, 1.4MB]. The position of the actuated tactor travels up the trunk as the amount of body tilt in that direction increases.
Improved Postural Control
We studied 17 vestibulopathic subjects' performance using computerized dynamic posturography sensor organization tests, with and without VTTF. In subjects with moderate deficits, VTTF significantly reduced their body sway, while in subjects with severe deficits, VTTF significantly reduced the percentage of falls under the challenging conditions of sensory organization tests 5 and 6, which make other sensory inputs unreliable (Wall & Kentala, 2006).
Reducing Falls in Older Adults
Can older adults who are prone to falling benefit from VTTF while walking? An estimated 7.4 million people between the ages of 65–85 in the United States have experienced recurrent falls (Tinetti, Speechley, & Ginter, 1988), a number that is increasing by about 240,000 per year. Our group compared the dynamic gait index in 12 healthy elderly people with an average age of 79 with and without VTTF while they performed typical locomotor tasks. Subjects with initial gait index scores of 19 or less—scores indicative of an increased propensity to fall—were trained for 30 minutes. When VTTF was turned on, the gait index showed a three-point increase that elevated the scores into the "no fall zone" for this index, a finding that was both statistically and clinically significant for this study population.
VTTF may help those with acute imbalancefrom stroke, surgery, accident, vestibular neuronitis, and amputation as well as those with chronicimbalance—elderly people prone to falling or those with uncompensated balance dysfunction. For these populations, we envision a logical progression of VTTF devices. The first device might be a laboratory-based device dedicated to balance/vestibular rehabilitation. Clinicians may want to provide their patients with a take-home version of the device to supplement the clinical training. Finally, there might be a device for elderly individuals who are prone to falling that could be worn full-time under clothing and would need only a simple alignment calibration once a day. A prototype of this device is shown in Figure 3 [PDF, 1.4MB].
The VTTF device may be used as a sensory substitution device to prevent falls or to augment rehabilitation. Preliminary results indicate improved balance when standing and walking, and rehabilitation specialists will play a vital role in all aspects of the device. The emerging approaches to vestibular prostheses will open up new avenues, in addition to balance testing, for audiologists to help people with balance problems.
Cochlear implant technology has led the way for the development of other implantable medical devices that use electrical stimulation to help people with stroke, blindness, incontinence, glaucoma, hydrocephalus, and chronic pain as well as balance impairments. Borrowing technology from other biomedical electrical implant applications, researchers are on their way to bringing vestibular implants to market that would provide electrical stimulation to the vestibular nerve. The following four pioneering studies highlight significant research milestones over the last 40 years and provide support for fundamental issues that laid the foundation for the development of vestibular implants.
Eye Movements in Planes of Electrically Stimulated Canals
More than 40 years ago, Suzuki and Cohen (1962) conducted experiments that were critical to the development of some vestibular prosthesis applications. The researchers showed that in a squirrel monkey, electric stimulation of the bundles that innervated each semicircular canal produced reflexive eye movements whose direction was specific to each canal. Stimulation of the horizontal semicircular nerve bundle produced eye movements that were in the plane of that canal, essentially resulting in a horizontal eye movement. This same result held true for stimulation of two other semicircular canals. Theoretically, it is possible to drive the motion of the eyes in any arbitrary direction by selectively stimulating the three semicircular canal nerve fibers in combination.
Ability to Drive VOR with Unilateral Stimulation
Motion input modulates the spontaneous activity in the vestibular nerves. The two sides work in push-pull opposition in which the total input is the difference between right and left sides. A unilateral implant would introduce a new signal to one side, creating an imbalance of tonic activity between right and left ears.
Is unilateral electric stimulation sufficient for a functional response? Lewis and Merfeld (2002) plugged both horizontal canals in the squirrel monkey, making them much less sensitive to normal motion. They inserted a stimulating electrode near the sensory region or ampulla of the right horizontal canal, and modulated the frequency of electrical stimulation. They were able to drive the eye movements toward the right or the left using only unilateral stimulation. These initial responses did not produce very large eye movements because the experimenters purposely used a relatively small-amplitude electric stimulus. The crucial result was not the amplitude of the response, but the finding that a stimulus in one side (or ear) could actually drive eye velocity in both directions.
VOR Modulation and Directional Specificity
Della Santina (2007) assessed performance using three-dimensional angular VOR measurements in the canal-plugged chinchilla model, and rate-modulated trains of electric pulses to drive the movement of the eyes. Stimulation of the horizontal canal produced horizontal eye movement components that mimicked the "ideal" VOR response. But there were also other response components in other directions, suggesting the need to increase the selectivity of the stimulating electrodes.
Human Eye Movements in Plane of Electrically Stimulated Posterior Canal
At University Hospital of Geneva, Wall, Kos, and Guyot (2007) conducted short-duration experiments that stimulated the nerve bundle of the posterior semicircular canal in three patients immediately preceding cochlear implantation. The patients' responses demonstrate that compensatory eye movements can be produced that are both robust and controllable—nystagmus velocity is modulated by pulse rate and pulse amplitude.
Preliminary findings and developments support the idea that a vestibular implant is feasible, both technically and physiologically. Such an implant can reach patients only after thorough clinical, technical, and physiological investigations. The National Institutes of Health (NIH) has recognized the research progression from bench to bedside, and has taken a leadership role by holding a workshop on the electric stimulation of the vestibular nerve, sponsoring directed research to answer some remaining key questions using animal models, and by soliciting proposals for multi-electrode arrays for use in vestibular prostheses. These steps will play a strong part in accelerating vestibular implants toward clinical use.
Our team has received the most positive reaction to balance prostheses from those who suffer from mal de debarquement syndrome (MdDS). Typically, these people take a trip by ship or airplane, and experience "sea legs"—the feeling of constant motion—that does not go away after the trip, as it does for most people. There is no remedy for MdDS, but two subjects who have tried vibrotactile tilt feedback both reported that the feedback greatly reduced their motion symptoms by providing them with a true "artificial horizon."
Our team also hears from the relatives of elderly individuals who are beginning to experience unsteadiness and who are becoming apprehensive about going out alone because they might fall. As a result, the elders begin to limit their activities of everyday living and have an increased fear of falling. In the future, our greatest hope is to improve the quality of life for these elderly individuals who are prone to falling by using a balance prosthesis to help increase their confidence and their potential to live independently.
Disclosure: At this time, the author has no financial interest in vestibular prostheses. His research was supported by The Keck Foundation and the National Institutes of Health.
Contributors & Acknowledgements
This paper is based upon a talk I was invited to give at the 2007 ASHA Convention in Boston. I would like thank the meeting organizers for inviting me to present at this prestigious meeting, and to gratefully acknowledge the many contributors whose names and affiliations are listed below:Charley Della Santina, Johns Hopkins University; Marco Dozza, Neurological Sciences Institute (NSI); Erna Kentala, Massachusetts Eye and Ear Infirmary (MEEI), Helsinki University Central Hospital; Jean-Phillip Guyot, University Hospital of Geneva; Fay Horak, NSI, Oregon Health Sciences University (OHSU); Izabel Kos, University Hospital of Geneva; Heather Kubert, (MEEI); Rick Lewis, MEEI/ Harvard Medical School (HMS); Dan Merfeld , MEEI/HMS; Lars Oddsson, Boston University (BU); Robert Peterka, NSI, OHSU; Stephen Rauch, MEEI/HMS; Jimmy Robertsson, MEEI, BU; Angus Rupert, MD, US Navy; Kathleen Sienko MIT/HMS, MEEI; Beatriz Silveira MEEI; Marilee Stephens, University of Buffalo (UB); Marc Weinberg, Draper Lab; Diane Wrisley, UB; Neurocom International for donation of Balance Master software; Funding: the Keck Foundation & NIDCD R01 DC 06201.Finally, I would especially like to acknowledge and thank Dr. Charles Berlin for his initial interest and encouragement to me in my career. He has been a great inspiration.