"Listen carefully and tell me where the sound is coming from," I explained to "Emma," a 12-year-old, bilateral cochlear implant user whose second CI had recently been activated. "What do you mean by 'where'?" Emma asked. "When I hear sounds, they are in my head or very near my ear."
The experience of Emma (not her real name) is typical of tens of thousands of unilateral cochlear implant (CI) users, who generally do not experience spatial hearing. For these individuals, the spatial auditory world often collapses into the head rather than being perceived at external locations. Unilateral CIs provide children the opportunity to attain spoken language, and offer adults renewed access to a hearing world. However, the lack of input to both ears can be a challenge.
In the spring of 2002, Emma embarked on a journey experienced by few children at that time when she became one of the first children in the United States to receive a second CI. When pediatric bilateral CIs were first provided, the candidates typically were children between 5 and 12 years of age who already had attained notable success with their first CI. By participating in research at the Binaural Hearing and Speech Lab in Madison, Wis., Emma and her cohort of pioneer pediatric bilateral CI users contributed to a growing body of knowledge about the emergence of spatial hearing skills in this population. Prior research had focused on communication milestones and the importance of the age of implantation (for the first- and typically only-implanted ear); this new line of research was aimed at determining whether bilateral implantation would provide measurable, significant benefits on tasks required for functioning in complex multi-source auditory environments.
In the process of developing this research program, we considered what was known about binaural hearing in people with normal hearing and in adults who had received bilateral CIs. The ability of humans to function in complex listening situations depends on whether they can solve the "what" and "where" problems for everyday sounds. The well-known "cocktail-party effect," in which listeners struggle to focus on the target speech signal and ignore other talkers in the room, has been studied for decades. In these studies, people with normal hearing listen through headphones to sounds that are presented from simulated locations in space. With virtual space technology, tight control can be exerted over which ears are activated, and monaural deafness can be simulated. Adults can understand speech in the presence of background sounds much better when both ears are activated (Hawley et al., 1999; 2004).
In particular, binaural hearing offers benefits in segregating sounds that arrive from different locations. Consider the cocktail party example depicted in Figure A [PDF], in which the target speech signal and the competing talkers are spatially close, rendering the target difficult to understand. In the second panel of Figure A [PDF] the position of the competing talkers is shifted by 90 degrees to the side, leading to significant improvement in the listener's ability to understand the target. The benefit is due to the spatial separation of the target signal and competing speech, which is known as spatial release from masking. Some advantage is due to the head shadow effect, in which the ear farther from the competitors (the left ear in the figure) has an improved signal-to-noise ratio. However, the advantage also is due to binaural mechanisms that are mediated by neural circuits in which the inputs from the right and left ears excite spatially tuned neurons in the auditory pathway.
In addition to enhancing speech understanding, binaural hearing enables listeners to identify sound source locations by utilizing neural circuits that integrate inputs from both ears and using acoustic cues known as inter-aural differences in time and intensity (level).
Bilateral CIs in Adults
In the 1990s a small group of audiologists and surgeons in Australia, the United States, and Europe initiated an effort to provide CI users with bilateral devices. Initial funding was provided by the National Institutes of Health and some cochlear implant manufacturers to determine feasibility and to develop measure outcomes. This work was initiated in adults to improve functional abilities in complex auditory environments. Studies on adults in controlled laboratory environments using loudspeaker arrays have demonstrated significant benefits in performance attributed to the use of two implants compared with one implant (e.g., van Hoesel & Tyler, 2003; Litovsky et al., 2004; 2006a; 2009; Nopp et al., 2004).
For example, when adults are asked to identify the location of sound sources positioned in a horizontal plane, localization errors (quantified as root-mean-square error, or RMS) typically average 20–30 degrees in bilateral listening modes and 50–60 degrees in monaural listening modes, demonstrating a significant benefit for bilateral CI listening. However, performance in the bilateral mode is still poorer than that seen in listeners with normal hearing, whose errors are 5–10 degrees (see Figure B [PDF]).
Studies with adults also have shown that bilateral CIs lead to improved speech understanding in the presence of competing speech. However, the magnitude and type of advantage for individual patients vary greatly. Although most patients benefit from the head shadow effect as measured with an improved signal-to-noise ratio at which speech is understood, the size of effect in patients varies from small (1–2 dB improvement) to large (>10 dB improvement). In addition, a relatively small number of bilateral CI users show benefits that require binaural processing, such as the "squelch effect" and binaural summation. Both effects lead to better speech understanding with two ears than with a single ear.
In the case of squelch, improvement is seen even if the "added" ear has a poor signal-to-noise ratio. For example, if a person listens to speech from the front and the noise is on the left, one might anticipate that listening monaurally with the right ear alone might be best. However, if squelch occurs, the addition of the left ear creates binaural integration such that the auditory system can suppress the noise better binaurally than monaurally (van Hoesel, 2004; Litovsky et al., 2006a; 2009).
Reduced benefits may be due to hardware and software limitations. Bilateral CI users essentially are fitted with two separate monaural CIs that are fitted separately and lack the time-dependent synchronization of normal-hearing listeners. Furthermore, the incoming signal is filtered into numerous frequency bands; the envelope of the signal is extracted from the output of each band and is used to set stimulation levels. However, processors discard low-frequency (fine-structure) information used by the binaural system. Binaural benefits also may be reduced because many CI users experience auditory deprivation prior to being activated, with the possibility of auditory neuron degeneration.
To clarify the importance of the loss of binaural synchrony, we conducted studies that bypassed clinical processors and introduced fine-structure cues. To achieve this condition, we used research processors (not commercially available) that were engineered to provide precise binaural cues directly with precise synchronization to pairs of electrodes in the right and left ears. This experimental approach enabled us to measure the sensitivity of bilateral CI users to important auditory cues that we know listeners with normal hearing rely on. We discovered that the age at which adults become deaf is a predictor of how well they can utilize binaural cues that rely on inter-aural time differences in localization. People with prelingual deafness are unable to use these cues. However, prelingually deaf adults, like postlingually deaf adults, are able to use inter-aural differences in signal intensity level at the two ears (Litovsky et al., 2010). Level cues are robustly preserved in today's clinical speech processors; adults with prelingual deafness, therefore, also can demonstrate benefits from bilateral CIs on some functional measures. With further research we hope to understand better the conditions under which binaural hearing can be restored to bilateral CI users.
Bilateral CIs in Children
Despite hardware and software limitations, a growing number of children worldwide have received bilateral CIs. These children differ from their adult counterparts; unlike the majority of adults who had been exposed to acoustic hearing early in life, most of these children are deaf from birth or from a very young age. A number of factors have motivated clinicians and parents to provide bilateral CIs to
- In absence of an alternative technology such as hair cell regenerationin the near future, bican be provided so that the auditory pathways can
be stimulated; animal studies suggest that unstimulated auditory nerves lead to degeneration of neural connections in the auditory pathway.
- In the event of a device failure, a child with bilateral CIs may be less likely to be deprived of sound for a period of time, which may be traumatic or isolating for an individual accustomed to auditory input.
- By implanting both ears, the "best-functioning" ear will be stimulated.
Bilateral CIs have the potential to provide children with improved ability to function in complex listening environments in daily life, such as classrooms, playgrounds, and sports environments. Safety also may be a concern when there is a need to be aware of moving objects, such as traffic when crossing a busy street.
Aspects of these real-world complex listening environments have been simulated in our laboratory studies (Litovsky et al., 2004, 2006a, 2006b; Grieco-Calub & Litovsky, 2010; Godar & Litovsky, 2010) in a soundproof booth by positioning arrays of loudspeakers at various locations in the horizontal plane and engaging children in "listening game" approaches (looking at and pointing to a signal or using verbal-response computer interactive games).
In one useful paradigm, children discriminate source locations to the right vs. left, and we evaluate the smallest angle that can be reliably discriminated (the minimum audible angle, or MAA). Emma was tested at two months, one year, and two years after the activation of her second CI. Given a starting point of not understanding the concept of spatial hearing, she experienced remarkable improvement in a two-year period, eventually attaining an MAA of 20 degrees. This level is significantly higher than the MAA of 2–5 degrees attained by children with normal hearing and also is significantly better than the MAA of 50 degrees attained when listening through her first-implanted ear alone. The bilateral benefit measured with the MAA in a group of 18 children ages 5–12 is shown in Figure C [PDF] (see also Litovsky et al., 2006b).
In a second study, children were tested prior to transitioning from single- to bilateral-CI use (see "Baseline" data set, Figure 1 [PDF]). Some of the children used a hearing aid in the other ear, although the hearing aid did not seem to lead to improved performance on this measure compared with the CI-alone users. Re-testing was done at three months and one year following activation of the second CI. The MAA is substantially higher, and more variable, at baseline; smaller and less variable with three months of bilateral experience; and significantly smaller and least variable at one year. The MAA thresholds for the first CI averaged 50 degrees, significantly worse than the bilateral condition. These results suggest that, despite receiving little or no hearing in one ear for a number of years, the children were able to perform a basic spatial hearing task with one year of bilateral hearing experience (Godar & Litovsky, 2010).
It is important to remember that bilateral CI candidacy is still being established, and also to consider the group of children who have residual hearing in the non-implanted ear. The usability of residual hearing and its functional consequences need to be well understood so that if children make the transition from bimodal (CI plus hearing aid in the other ear) to bilateral, they are likely to experience benefits following the transition.
In a third study, spatial hearing skills were tested using more complex sound localization tasks, in which children identified where sounds were located from an array of up to 15 loudspeakers positioned in the horizontal plane. Children who were sequentially implanted with a separation of several years between the activation in the first- and second-implanted ears appeared to be capable of developing this spatial hearing skill. Figure 2 [PDF] shows data from a recent study in which the best-performing children with bilateral CIs (panel B) with low RMS errors fall within the range of errors seen in children with normal hearing (panel A). The large variability in performance among these children suggests that the emergence of sound localization skills in this population likely depends on many factors that are yet to be fully understood. Preliminary data from our lab suggest that factors such as age of implantation and chronological age at the time of testing should be investigated.
As clinical criteria change, children are implanted bilaterally at younger ages at many clinics. The notion that "earlier is best," which took root in the pediatric cochlear implantation literature more than a decade ago, is now being embraced by many clinicians to initiate early activation of both ears.
At the 2009 meeting of the Association for Research in Otolaryngology we reported data from a group of 2.5-year-olds who received their first CI by 1 year of age; 26 of these 34 children had received a second CI by age 1.5–2 years, and a control group of eight children were unilateral CI users. The results, which are not yet published, suggest that none of the unilateral CI users could localize sounds to the right or left, even at large angle separations. Of the 26 bilateral CI users, five also were unable to perform this task. An additional 14 were able to recognize sounds to the right or left, with a variable level of performance. The best performers were seven children with MAA thresholds similar to those measured in 2.5-year-old children with normal hearing; five of the seven had more than a year of bilateral experience, suggesting that early bilateral stimulation may be an important factor in the acquisition of spatial hearing skills. However, factors that account for the variability and lack of spatial hearing in some of these children remain unclear. More research must be conducted prior to determining whether this finding can be generalized to the greater population of young CI candidates.
Quality of Life Considerations
Objective behavioral and perceptual outcomes in adults and children with bilateral CIs generally do not meet the same level of performance seen in listeners with normal hearing. However, subjective reports gathered from questionnaire data about improvements in the quality of life are noteworthy (e.g., Litovsky et al., 2006a; Noble et al., 2008; Summerfield et al., 2006). Adult patients report that they prefer to wear both devices regularly, they feel more "balanced" in their perception of sound, their conversations are more enjoyable and less strenuous, their ease of listening is enhanced, and their incidental listening and learning occur more frequently.
Although no standardized rehabilitative programs exist for training bilateral CI users on spatial hearing tasks, we believe that such programs should be implemented in the clinic and at home. Anecdotal evidence suggests that listeners who struggle with spatial hearing can discover strategies to improve their listening skills if they engage in situations that provide active feedback.
Acknowledgment: This work was supported primarily by grants from the National Institute on Deafness and Other Communication Disorders at the National Institutes of Health. Travels costs for some children were funded by Advanced Bionics Corp., Cochlear LTD, and Med-El Corp.