Page 31 - Spring 2015

Figure 3. A, left to right: George Washington transitioning from low resolution to high resolution. B: low-resolution (20 pixels verti- cal) versions of George W. Bush, Bill Clinton, Barack Obama, and Ronald Reagan. Photos were taken from publically available official presidential portraits.
The Two big Problems in the field
As mentioned above, CIs remove a massive amount of acoustic information before transferring signals to the audi- tory system. The field of CIs has many of areas in which it could improve but the two largest are probably (1) the issue of music, pitch, and noise; and (2) the issue of spatial hearing and noise.
Yes, “noise” was mentioned twice. But life is noisy and messy, and understanding speech with CIs does not work well in noise (e.g., Loizou et al., 2009). Because the major reason to get a CI is to provide/regain/restore the ability to understand speech and most of the time we are in noisy situations, then there is the need for CIs to work better in noise.
It is not uncommon to hear anecdotes about people who do not use their CIs because they cannot enjoy music any more. This is true; for most CI users, music is a bust (perhaps more so for those who enjoy classical music and symphonies that rely on good melodies and less so for those who like rap and dance music that rely on rhythm and beats). Because CIs only convey temporal envelope information and replace the carrier with a pulse train (Figure 2), much of the “temporal pitch” (i.e., the pitch that results from the periodicity of the waveform) is obliterated. Some of the “place pitch” (i.e., the pitch that results from the characteristic frequency of the neurons that are excited) remains. Not all temporal pitch is absent because some of it is conveyed in the temporal en- velope (enough to provide the ability to discriminate gen- ders, albeit at a relatively low level compared with acoustic
hearing) (Fu et al., 2004). Because we want CI users to be out in the world socializing, perhaps at a noisy restaurant or cocktail party, the lack of fine structure or periodicity in- formation severely hampers any fun in such situations. Fine structure or periodicity information is a major or perhaps “the” auditory grouping cue (i.e., grouping meaning how our brain organizes an auditory scene into single perceived objects), which is missing for a person with a CI. Without it, the envelopes mush together into a mostly unintelligible mess.
In the absence of temporal fine structure and periodicity grouping cues, grouping cues associated with spatial hear- ing (i.e., cues provided by having two ears that allow you to perceive sounds as coming from different locations) become incredibly important, which is the second major problem in the field. People have two ears because they help us locate sounds in space by calculating timing and level differences of the sound as it reaches the two ears.
In addition, two ears allow us to better understand speech in background noise. An increasing number people are receiv- ing bilateral CIs with the hopes of gaining access to spatial hearing. In normal hearing, when other grouping cues like pitch fail, spatial cues allow for a good auditory scene or- ganization (e.g., Brungart, 2001). Such cues allow listeners to pick out the person they are attending to at a loud and crowded party, even if they are surrounded by a number of people all of the same gender, who generally have the same voice pitch.
Thus, until we restore temporal fine structure and periodic- ity grouping cues (really, the “holy grail” of CI research; if we can fix periodicity and pitch, we can probably fix any aspect of hearing), CI users need to have some other cue to orga- nize auditory scenes. The lowest hanging fruit may be spatial cues. Currently, bilateral CI users can locate sounds in space and better understand speech in noise than unilateral CI us- ers. But the field is far from doing a great job of providing spatial cues.
How might we improve these percepts for CIs? There are limitations in several domains, including biological, surgi- cal, and device related (Litovsky et al., 2012). More broadly though, the field has used an “engineer’s approach” to this problem. The damaged ear is not functioning as designed. To fix the ear, we need to figure out what information to con- vey to the auditory system and then convey it.
| 29