Page 22 - Volume 12, Issue 2 - Spring 2012

two-masker speech stimuli were used than when one-masker was used, suggesting that, when the listening environment is more complex, spatial cues play an increasingly more impor- tant role in assisting the listener with source segregation. Follow-up studies have been conducted in which other aspects of the auditory environment have been manipulated, such as the type of masker (Johnstone, 2006; Johnstone and Litovsky, 2006), with evidence to suggest SRM in children is greatest when target-masker similarity is high (i.e., when both the target and masker were produced by a male talker), similar to what has been seen in adults (Bronkhorst, 2000; Brungart et al., 2001). This effect has been attributed to the fact that target-masker similarity renders speech segregation extremely difficult, making spatial cues the most salient cues that listeners can use to segregate target from maskers. By comparison, SRM is smaller with different-sex and non- speech maskers, because cues resulting from differences in the spectra of the signals can be useful for source segregation. A somewhat different variation on the target-masker similar- ity can be introduced with a reversed-speech masker, which contains the same temporal amplitude fluctuations and long- term spectrum as speech. In children, this masker produced SRM that was similar to that with the speech masker from which it was created, even though it carried no linguistic con- tent (Johnstone and Litovsky 2006), a finding that is consis- tent with observations in adults (Hawley et al., 2004). Children provided anecdotal reports that the masker had a novel feature with some resemblance to a person speaking in a foreign language. This may have added to the interference that would have been produced simply by a modulated speech-shaped noise masker that bears similarity in the spec- tral and amplitude-modulation domain, but carries no resemblance to spoken language.
Spatial release from masking in special populations
A growing number of adults and children with hearing impairment have been receiving stimulation in both ears in an effort to provide them with perceptual benefits on audito- ry tasks that are known to rely on having inputs in both ears. For many years, the standard of care in acoustic amplification has been to provide bilateral hearing aids to people with bilateral hearing loss (Litovsky and Madell, 2009). Another population of patients with hearing loss who cannot benefit from amplification is a population of people with severe-to- profound hearing loss. These patients are often candidates for receiving electrical stimulation through cochlear implants (CIs). These devices have been clinically available for the past few decades, and can provide auditory input by electrically stimulating the auditory nerve, bypassing the damaged sen- sory organ of hearing, known as the cochlea.
For many patients, electrical input through a CI is suf- ficient for attainment of speech perception and production within the normal range, allowing aural-verbal communi- cation. The standard of care for many years was considered to be provision of adequate speech perception and language acquisition through the use of a monaural CI. Deciding which ear to implant has been a complex decision, one which has undergone a series of changes throughout the
years. In considering this choice, it is important to note that many patients do not have symmetrical hearing loss in the two ears, thus the ear chosen for implantation has depend- ed on numerous factors, including the etiology of the hear- ing impairment and various clinical considerations. In some cases, patients with residual acoustic hearing in one ear but not the other receive the implant in the “worse” ear to preserve the residual hearing, which can otherwise be destroyed with insertion of the CI. Alternatively, consider a patient with long-term hearing loss in at least one ear; thus, the ear with residual hearing may also be the ear that has had less auditory deprivation and thus responds best to stimulation with the CI.
Either way, many patients or parents of children who are eligible for CIs have reported that monaural hearing can be challenging, with poor speech comprehension in complex noisy environments, and poor sound localization. Clinical care has undergone a transformation, whereby many patients, or parents of young patients, are electing bilateral CIs (one in each ear), with the goal of providing an improved ability to segregate speech from background noise and to localize sounds (e.g., van Hoesel, 2004, 2011; Litovsky et al., 2009; Litovsky, 2011). Sometimes the surgical procedures are simultaneous, and other times they are sequential, with months or years between procedures. Research to date has shown that the vast majority of adults who became deaf post- lingually and were implanted bilaterally show significant improvement on the desired abilities when their perform- ance is compared in bilateral vs. monaural conditions (van Hoesel and Tyler, 2003; Litovsky et al., 2009). Typically, adults who are post-lingually deaf will have had exposure to acoustic hearing for many years prior to becoming deaf, and the activation of bilateral CIs most likely re-activates some aspects of their previously established spatial-hearing abili- ties. In children the issues are quite different, because most of them are congenitally deaf and will not have been exposed to acoustic input prior to becoming deaf.
Referring to the above-mentioned spatial cues (monaur- al head shadow and binaural), studies have been conducted (Litovsky et al., 2006a, 2009) in which SRTs or percent cor- rect measures are obtained with various spatial distributions of target and maskers. Rather than measuring SRM per se, the focus in many studies has been to measure bilateral benefit, i.e., improvement in speech understanding when patients are using bilateral vs. monaural CIs. As for studies described above with normal-hearing listeners, studies with bilateral CI users have compared performance for conditions with target speech in front, and masker(s) either were co-located with the target or spatially separated. Because one of the two CIs can be turned off, rendering the patient monaurally deaf, the spatial configuration of the target and maskers must be con- sidered along with the active/inactive ear. A schematic dia- gram of the three different masker configurations (right, left, or front) combined with each of the three listening modes (right CI only, left CI only, or bilateral) is shown in Fig. 3.
Results from the vast majority of patients to date (Schleich et al., 2004; Litovsky et al., 2006a, 2009) suggest that the primary benefit from bilateral CIs can be attributed to the
Spatial Release from Masking 21