Cortical Auditory-Evoked Potentials (CAEPs) in Adults in Response to Filtered Speech Stimuli DOI: 10.3766/jaaa.24.9.5 Lyndal Carter* Harvey Dillon* John Seymour* Mark Seeto* Bram Van Dun*† Abstract Background: Previous studies have demonstrated that cortical auditory-evoked potentials (CAEPs) can be reliably elicited in response to speech stimuli in listeners wearing hearing aids. It is unclear, however, how close to the aided behavioral threshold (i.e., at what behavioral sensation level) a sound must be before a cortical response can reliably be detected. Purpose: The purpose of this study was to systematically examine the relationship between CAEP detec- tion and the audibility of speech sounds (as measured behaviorally), when the listener is wearing a hearing aid fitted to prescriptive targets. A secondary aim was to investigate whether CAEP detection is affected by varying the frequency emphasis of stimuli, so as to simulate variations to the prescribed gain-frequency response of a hearing aid. The results have direct implications for the evaluation of hearing aid fittings in nonresponsive adult clients, and indirect implications for the evaluation of hearing aid fittings in infants. Research Design: Participants wore hearing aids while listening to speech sounds presented in a sound field. Aided thresholds were measured, and cortical responses evoked, under a range of stimulus con- ditions. The presence or absence of CAEPs was determined by an automated statistic. Study Sample: Participants were adults (6 females and 4 males). Participants had sensorineural hearing loss ranging from mild to severe-profound in degree. Data Collection and Analysis: Participants’ own hearing aids were replaced with a test hearing aid, with linear processing, during assessments. Pure-tone thresholds and hearing aid gain measurements were obtained, and a theoretical prediction of speech stimulus audibility for each participant (similar to those used for audibility predictions in infant hearing aid fittings) was calculated. Three speech stimuli, (/m/, /t/, and /g/) were presented aided (monaurally, nontest ear occluded), free field, under three conditions (14 dB/octave, 24 dB/octave, and without filtering), at levels of 40, 50, and 60 dB SPL (measured for the unfiltered condition). Behavioral thresholds were obtained, and CAEP recordings were made using these stimuli. The interaction of hearing loss, presentation levels, and filtering conditions resulted in a range of CAEP test behavioral sensation levels (SLs), from 225 to 140 dB. Results: Statistically significant CAEPs ( p , .05) were obtained for virtually every presentation where the behavioral sensation level was .10 dB, and for only 5% of occasions when the sensation level was negative. In these (“false-positive”) cases, the greatest (negative) sensation level at which a CAEP was judged to be present was 26 dB SL. Conclusions: CAEPs are a sensitive tool for directly evaluating the audibility of speech sounds, at least for adult listeners. CAEP evaluation was found to be more accurate than audibility predictions, based on threshold and hearing aid response measures. *National Acoustic Laboratories, Sydney, Australia; †The Hearing CRC, Melbourne, Australia Lyndal Carter, National Acoustic Laboratories (NAL), Australian Hearing Hub, Level 5, 16 University Avenue, Macquarie University, NSW 2109, Australia; E-mail: lyndal.carter@nal.gov.au Selected results from this article were presented at the Audiology Australia XIX National Conference, Sydney, May 16–19, 2010. The author has no direct conflict of interest. However, the National Acoustic Laboratories (NAL) developed the HEARLab test suite described in the methodology of this paper. Since the time of the experiment described, HEARLab has been commercialized, and NAL receives a royalty for each unit sold. J Am Acad Audiol 24:807–822 (2013) 807