For humans, the ability to locate the direction of the sound source reflects central auditory function as signals gathered by each ear are integrated in the central nervous system beginning at the level of the brainstem. With binaural auditory input, listeners can understand speech at lower (i.e., poorer) signal-to-noise ratios as compared to monaural input (e.g., Cosentino et al., 2014; Garadat et al., 2009; Hawley et al., 2004; Tollin & Yin, 2009), particularly for spatially separated speech and noise. This binaural benefit is attributed to accessibility of binaural cues including interaural time differences (ITDs) and interaural level differences (ILDs) which collectively contribute to spatial release from masking, binaural unmasking of speech (also referred to as squelch), as well as head shadow or “better-ear listening”. While a listener can benefit from head shadow with monaural auditory input—provided the noise is originating from the side of the poor ear—symmetrical interaural head shadow yields significantly better speech understanding irrespective of speech and noise source location (e.g., Gifford et al., 2018).

As indicated above, sounds originating in space typically arrive at the two ears at different times and at different levels. ITDs result from sound sources travelling different distances to each ear and are first resolved in the brainstem—primarily in the medial superior olivary complex—via peripheral neural phase locking. ILDs are primarily resolved in the lateral superior olivary complex and originate from differences in the level of sound at each ear resulting from the head acting as a physical barrier (i.e. head shadow or better-ear listening). In the low-frequency region, sound localization relies more on ITD sensitivity, and in the high-frequency region, sound localization ability relies more on ILD cues (e.g., Dietz et al., 2008; Macaulay et al., 2010; Stevens & Newman, 1936). Speech and suprasegmental perception (e.g., intonation, stress, rhythm, emotion) depends on resolution of fundamental frequency (F0), formants, and spectrotemporal formant transitions and their resultant harmonics which tend to be concentrated in the lower frequencies (< 1500 Hz); as such, periodicity and temporal fine structure—including fine structure ITDs—are critical for auditory tasks in listening scenarios including source segregation (e.g., Schwartz et al., 2012), binaural unmasking of speech (Culling et al., 2004; Gallun et al., 2005), and horizontal plane localization (Macpherson & Middlebrooks, 2002; Wightman & Kistler, 1992).

Electric and acoustic stimulation (EAS) is an auditory intervention combining cochlear implant (CI) and hearing aid (HA) technology in the implanted ear(s) to optimize broadband auditory access. Though EAS includes two modes of auditory stimulation, the term “bimodal hearing” is a descriptor typically reserved for individuals combining a CI and contralateral HA. That is, bimodal listeners do not have preserved acoustic hearing to be aided in the implanted ear.

EAS listening offers high-frequency access from the CI(s) and binaural acoustic hearing in the low-to-mid frequency range via use of a HA integrated in the CI sound processor. Cochlear implantation resulting in EAS listening is an increasingly prevalent intervention for individuals with precipitously sloping high-frequency hearing losses for whom conventional amplification and frequency lowering technologies provide marginal benefit (e.g., Gifford et al., 2007; Salorio-Corbetto et al., 2017b, 2019) likely due to increased prevalence of cochlear dead regions with this audiometric configuration (e.g. Hornsby & Dundas, 2009; Moore, 2001; Salorio-Corbetto et al., 2017a; Vinay & Moore, 2007). Recent research indicates that approximately 20 to 25% of adults seen for preoperative CI evaluation meet audiometric candidacy for EAS or hybrid CI systems (Holder et al., 2018; Patro et al., 2022). Individuals using EAS are most commonly implanted unilaterally with the goal of preserving low-frequency acoustic hearing in the implanted ear. However, there is also growing population of bilateral CI recipients with binaural acoustic hearing preservation using bilateral EAS technology (e.g., Dorman et al., 2013; Gifford et al., 2015; Roberts et al., 2021).

Current CI signal processing strategies largely discard temporal fine structure—rapid spectrotemporal fluctuations in ongoing speech imposed by the articulators (e.g., lips, tongue)—in favor of envelope derivation and extraction; however, EAS users still have access to low-frequency temporal fine structure via binaural acoustic audibility afforded by cochlear implantation with hearing preservation. Several studies have reported significant EAS benefit as compared to the bimodal hearing condition (CI plus contralateral HA) for speech recognition in multi-source noise (Dunn, Perreau, et al., 2010; Gifford et al., 2017; Gifford et al., 2010; Gifford et al., 2013; Gifford et al., 2022), reverberation (Gifford et al., 2013), horizontal-plane localization (Dunn, Noble, et al., 2010; Gifford et al., 2014; Plant & Babic, 2016), as well as subjective reports of listening difficulty (Gifford et al., 2017; Gifford et al., 2022). However, much like other CI configurations, EAS yields outcomes that are highly variable across patients and this variability is not explained by the underlying audiogram (e.g., Gifford et al., 2013; Gifford et al., 2014). Instead, research has shown that EAS benefit is significantly correlated with binaural cue (ITD and ILD) sensitivity in the acoustic hearing domain with the most robust effect observed for fine structure ITD sensitivity (Gifford et al., 2013; Gifford et al., 2014; Gifford & Stecker, 2020). Thus, it is likely the case that individuals with better ITD sensitivity in the acoustic hearing domain will benefit more from EAS.

ITD sensitivity is typically measured using behavioral methods that are time-consuming and require listener training, which limits clinical feasibility as well as the applicability of this measure to the pediatric population. Given the relationship between ITD sensitivity and EAS benefit, there is a clinical need for measures of ITD sensitivity that are quick and easy to obtain. Specifically, identification of an objective measure of fine structure ITD sensitivity could prove useful for clinicians to identify individuals who stand to benefit most from EAS technology and provide data-driven guidance for fitting of EAS technology; the latter point holds clinical significance given recent reports that only 43 to 50% of adult CI recipients with aidable acoustic hearing in the implanted ear are routinely fitted with EAS systems (Perkins et al., 2021; Spitzer et al., 2021). In other words, 50 to 67% of patients who have aidable acoustic hearing in the implanted ear(s) are not fitted with EAS technology despite all ear-level CI processors having integrated HA circuitry. Though there are various reasons for poor EAS utilization, with a split between patient and clinician factors. Patient-related factors include patient preference for an off-the-ear processor which does not have a HA attachment, difficulty maintaining integrated HA functionality (magnetic headpiece attracts ferrous material in the HA attachment which can damage the HA receiver), and patient refusal of earmold for EAS fitting (Spitzer et al., 2020; Perkins et al., 2021). Clinician-related factors include lack of experience fitting EAS, lack of data-driven guidelines for EAS fittings (e.g., HA prescriptive fitting formula, acoustic and electric overlap, CI frequency allocation, etc.) and clinicians being unsure whether the patient will derive benefit. Given the potential for significant benefit afforded by EAS, variability in EAS benefit, and lack of a reliable relationship between EAS benefit and the underlying audiogram, identification of an objective novel measure could be leveraged to improve our predictions of EAS outcomes and even possibly guide electrode selection for patients with astute ITD sensitivity measured preoperatively to maximize acoustic hearing preservation.

Previous research has demonstrated the efficacy of using interaural phase differences (IPDs)—equivalent to ITDs for steady periodic tones—to identify binaural cue sensitivity with cortical auditory evoked potentials (Papesh et al., 2017; Ross et al., 2007). Specifically, researchers have used an acoustic change complex (ACC) paradigm applying an IPD to stimulus midpoint or close to midpoint (Ozmeral et al., 2016; Papesh et al., 2017; Ross et al., 2007; So & Smith, 2021). In these studies, cortical responses were collected via electroencephalography (EEG) (e.g., Ozmeral et al., 2016; Papesh et al., 2017; So & Smith, 2021) or magnetoencephalography (MEG) (e.g., Ross et al., 2007) while participants were exposed to binaural amplitude modulated stimuli with imposed IPD. Previous research has observed robust responses to stimulus onset and offset known as the onset and offset response or the N1-P2 complex. The N1-P2 onset response (change from silence to sound) is an obligatory, event-related potential with N1 and P2 latencies occurring approximately 100 and 200 ms following stimulus onset, respectively. The offset response (change from sound to silence) is also an obligatory, event-related potential observed following sound cessation with similar amplitude and latency as the onset-elicited N1-P2 complex. Both N1 and P2 components have multiple cortical generators in primary and secondary auditory cortices. The ACC also reflects the brain's ability to detect and process a change in one's acoustic environment, which in this case, would be imposition of an IPD at stimulus midpoint. The previous research indicates that as the ITD/IPD increases, the strength of the ACC response also increases.

Past studies have used tonal (Papesh et al., 2017; Ross et al., 2007; So & Smith, 2021) and noise band carriers (Ozmeral et al., 2016) for these tasks in the range of 375 to 1500 Hz. The majority of tested frequency regions hold limited application to EAS patients who generally have the best acoustic hearing sensitivity below 500 Hz (e.g., Adunka et al., 2013; Lenarz et al., 2020; Roland et al., 2018; Skarzynski et al., 2012). Studies such as Ross et al. (2007) and Papesh et al. (2017) both used a 180° IPD focusing on carrier frequencies ≥ 500 Hz1. Ozmeral et al. (2016) used 500 – 750 Hz band-pass noise with 250 or 500 µs ITD. Band-pass noise is difficult to calculate accurate IPD when a fixed ITD is used. This is because each different frequency component (from 500 Hz – 750 Hz) will result in a different IPD, yet the ITD value was close to 90° or 180° IPD with the 1000-Hz carrier. So and Smith (2021) used 500 Hz pure tone as carrier and 83-Hz modulation. They also used multiple IPDs including 22.5°, 45°, 67.5°, and 90°.

It is worth noting that Ross (2018) demonstrated cortical auditory sensitivity to envelope ITDs for a 250-Hz signal using MEG functional neuroimaging. They imposed ITDs by manipulating interaural phase of the 40-Hz modulation. With relevance to our targeted population of EAS candidates, envelope ITD sensitivity holds limited application given that resolution of envelope ITDs has been demonstrated for listeners with normal acoustic and electric hearing using both behavioral (e.g., Bernstein and Trahoitis, 1994, 1996; Laback et al., 2011, 2015) and objective measures (e.g., Ross, 2018). Having functionally aidable binaural low-frequency acoustic hearing, EAS patients are uniquely suited for resolution of fine structure ITDs, which are poorly transmitted via CIs due both to envelope-based processing and channel interaction. Furthermore, envelope ITD resolution is inversely correlated with modulation rate with decreasing sensitivity and range of lateralization for rates >200 Hz (e.g., Bernstein and Trahoitis, 2014; Anderson et al., 2019) and additional limitations imposed by the envelope-extraction stage in CI signal processing including low-frequency filtering, typically ≤ 300 Hz and n-of-m strategies for which the audible spectrum may not be represented or equally represented across ears for each corresponding electrode/frequency region (e.g., Kan et al., 2018). The latter point is especially critical given that envelope ITD sensitivity is influenced by bandwith for narrow- and broad-band carriers (Mayo et al., 2021). Furthermore, fine structure ITDs in the lower frequency range are a robust, heavily weighted cue for localization (e.g., MacPherson and Middlebrooks, 2002). Thus, fine structure ITD resolution will be the focus of this manuscript henceforward.

Using a different approach, Haywood et al. (2015) introduced a direct method to measure sensitivity to interaural time differences, termed the interaural phase modulation following response (IPM-FR). IPM-FR involves presenting an ongoing amplitude-modulated signal and periodically changing the interaural phase difference at the modulation zero crossing point, leading to a steady state response; thus, in this case IPD-mediated ACC and IPM-FR tap into similar auditory mechanisms, including brainstem IPD sensitivity preserved in auditory cortical activity. So and Smith (2021) compared measuring ACC to IPM-FR on three different aspects. They performed scalp topography analysis, compared the time to detect the two responses, and compared the input-output characteristics of the two measurements. They pointed out that the IPM-FR has better time-to-detect performance on smaller IPDs (< 45o) but not larger (> 67.5o) IPDs. Undurraga et al. (2016) investigated participants’ behavioral responses and IPM-FR amplitude in response to a wide range of IPDs (45o - 315o). Although IPM-FR amplitude and behavioral discrimination of IPM and static ITD varied in similar pattern across IPDs, behavioral performance and IPM-FR amplitude were not correlated across all IPDs or for all study participants. Rather, the relationship between IPM-FR amplitude and behavioral IPM/ITD discrimination was strongest for phases ranging from ±67.7 to ±112.6 μs. Similarly, Vercammen et al. (2018) investigated the relationship between participants’ IPM-FR with 492-Hz carrier and behavioral IPD discrimination for young, middle aged, and older listeners both with and without sensorineural hearing loss. Their data indicated a significant moderate correlation between participants’ IPM-FR dynamic range and behavioral IPD thresholds (r = -0.47)—a relationship influenced both by listener age and hearing status. That is, as compared to young participants, middle aged and older individuals had significantly lower IPM-FR amplitudes and poorer IPD discrimination. The presence of hearing loss also significantly affected IPM-FR amplitude, but was not associated with poorer IPD discrimination, likely because the participants with hearing loss had broadband auditory access given their mild sloping to moderate-severe hearing losses.

It also worth mentioning that Borjigin et al. (2022) measured EEG in 42 participants and used intertrial coherence (ITC) to quantify the response to ITDs at 20, 60, 180, and 540 microseconds (equivalent to 3.6°, 10.8°, 32.4°, 97.2° with 500 Hz-carrier). The model they proposed used both EEG data and lapse rate—a parameter measured behaviorally that reflects participant's attention—to predict participants’ behavioral ITD threshold. Although they observed a strong, significant correlation (r = 0.71), their method also required a behavioral component.

Given the relationship between acoustic ITD sensitivity and EAS benefit for localization and speech understanding with spatially separated masker(s), the use of a quick, objective estimate of binaural cue processing could ultimately prove clinically useful. While there is a solid literature base describing efficacy of ACC and IPM-FR for objective quantification of binaural processing, no published studies exploring fine structure ITDs have used low-frequency carriers, such as 250 Hz, as this frequency range holds critical relevance to the EAS population. Prior to commencing experimentation for individuals with steeply sloping hearing losses who meet EAS labeled indications, we sought to describe this relationship in a group of young listeners with normal hearing. Thus, the primary purpose of the current study was to investigate efficacy of the ACC paradigm using lower and higher frequency carriers (250 and 1000 Hz) with 40-Hz sinusoidal amplitude modulation and the following IPDs imposed at stimulus onset: 0°, 45°, 90°, 135°, and 180°. At 250Hz, the corresponding ITDs are: 0, 500, 1000, 1500, and 2000 microseconds. At 250Hz, the corresponding ITDs are: 0, 125, 250, 375, and 500 microseconds. We also measured behavioral ITD thresholds at 250 and 1000 Hz allowing us to further explore the relationship between the objective and behavioral measures of binaural processing in humans, particularly for 250 Hz given its clinical application. Based on referenced studies in this space, our hypotheses were that 1) the ACC-IPD paradigm would yield higher amplitude N1-P2 responses to the IPD with a 250-Hz carrier given the known inverse relationship between degree of neural synchrony and stimulus frequency, 2) there would be a significant correlation between ACC amplitude and IPD, and 3) there would be a significant and strong correlation between participants’ ITD/IPD sensitivity measured using behavioral and objective methods.