Bone conduction (BC) hearing refers to the way vibratory sound waves pass through the skull to the cochlea to produce an auditory perception (Stenfelt, 2011; Stenfelt and Goode, 2005a). The transmission characteristics from a BC transducer (BT) to the cochlea are closely related to physiological parameters such as head structure and size, which is of great significance in the application of BC sound reproduction (Pollard et al., 2017).
Previous studies have shown that spatial sound replayed by binaural air conduction (AC) earphones or hearing aids have high inter-aural separation and nearly no problem with cross-talk (Xie, 2013; Zwislocki, 1953). On the contrary, the binaural separation of BC devices is poor leading to problems with cross-talk (Eeg-Olofsson et al., 2011; Nolan and Lyon, 1981; Stenfelt, 2005; Stenfelt and Zeitooni, 2013; Zeitooni et al., 2016). The issue frequently encountered in BC sound reproduction is that the vibration of the BT on either side of the head spreads to both ears. Consequently, the vibration of the left BT is transmitted to the left ear as well as the right ear, and vice versa. Such cross-talk impedes source localization. Specifically, the cross-talk distorts the spatial information of the sound and affect the perception of the spatial location of a sound source (Agterberg et al., 2019).
Transcranial attenuation (TA) is defined as the difference between the ipsilateral and contralateral response for a stimulation at a specific position (Rigato et al., 2019). Stenfelt measured the TA of patients with unilateral deafness at two stimulation positions, and found that the TA depends on the stimulation position and frequency (Stenfelt, 2012). In that study, similar to others, there were a high variability between different participants and between adjacent frequencies of the same participant (Stenfelt, 2012). Rigato et al. analysed the effect of different BT attachments on vibration transmission and TA, and the results showed that the TA of three attachments at the mastoid was similar (Rigato et al., 2019). Eeg-Olofsson et al. as well as Stenfelt and Goode measured the TA for multiple stimulation positions showing that the TA depend on the exact stimulation position with greater TAs for stimulation positions close to the cochlea compared to stimulation positions further away (Eeg-Olofsson et al., 2011; Stenfelt and Goode, 2005b). Studies have shown that the perception of spatial orientation for hearing-impaired people wearing bone-anchored hearing aids (BAHAs) is still worse than that for people with normal hearing (Bosman et al., 2001; Caspers et al., 2022; Iwasaki, 2010; Priwin et al., 2004). Some studies have even shown that people with normal hearing wearing BC devices cannot accurately lateralize sounds coming from right or left directions (Walker et al., 2005). It is therefore of great value to investigate cross-talk cancellation (CTC) with BC sound to improve space perception.
Liao investigated cross-talk cancellation in a dry skull, and designed a cross-talk cancellation system (CCS) by using a fast deconvolution algorithm. The design of that system was not based on a real person's BC transfer function (BCTF), and the experimental results could not be subjectively verified (Liao, 2010). Irwansyah used active control techniques on a BC headphone for estimating a cross-talk compensation filter (Usagawa, 2017). That study was compromised by both the stimulation and background noise when extracting otoacoustic emissions (OAEs) in the ear canal. The problems were still persistent in a subsequent study when a BC microphone was used to record the BC sound (Usagawa, 2019). Mcleod estimated the cross-head sound transfer by a cancellation experiment and used that estimate to cancel the BC cross-talk, but the measurements were reported as time consuming and complex (Mcleod and Culling, 2019, 2020).
As indicated in the previous paragraph, only a few studies have investigated the cross-talk problem of bilateral BC sound reproduction. One important reason for this lack of investigations is the difficulty to accurately measure the BCTF. Due to the complexity of BC sound transmission as well as the transformation from a skull bone vibration to a BC perception (Zhao et al., 2021), it is difficult to conduct live human experiments. At present, the exact transmission properties of BC sound are not clarified.
Because it is not possible to directly measure the BC sound in a real person's cochlea, BC sound transmission have been explored in dry skulls or cadaveric heads (Håkansson et al., 1994; Liao, 2010; Prodanovic and Stenfelt, 2021; Stenfelt and Goode, 2005b). Stenfelt conducted stimulation at multiple positions on the head of cadavers (Stenfelt and Goode, 2005b) and by measurement of the cochlear promontory vibration, it was found that the best stimulation point was located at the mastoid and the worst was at the mid-line of the skull. Several studies have investigated the effects of different positions of BTs (Frank, 1982; McBride et al., 2008; Osafo-Yeboah et al., 2006; Qin and Usagawa, 2017; Studebaker, 1962), and found that the closer to the cochlea, the greater the BC response.
When the inner ear is stimulated by an external sound, the active processes in the cochlea produces a sound termed OAE, and some investigations have used OAEs to explore the physiological characteristics of the inner ear with BC stimulation (Purcell et al., 2003; Qin et al., 2014). The OAEs can be extracted using different methods (Han et al., 2015; Ye et al., 2004). Deng used two swept tones as stimulus and a dynamic tracking filter to extract distortion product otoacoustic emissions (DPOAEs) for an entire frequency range with high efficiency (Deng et al., 2013). Chen used a swept-tone to extract stimulus frequency otoacoustic emissions (SFOAEs) (Chen et al., 2013; Wang et al., 2018), which was efficient as the response for an entire frequency range could be obtained. Even if the above studies used AC stimulation to extract OAEs, OAEs can been used for objective investigations of human BCTF (Purcell et al., 2003; Usagawa, 2017). For example, Purcell used discrete frequency DPOAEs and an auditory peripheral model to estimate the BC cochlear response, and thereby compute the BCTF (Purcell et al., 2003).
In this study, we used Chen's method of swept-tone SFOAEs but adopted for BC stimulation (Chen et al., 2013) where the swept-tone SFOAEs was evoked by BC. The relative relationship between SFOAEs and swept-tone stimulus was defined as the BC response function (BCRF), which is regarded as an objective characterization of the BCTF.
This aim of the current study is to investigate the BC transmission properties in human heads with BC stimulation at the mastoid using swept-tone SFOAEs to estimate the BCRF. In addition, a cross-talk cancellation filter was designed according to the individually estimated BCRF to improve sound source localization.
留言 (0)