Development and evaluation of binaural and transcranial bone conduction stimulation on anatomic head simulator

Abstract

Previous research has demonstrated that patients with bone conduction hearing loss can benefit from binaural fitting using bone conduction hearing devices [1]. However, certain limitations persist due to the complex vibrational patterns that arise from multiple transmission pathways within the complex skull. Notably, transcranial attenuation (TA) below 1 kHz has been reported to be approximately 0 dB SPL, with higher attenuation towards higher frequencies, allowing sound energy to propagate across the skull and stimulate both cochleae simultaneously [2]. To investigate a more realistic perception of different sound scenarios with bone conduction hearing aids, a specialized head simulator was developed by Cochlear Bone Anchored Solutions AB, modeled after a human skull and equipped with three-dimensional accelerometers positioned at the cochlear locations. This Master’s thesis aimed to enable real-time audio playback from bone conduction hearing aids and simulate binaural fittings. To achieve this, digital signal processing was applied, a vector projection of the axis onto the excitation force was performed, and filters were designed to align the output with the air conduction threshold. Finally, a listening experiment involving 37 normal-hearing participants was conducted to evaluate the simulator’s performance. Speech reception thresholds (SRTs) in noise were measured using the standardized Swedish Matrix (Hagerman) test, and binaural intelligibility level differences (BILD) were calculated to assess participants’ ability to utilize spatial cues during headphone-based listening. Four test conditions were evaluated, involving co-located sources as well as configurations with speech presented from the front and noise positions at 270° azimuth. Measurements were conducted with both one and two BAHIs attached. Initial measurements confirmed that transcranial attenuation (TA) values are consistent with literature findings, showing minimal attenuation below 1 kHz and increasing at higher frequencies. The resulting speech reception threshold (SRT) scores revealed a systematic shift of approximately 3 dB toward better signal-to-noise ratios (SNRs) compared to previous studies. An ILD and BILD of approximately 3.6 dB SPL confirmed the simulator’s ability to reliably replicate binaural hearing.