Evaluation Of Joint Auditory Attention Decoding And Adaptive Binaural Beamforming Approach For Hearing Devices With Attention Switching

Beamforming is a common technique used to improve speech intelligibility and listening comfort of hearing aids users in a noisy environment. Traditional hearing aids beamforming algorithms require the a priori knowledge of the auditory attention of the listener, which may not be available in real applications. Recent advances in electroencephalography (EEG) offer a potential non-invasive solution to this problem. The listener’s auditory attention is derived from the EEG signals through auditory attention decoding algorithms and can be used as an input to the beamforming algorithms. Recently, a joint auditory attention decoding and adaptive beamforming algorithm framework by correlating the envelope of beamforming output and the EEG signal was proposed to improve the beamformer’s robustness against attention decoding error. Consistent performance improvement was demonstrated on an EEG database recorded on listeners with fixed attention. In this study, we present the evaluation results of this joint formulation on a newly collected EEG dataset collected on subjects with dynamic attention switch. We demonstrate not only the joint framework’s performance improvement against decoding errors, but also its ability to capture listener’s dynamic attention switch.