Brain-informed speech separation (BISS) for enhancement of target speaker in multitalker speech perception

• Published in: NeuroImage (Orlando, Fla.) Vol. 223; p. 117282
• Main Authors: Ceolini, Enea, Hjortkjær, Jens, Wong, Daniel D.E., O’Sullivan, James, Raghavan, Vinay S., Herrero, Jose, Mehta, Ashesh D., Liu, Shih-Chii, Mesgarani, Nima
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.12.2020; Elsevier Limited; Elsevier
• Subjects: Accuracy; Acoustic Stimulation; Acoustics; Adult; Age; Algorithms; Brain - physiology; Cognitive control; Deep Learning; EEG; Electroencephalography; Experiments; Hearing; Hearing aid; Hearing aids; Hearing loss; Hearing Loss - physiopathology; Humans; Middle Aged; Neural networks; Neuro-steered; Signal processing; Signal Processing, Computer-Assisted; Speech; Speech Acoustics; Speech perception; Speech Perception - physiology; Speech separation; Deep learning; Speech separation; Neuro-steered; EEG; Cognitive control; Hearing aid
• ISSN: 1053-8119; 1095-9572; 1095-9572
• DOI: 10.1016/j.neuroimage.2020.117282

•Multi-talker speech perception is challenging for people with hearing loss.•Automatic speech separation cannot help without first identifying the target speaker.•We used the brain signal of listeners to jointly identify and extract target speech.•This method eliminates the need for separating sound sources or knowing their number.•We show the efficacy of this method in both normal and hearing impaired subjects. [Display omitted] Hearing-impaired people often struggle to follow the speech stream of an individual talker in noisy environments. Recent studies show that the brain tracks attended speech and that the attended talker can be decoded from neural data on a single-trial level. This raises the possibility of “neuro-steered” hearing devices in which the brain-decoded intention of a hearing-impaired listener is used to enhance the voice of the attended speaker from a speech separation front-end. So far, methods that use this paradigm have focused on optimizing the brain decoding and the acoustic speech separation independently. In this work, we propose a novel framework called brain-informed speech separation (BISS)11BISS: brain-informed speech separation. in which the information about the attended speech, as decoded from the subject’s brain, is directly used to perform speech separation in the front-end. We present a deep learning model that uses neural data to extract the clean audio signal that a listener is attending to from a multi-talker speech mixture. We show that the framework can be applied successfully to the decoded output from either invasive intracranial electroencephalography (iEEG) or non-invasive electroencephalography (EEG) recordings from hearing-impaired subjects. It also results in improved speech separation, even in scenes with background noise. The generalization capability of the system renders it a perfect candidate for neuro-steered hearing-assistive devices.