Focus Detection Using Spatial Release From Masking

Abstract

Individuals are often subjected to environments where multiple conversations occur simultaneously. In these situations, most hearing-abled individuals are able to focus on the auditory stimulus of their choice by filtering out other present auditory stimuli. This ability is also referred to as ‘The Cocktail Party Effect’. Unfortunately, this ability is not yet applicable for people who use assistive listening devices or digital communications devices to communicate with more than one individual [1]. In this study, Spatial Release from Masking techniques are used within the context of its influence on Speech Intelligibility. A Brain-Computer Interface (BCI) system was used to take electroencephalogram (EEG) signals, through noninvasive methods, for machine learning classification training. The goal of using EEG signals to train a machine learning classifier is to find a model that can accurately predict if a subject is listening to a particular auditory stimulus in the presence of multiple auditory stimuli. A similar study has been conducted before but without the use of machine learning for data processing [2].