Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Low-latency auditory spatial attention detection based on spectro-spatial features from EEG

Siqi Cai, Pengcheng Sun, Tanja Schultz, Haizhou Li

TL;DR

This work proposes a spectro-spatial feature extraction technique to detect auditory spatial attention (left/right) based on the topographic specificity of alpha power and shows that this neural approach outperforms other competitive models by a large margin in all test cases.

Abstract

Detecting auditory attention based on brain signals enables many everyday applications, and serves as part of the solution to the cocktail party effect in speech processing. Several studies leverage the correlation between brain signals and auditory stimuli to detect the auditory attention of listeners. Recently, studies show that the alpha band (8-13 Hz) EEG signals enable the localization of auditory stimuli. We believe that it is possible to detect auditory spatial attention without the need of auditory stimuli as references. In this work, we use alpha power signals for automatic auditory spatial attention detection. To the best of our knowledge, this is the first attempt to detect spatial attention based on alpha power neural signals. We propose a spectro-spatial feature extraction technique to detect the auditory spatial attention (left/right) based on the topographic specificity of alpha power. Experiments show that the proposed neural approach achieves 81.7% and 94.6% accuracy for 1-second and 10-second decision windows, respectively. Our comparative results show that this neural approach outperforms other competitive models by a large margin in all test cases.

Low-latency auditory spatial attention detection based on spectro-spatial features from EEG

TL;DR

This work proposes a spectro-spatial feature extraction technique to detect auditory spatial attention (left/right) based on the topographic specificity of alpha power and shows that this neural approach outperforms other competitive models by a large margin in all test cases.

Abstract

Detecting auditory attention based on brain signals enables many everyday applications, and serves as part of the solution to the cocktail party effect in speech processing. Several studies leverage the correlation between brain signals and auditory stimuli to detect the auditory attention of listeners. Recently, studies show that the alpha band (8-13 Hz) EEG signals enable the localization of auditory stimuli. We believe that it is possible to detect auditory spatial attention without the need of auditory stimuli as references. In this work, we use alpha power signals for automatic auditory spatial attention detection. To the best of our knowledge, this is the first attempt to detect spatial attention based on alpha power neural signals. We propose a spectro-spatial feature extraction technique to detect the auditory spatial attention (left/right) based on the topographic specificity of alpha power. Experiments show that the proposed neural approach achieves 81.7% and 94.6% accuracy for 1-second and 10-second decision windows, respectively. Our comparative results show that this neural approach outperforms other competitive models by a large margin in all test cases.

Paper Structure

This paper contains 10 sections, 2 figures, 1 table.

Table of Contents

  1. INTRODUCTION
  2. Auditory Spatial Attention Detection
  3. Spectro-spatial Feature (SSF)
  4. Attention Detection with Convolutional Neural Network
  5. Experiments and Results
  6. Experimental Setup
  7. Data Preparation
  8. Experiments with EEG of 64 Channels
  9. Experiments with EEG of 32 Channels
  10. CONCLUSIONS

Figures (2)

  • Figure 1: The proposed convolutional neural network (CNN) with spectro-spatial feature (SSF) for auditory spatial attention detection, that is referred to as SSF-CNN model. The SSF-CNN network is trained to output two values, i.e., 0 and 1, to indicate the spatial location of the attended speaker.
  • Figure 2: Auditory spatial attention detection accuracy of the proposed SSF-CNN model with 64-channel EEG over different decision windows and for all subjects. The subjects are ranked according to the accuracy for the 10-second decision window. The horizontal dotted line shows a reference point of high accuracy at 90%.