Title: Improving ASSR detection using Independent Component Analysis
Authors: Van Dun, Bram
Wouters, Jan
Moonen, Marc #
Issue Date: 2005
Conference: International Evoked Response Audiometry Study Group (IERASG) edition:19th Biennal location:Havana, Cuba date:Jun 12 16/06/2005
Abstract: The technique of frequency specific hearing threshold assessment using Auditory Steady State Responses (ASSR) makes use of single or multi channel EEG recordings obtained from different scalp electrodes located on the skull. These recordings unfortunately go with excessive measurement duration. Sessions of 90 minutes are no exception, which is considered unpractical for widespread clinical application. This is inherent in the nature of the responses, which are small compared to the noise. Extended measurement times are needed to obtain an acceptable signal-to-noise ratio (SNR). To reduce this problem, several signal processing techniques were investigated, from which Independent Component Analysis (ICA) gave very good results. In general, the technique supposes the recorded channels are a linear combination of underlying unknown generators that are assumed to be statistically independent. If a steady state response to an auditory stimulus is present in more than one channel, the algorithm is able to separate this response from noise because statistically both signals (response and noise) are totally different. Results show a recording time reduction of 35% is possible for single channel data (partitioned in several channels to make ICA input possible). Especially for hearing impaired patients this improvement is significant. Supported by these findings and the optimal applicability of ICA to multiple simultaneous recordings, a multi channel ASSR setup has been developed and measurements carried out. Results of ICA on these multi channel recordings will also be presented.
Publication status: published
KU Leuven publication type: IMa
Appears in Collections:Research Group Experimental Oto-rhino-laryngology
ESAT - STADIUS, Stadius Centre for Dynamical Systems, Signal Processing and Data Analytics
# (joint) last author

Files in This Item:

There are no files associated with this item.


All items in Lirias are protected by copyright, with all rights reserved.