PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Auditory ERP speller applications as a tool for BCI end-users
Johannes Höhne, Martijn Schreuder and Michael Tangermann
In: SAN 2011, Thessaloniki, Greece(2011).

Abstract

With Brain-Computer Interfaces (BCI), it is possible to establish a communication pathway that solely relies on neural activity. Since visual BCI paradigms are not suitable for users with vision impairments (e.g. late-stage ALS patients), auditory speller paradigms have recently been investigated. We propose two auditory ERP speller paradigms that were recently developed at the BBCI group in Berlin. Both paradigms utilize spatial auditory stimuli and were successfully tested in online studies with healthy participants. The AMUSE paradigm (Schreuder et al., 2010) consists of a ring with 6 audio speakers around the subject which produces stimuli in a pseudo-random order. In the PASS2D paradigm (Höhne et al., 2010), nine auditory stimuli are presented using standard headphones in a pseudo-random order. Stimuli vary in pitch (high/medium/low) and direction (left/middle/right). For both paradigms it was found that focusing attention to any of the stimuli leads to reliable elicitations of class-discriminative N200 and P300 signals in the EEG, that are strong enough to drive a speller application. For the AMUSE paradigm, the spelling procedure is based on six control signals and follows an adapted version of the Hex-o-Spell text entry system. In the PASS2D paradigm, the spelling procedure is based on nine control signals, being supported by a predictive text system that is intuitive to understand and well known from mobile phones. On average, a spelling speed of more than 0.8 char/min and more than 3 bits/min was achieved. Thus, both spellers are competitive paradigms even though they are independent of visual input. These paradigms are promising to provide a new communication channel for BCI end-users such as late-stage ALS patients that are unable to use visual spellers.

EPrint Type:Conference or Workshop Item (Poster)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Brain Computer Interfaces
ID Code:9470
Deposited By:Martijn Schreuder
Deposited On:16 March 2012