A Review of Factors and Display Characteristics Influencing P300 Brain-Computer Interface Performance for AAC.

Authors

Kevin Pitt, University of Nebraska–LincolnFollow

Document Type

Article

Date of this Version

2020

Citation

Pitt, K. (2020). A Review of Factors and Display Characteristics Influencing P300 Brain-Computer Interface Performance for AAC. Accepted for presentation at the 2020 International Society for Augmentative and Alternative Communication (ISAAC) Conference, Cancun, Mexico. (convention cancelled). doi: 10.32873/unl.dc.oth.002

https://doi.org/10.32873/unl.dc.oth.002

Comments

Copyright 2020 Kevin Pitt

Abstract

For individuals with amyotrophic lateral sclerosis (ALS), neurodegeneration can lead to a state of near-total paralysis, with intact cognition and sensation (Fried-Oken. et al., 2013). This loss of voluntary speech and motor system control (including oculomotor) can cause individuals with ALS to lose a reliable form of communication. Brain-computer interfaces for AAC (BCI-AAC) can provide a non-invasive link between an individual’s neurologic activity and a communicational device via electroencephalography (EEG), circumventing the necessity of reliable motor control for AAC access. Different BCI techniques may afford communication access through a range of techniques such as the P300 (including rapid serial visual presentation (RSVP) of items), steady state evoked potential, and motor imagery (see Brumberg, Pitt, Mantie-Kozlowski & Burnison, 2018; and Rezeika et al., 2018, for a full review of BCI techniques).

In the past decade there has been an increased focus into the application of BCI for AAC, with P300-based BCI-AAC devices becoming commercially available (i.e., the P300 Intendix Speller from g.tec medical engineering). However, there are currently limited guidelines regarding how to effectively feature match an individual to an AAC-BCI technique and display (e.g., Pitt, Brumberg & Pitt., 2019). Similar to existing AAC techniques, the best BCI match for an individual may vary between different types of BCI devices (e.g., P300 versus motor imagery), due to factors including; the highly contrasting user interfaces, and neural signals upon which the BCI functions (see Pitt & Brumberg, 2018 for review). Therefore, further research is needed to identify factors that may be of important consideration in BCI assessment, especially for visually based P300-based BCI techniques such as the Intendix speller, which will likely to be the first BCIs to enter clinical practice.