BACKGROUND: Some noninvasive brain-computer interface (BCI) systems are currently available for locked-in syndrome (LIS) but none have incorporated a statistical language model during text generation. OBJECTIVE: To begin to address the communication needs of individuals with LIS using a noninvasive BCI that involves rapid serial visual presentation (RSVP) of symbols and a unique classifier with electroencephalography (EEG) and language model fusion. METHODS: The RSVP Keyboard was developed with several unique features. Individual letters are presented at 2.5 per second. Computer classification of letters as targets or nontargets based on EEG is performed using machine learning that incorporates a language model for letter prediction via Bayesian fusion enabling targets to be presented only 1 to 4 times. Nine participants with LIS and 9 healthy controls were enrolled. After screening, subjects first calibrated the system, and then completed a series of balanced word generation mastery tasks that were designed with 5 incremental levels of difficulty, which increased by selecting phrases for which the utility of the language model decreased naturally. RESULTS: Six participants with LIS and 9 controls completed the experiment. All LIS participants successfully mastered spelling at level 1 and one subject achieved level 5. Six of 9 control participants achieved level 5. CONCLUSIONS: Individuals who have incomplete LIS may benefit from an EEG-based BCI system, which relies on EEG classification and a statistical language model. Steps to further improve the system are discussed.
BACKGROUND: Some noninvasive brain-computer interface (BCI) systems are currently available for locked-in syndrome (LIS) but none have incorporated a statistical language model during text generation. OBJECTIVE: To begin to address the communication needs of individuals with LIS using a noninvasive BCI that involves rapid serial visual presentation (RSVP) of symbols and a unique classifier with electroencephalography (EEG) and language model fusion. METHODS: The RSVP Keyboard was developed with several unique features. Individual letters are presented at 2.5 per second. Computer classification of letters as targets or nontargets based on EEG is performed using machine learning that incorporates a language model for letter prediction via Bayesian fusion enabling targets to be presented only 1 to 4 times. Nine participants with LIS and 9 healthy controls were enrolled. After screening, subjects first calibrated the system, and then completed a series of balanced word generation mastery tasks that were designed with 5 incremental levels of difficulty, which increased by selecting phrases for which the utility of the language model decreased naturally. RESULTS: Six participants with LIS and 9 controls completed the experiment. All LIS participants successfully mastered spelling at level 1 and one subject achieved level 5. Six of 9 control participants achieved level 5. CONCLUSIONS: Individuals who have incomplete LIS may benefit from an EEG-based BCI system, which relies on EEG classification and a statistical language model. Steps to further improve the system are discussed.
Entities:
Keywords:
brain–computer interface; electroencephalography (EEG); language model; locked-in syndrome; neuroengineering
Authors: D B Ryan; G E Frye; G Townsend; D R Berry; S Mesa-G; N A Gates; E W Sellers Journal: Int J Hum Comput Interact Date: 2011-01-01 Impact factor: 3.353
Authors: Jonathan S Brumberg; Kevin M Pitt; Alana Mantie-Kozlowski; Jeremy D Burnison Journal: Am J Speech Lang Pathol Date: 2018-02-06 Impact factor: 2.408
Authors: Umut Orhan; Hooman Nezamfar; Murat Akcakaya; Deniz Erdogmus; Matt Higger; Mohammad Moghadamfalahi; Andrew Fowler; Brian Roark; Barry Oken; Melanie Fried-Oken Journal: Brain Comput Interfaces (Abingdon) Date: 2016-12-05