Event Title

Classification of Covert Visuospatial Attention Direction to Assist People with No Motor Function

Mentor 1

Roger Smith

Start Date

1-5-2020 12:00 AM

Description

People with minimal to no motor function lack the means necessary to communicate and control their environment. Our literature review revealed that Covert Visuospatial Attention (CVSA) classification, although thoroughly studied, has rarely been used in addressing the needs of these individuals (such as individuals with Locked-in Syndrome). CVSA is the process of attending to a target without overtly looking at it, which can be classified through noninvasive Electroencephalography (EEG). We collected EEG data from 4 healthy participants (all female, ages 21-27) performing a CVSA task by attending to the left or right. We then preprocessed the data and applied machine learning algorithms to classify the direction of CVSA. Due to human error, we discarded one participant’s recorded data. We discovered that the classification accuracy ranged between 70%—74.07% for the remaining 3 participants. Our classification of CVSA direction met the accuracy requirements set by pioneers in the Brain-Computer Interface (BCI) field. Therefore, in this pilot study, we conclude that CVSA direction classification through EEG can be utilized and has potential to assist individuals with minimal to no motor function with communication and control tasks.

This document is currently not available here.

Share

COinS
 
May 1st, 12:00 AM

Classification of Covert Visuospatial Attention Direction to Assist People with No Motor Function

People with minimal to no motor function lack the means necessary to communicate and control their environment. Our literature review revealed that Covert Visuospatial Attention (CVSA) classification, although thoroughly studied, has rarely been used in addressing the needs of these individuals (such as individuals with Locked-in Syndrome). CVSA is the process of attending to a target without overtly looking at it, which can be classified through noninvasive Electroencephalography (EEG). We collected EEG data from 4 healthy participants (all female, ages 21-27) performing a CVSA task by attending to the left or right. We then preprocessed the data and applied machine learning algorithms to classify the direction of CVSA. Due to human error, we discarded one participant’s recorded data. We discovered that the classification accuracy ranged between 70%—74.07% for the remaining 3 participants. Our classification of CVSA direction met the accuracy requirements set by pioneers in the Brain-Computer Interface (BCI) field. Therefore, in this pilot study, we conclude that CVSA direction classification through EEG can be utilized and has potential to assist individuals with minimal to no motor function with communication and control tasks.