Logistic regression of working memory impairments in children based on single-trial ERP features

• Main Authors: Mohd. Tumari, Siti Zubaidah, Sudirman, Rubita
• Format: Conference or Workshop Item
• Published: 2015
• Subjects: QA75 Electronic computers. Computer science; TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.utm.my/id/eprint/59359/; http://dx.doi.org/10.1109/TIME-E.2014.7011629
• Institution: Universiti Teknologi Malaysia

The purpose of this study is to determine the performance of the working memory in processing visual stimuli based on the single-trial Event-Related Potential (ERP) signal. This signal was further analyzed using the logistic regression method to investigate children's working memory ability. This study involved 54 children aged 10 to 12 years old who were given computer-based visual stimuli while their working memory activities were simultaneously recorded using the Neurofax 9200 EEG acquisition machine. Redundant noises and artifacts were removed using wavelet algorithm and ICA techniques. The EEG signals were segmented to obtain the average mean of single-trial ERP and necessary input for the logistic regression classifier. Results indicated that the resultant parameters as well as biological data were significant in predicting whether a child had working impairment or not. Moreover, the signal's sensitivity value was more than the specificity value. In conclusion, our results indicated the working memory impairment in children tends to worsen with age.