Comparing the stimulus time of the P300 Based Brain Computer Interface Systems with the Deep Learning Method

SELVİ A. O., Ferikoglu A., Guzel D.

2nd International Symposium on Multidisciplinary Studies and Innovative Technologies, ISMSIT 2018, Kizilcahamam, Ankara, Türkiye, 19 - 21 Ekim 2018 identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/ismsit.2018.8567064
  • Basıldığı Şehir: Kizilcahamam, Ankara
  • Basıldığı Ülke: Türkiye
  • Anahtar Kelimeler: Brain Computer Interface, Deep Learning, Electroencephalography, Emotiv, P300
  • Bilecik Şeyh Edebali Üniversitesi Adresli: Evet

Özet

© 2018 IEEE.With the help of Brain Computer Interface systems, people can generate commands to the computer environment by using their ability to think and focus. Some of these systems are designed using P300 signals. One of the components of the electroencephalography (EEG) signal is the positive deflection potential of approximately 300ms after the stimulus. The participants were asked to follow two different scenarios by using a computer with the help of the software prepared on Unity. The study was performed on 6 participants. In this study, stimulus time which is one of the basic elements of BBA systems was compared by using deep learning method. Transition time between stimulus in tasks was chosen from 125 to 250 milliseconds. In the classification made with deep learning, the transition time between the two stimuli resulted in 100% performance in the training data. In test data, the 125 milisecond transition time achieved 80% performance. In test data, the 250 milisecond transition time achieved 40% performance.