Straathof, Pjotr (2024) Mindreading using EEG data: Decoding visually perceived and imagined numbers in the brain. Bachelor's Thesis, Artificial Intelligence.
|
Text
Thesisfinal.pdf Download (847kB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
toestemming.pdf Restricted to Registered users only Download (125kB) |
Abstract
Recent advances in brain-computer interfaces (BCIs) have shown great potential in improving the lives of people with limited capabilities such as paraplegics or people suffering from locked-in syndrome (LIS). This study was done to further the development of decoding electroencephalogram (EEG) data by researching the difference in accuracy of a Recurrent Neural Network (RNN) trained on two different datasets. One dataset consisted of EEG data of participants imagining a number, whereas the other consisted of EEG data of participants visually perceiving a number. By comparing the results, I aimed get a decisive difference that helps pave the way for more efficient EEG decoding by creating a better understanding of it. I used the consumer-grade MUSE 2 headband, as it had been proven to achieve high results with only four electrodes and could be used at home by anyone. After training, the classifiers were evaluated using their accuracy, but neither received an accuracy score significantly above chance level. Therefore, the results of this study remained inconclusive. Further research should be done using the EPOC headset with 15 electrodes and a Convolutional Neural Network (CNN), as these have been proven to achieve better results.
| Item Type: | Thesis (Bachelor's Thesis) |
|---|---|
| Supervisor name: | Vortmann, L. |
| Degree programme: | Artificial Intelligence |
| Thesis type: | Bachelor's Thesis |
| Language: | English |
| Date Deposited: | 02 Aug 2024 12:50 |
| Last Modified: | 02 Aug 2024 12:51 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/33768 |
Actions (login required)
 |
View Item |