Kalafatovich Jenifer, Lee Minji, Lee Seong-Whan
Annu Int Conf IEEE Eng Med Biol Soc. 2020 Jul;2020:3363-3366. doi: 10.1109/EMBC44109.2020.9175990.
Many studies have explored brain signals during the performance of a memory task to predict later remembered items. However, prediction methods are still poorly used in real life and are not practical due to the use of electroencephalography (EEG) recorded from the scalp. Ear-EEG has been recently used to measure brain signals due to its flexibility when applying it to real world environments. In this study, we attempt to predict whether a shown stimulus is going to be remembered or forgotten using ear-EEG and compared its performance with scalp-EEG. Our results showed that there was no significant difference between ear-EEG and scalp-EEG. In addition, the higher prediction accuracy was obtained using a convolutional neural network (pre-stimulus: 74.06%, on-going stimulus: 69.53%) and it was compared to other baseline methods. These results showed that it is possible to predict performance of a memory task using ear-EEG signals and it could be used for predicting memory retrieval in a practical brain-computer interface.
许多研究探索了在执行记忆任务期间的脑信号，以预测随后被记住的项目。然而，预测方法在现实生活中的应用仍然很少，并且由于使用从头皮记录的脑电图（EEG）而不实用。由于耳脑电图在应用于现实世界环境时具有灵活性，最近已被用于测量脑信号。在本研究中，我们试图使用耳脑电图预测所呈现的刺激是会被记住还是被遗忘，并将其性能与头皮脑电图进行比较。我们的结果表明，耳脑电图和头皮脑电图之间没有显著差异。此外，使用卷积神经网络获得了更高的预测准确率（刺激前：74.06%，刺激进行中：69.53%），并与其他基线方法进行了比较。这些结果表明，使用耳脑电图信号预测记忆任务的表现是可能的，并且它可用于在实际的脑机接口中预测记忆检索。
