Ahmad Rana Fayyaz, Malik Aamir Saeed, Kamel Nidal, Reza Faruque, Amin Hafeez Ullah, Hussain Muhammad
Centre for Intelligent Signal and Imaging Research, Department of Electrical and Electronic Engineering, Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Perak, Malaysia.
Department of Neuroscience, Universiti Sains Malaysia, Kota Bharu, Kelantan, Malaysia.
Technol Health Care. 2017;25(3):471-485. doi: 10.3233/THC-161286.
Classification of the visual information from the brain activity data is a challenging task. Many studies reported in the literature are based on the brain activity patterns using either fMRI or EEG/MEG only. EEG and fMRI considered as two complementary neuroimaging modalities in terms of their temporal and spatial resolution to map the brain activity. For getting a high spatial and temporal resolution of the brain at the same time, simultaneous EEG-fMRI seems to be fruitful.
In this article, we propose a new method based on simultaneous EEG-fMRI data and machine learning approach to classify the visual brain activity patterns. We acquired EEG-fMRI data simultaneously on the ten healthy human participants by showing them visual stimuli. Data fusion approach is used to merge EEG and fMRI data. Machine learning classifier is used for the classification purposes.
Results showed that superior classification performance has been achieved with simultaneous EEG-fMRI data as compared to the EEG and fMRI data standalone. This shows that multimodal approach improved the classification accuracy results as compared with other approaches reported in the literature.
The proposed simultaneous EEG-fMRI approach for classifying the brain activity patterns can be helpful to predict or fully decode the brain activity patterns.
从大脑活动数据中对视觉信息进行分类是一项具有挑战性的任务。文献中报道的许多研究仅基于使用功能磁共振成像（fMRI）或脑电图（EEG）/脑磁图（MEG）的大脑活动模式。就绘制大脑活动的时间和空间分辨率而言，EEG和fMRI被视为两种互补的神经成像方式。为了同时获得大脑的高空间和时间分辨率，同步EEG-fMRI似乎卓有成效。
在本文中，我们提出了一种基于同步EEG-fMRI数据和机器学习方法的新方法，用于对视觉大脑活动模式进行分类。我们通过向十名健康人类受试者展示视觉刺激，同时获取了他们的EEG-fMRI数据。采用数据融合方法来合并EEG和fMRI数据。使用机器学习分类器进行分类。
结果表明，与单独使用EEG和fMRI数据相比，同步EEG-fMRI数据实现了更高的分类性能。这表明与文献中报道的其他方法相比，多模态方法提高了分类准确性结果。
所提出的用于对大脑活动模式进行分类的同步EEG-fMRI方法有助于预测或完全解码大脑活动模式。
