Transfer Learning in Motor Imagery Brain Computer Interface: A Review

Abstract

Sparse-view computed tomography (CT)-using a small number of projections for tomographic reconstructionenables much lower radiation dose to patients and accelerated data acquisition. The reconstructed images, however, suffer from strong artifacts, greatly limiting their diagnostic value. Current trends for sparse-view CT turn to the raw data for better information recovery. The resultant dual-domain methods, nonetheless, suffer from secondary artifacts, especially in ultra-sparse view scenarios, and their generalization to other scanners/protocols is greatly limited. A crucial question arises: have the image post-processing methods reached the limit? Our answer is not yet. In this paper, we stick to image post-processing methods due to great flexibility and propose global representation(GloRe) distillation framework for sparse-view CT, termed GloReDi. First, we propose to learn GloRe with Fourier convolution, so each element in GloRe has an image-wide receptive field. Second, unlike methods that only use the full-view images for supervision, we propose to distill GloRe from intermediate-view reconstructed images that are readily available but not explored in previous literature. The success of GloRe distillation is attributed to two key components: representation directional distillation to align the GloRe directions, and band-pass-specific contrastive distillation to gain clinically important details. Extensive experiments demonstrate the superiority of the proposed GloReDi over the state-of-the-art methods, including dual-domain ones. The source code is available at https: //github.com/longzilicart/GloReDi.

Effects of Data Including Visual Presentation and Rest Time on Classification of Motor Imagery of Using Brain-Computer Interface Competition Datasets

“…In deep learning, the neural network most commonly used for motor imagery classification is the CNN [14], [24], [25], [26], [27]. In the imaging field, AlexNet [43], Visual geometry Group Network (VGGNet) [44], GoogLeNet [45], and Residual NN (ResNet) [46] have been proposed.…”

“…We employed CNN, which are often used as classifiers in deep learning for motor imagery [24], [25], [26], [27]. We created CNNs with two different input formats.…”

“…Modern models used in classifying such motor imagery have various techniques, including deep learning, transfer learning, fuzzy models, and machine learning using linear discriminant models [11], [23], [24], [25]. To the best of our knowledge, from several reviews and recent papers [14], [24], [25], [26], [27], training models with high classification accuracies of 90 % or higher exist for BCI Competition IV 2a [10], [13], [14], [20] and IV 2b [10], [12], [16].…”