IEEE/ACM Trans Comput Biol Bioinform. 2021 Sep-Oct;18(5):1699-1709. doi: 10.1109/TCBB.2020.3024228. Epub 2021 Oct 7.
Electroencephalogram (EEG) is a non-invasive collection method for brain signals. It has broad prospects in brain-computer interface (BCI) applications. Recent advances have shown the effectiveness of the widely used convolutional neural network (CNN) in EEG decoding. However, some studies reveal that a slight disturbance to the inputs, e.g., data translation, can change CNN's outputs. Such instability is dangerous for EEG-based BCI applications because signals in practice are different from training data. In this study, we propose a multi-scale activity transition network (MSATNet) to alleviate the influence of the translation problem in convolution-based models. MSATNet provides an activity state pyramid consisting of multi-scale recurrent neural networks to capture the relationship between brain activities, which is a translation-invariant feature. In the experiment, Kullback-Leibler divergence is applied to measure the degree of translation. The comprehensive results demonstrate that our method surpasses the AUC of 0.0080, 0.0254, 0.0393 in 1, 5, and 10 KL divergence compared to competitors with various convolution structures.
脑电图（EEG）是一种用于采集脑信号的非侵入式方法。它在脑机接口（BCI）应用中具有广阔的前景。最近的研究表明，广泛使用的卷积神经网络（CNN）在 EEG 解码中是有效的。然而，一些研究表明，输入的微小干扰，例如数据转换，会改变 CNN 的输出。这种不稳定性对于基于 EEG 的 BCI 应用来说是危险的，因为实际信号与训练数据不同。在这项研究中，我们提出了一种多尺度活动转换网络（MSATNet），以减轻卷积模型中翻译问题的影响。MSATNet 提供了一个由多尺度递归神经网络组成的活动状态金字塔，以捕获大脑活动之间的关系，这是一种平移不变的特征。在实验中，我们应用了 Kullback-Leibler 散度来衡量翻译的程度。综合结果表明，与具有各种卷积结构的竞争对手相比，我们的方法在 1、5 和 10 KL 散度下的 AUC 分别超过了 0.0080、0.0254 和 0.0393。
