EEG-inception: an accurate and robust end-to-end neural network for EEG-based motor imagery classification

Abstract

Objective. Classification of electroencephalography (EEG)-based motor imagery (MI) is a crucial non-invasive application in brain–computer interface (BCI) research. This paper proposes a novel convolutional neural network (CNN) architecture for accurate and robust EEG-based MI classification that outperforms the state-of-the-art methods. Approach. The proposed CNN model, namely EEG-inception, is built on the backbone of the inception-time network, which has showed to be highly efficient and accurate for time-series classification. Also, the proposed network is an end-to-end classification, as it takes the raw EEG signals as the input and does not require complex EEG signal-preprocessing. Furthermore, this paper proposes a novel data augmentation method for EEG signals to enhance the accuracy, at least by 3%, and reduce overfitting with limited BCI datasets. Main results. The proposed model outperforms all state-of-the-art methods by achieving the average accuracy of 88.4% and 88.6% on the 2008 BCI Competition IV 2a (four-classes) and 2b datasets (binary-classes), respectively. Furthermore, it takes less than 0.025 s to test a sample suitable for real-time processing. Moreover, the classification standard deviation for nine different subjects achieves the lowest value of 5.5 for the 2b dataset and 7.1 for the 2a dataset, which validates that the proposed method is highly robust. Significance. From the experiment results, it can be inferred that the EEG-inception network exhibits a strong potential as a subject-independent classifier for EEG-based MI tasks.

Image-Based Learning Using Gradient Class Activation Maps for Enhanced Physiological Interpretability of Motor Imagery Skills

“…That is, fixing the sliding short-time window length parameter τ =2 s with an overlapping step of 1 s, resulting in N τ =5 EEG segments. For implementing the filter bank strategy, the following bandwidths of interest: ∆ f ∈{µ∈ [8][9][10][11][12], β∈ [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30]} Hz. These bandwidths belong to µ, and β rhythms, commonly associated with electrical brain activities provoked by MI tasks [53].…”

“…DL architectures can also be adapted to allow learning models to mimic extracting EEG features by imposing explicit properties on the representations learned [19]. However, a few issues remain challenging for applying CNN learners to achieve accurate and reliable single-trial detection of MI tasks: (i) The DL outputs require collecting substantial training data for avoiding overfitting inherent of small datasets so that the provided superior performance comes at the expense of much higher time and computational costs [20]; (ii) finding representations extracted from EEG data invariant to inter-and intra-subject differences [21,22]; and (iii) for dealing with non-stationary and corrupted by noise artifacts, EEG-based training frameworks involve complex, nonlinear transformations that generate many trainable DL parameters, which in turn requires a considerable number of examples to calibrate them [23], fitting the data with inconvenient understanding. Consequently, the use of weights learned by CNNs tends to be highly non-explainable [24].…”

Brain-Computer Interface: Advancement and Challenges

“…Since 2016 transfer learning is used for using MI classification tasks [317]. Some ground-breaking architectures are established in recent years, such as EEG-inception, an end-to-end Neural network [318], cluster decomposing, and multi-object optimizationbased-ensemble learning framework [319]; RFNet is a fusion network that learns from attention weights and used in embedding-specific features for decision making [179]. Now, a better understanding of the performance of commonly known classifiers with some popular datasets are given in Table 9.…”

Improving Pre-movement Pattern Detection with Filter Bank Selection

“…The movements of human limbs arouse potential changes on the human scalp, which can be observed with noninvasive brain-computer interface-based electroencephalogram (EEG) signals [1,2]. In studies on the movement detection with EEG signals, motor imagery (MI) is one of the most frequently used brain activities in the motor cortex [3][4][5]. When the limbs begin moving, the power of EEG signals in alpha rhythm (frequency range: 8∼12 Hz) and beta rhythm (frequency range: 13∼30 Hz) shows an upward or downward trend, which is called event-related desynchronization/synchronization.…”