An end-to-end multi-task motor imagery EEG classification neural network based on dynamic fusion of spectral-temporal features.

Electroencephalograph (EEG) brain-computer interfaces (BCI) have potential to provide new paradigms for controlling computers and devices. The accuracy of brain pattern classification in EEG BCI is directly affected by the quality of features extracted from EEG signals. Currently, feature extraction heavily relies on prior knowledge to engineer features (for example from specific frequency bands); therefore, better extraction of EEG features is an important research direction. In this work, we propose an end-to-end deep neural network that automatically finds and combines features for motor imagery (MI) based EEG BCI with 4 or more imagery classes (multi-task). First, spectral domain features of EEG signals are learned by compact convolutional neural network (CCNN) layers. Then, gated recurrent unit (GRU) neural network layers automatically learn temporal patterns. Lastly, an attention mechanism dynamically combines (across EEG channels) the extracted spectral-temporal features, reducing redundancy. We test our method using BCI Competition IV-2a and a data set we collected. The average classification accuracy on 4-class BCI Competition IV-2a was 85.1 % ± 6.19 %, comparable to recent work in the field and showing low variability among participants; average classification accuracy on our 6-class data was 64.4 % ± 8.35 %. Our dynamic fusion of spectral-temporal features is end-to-end and has relatively few network parameters, and the experimental results show its effectiveness and potential.

Effects of rhythmic visual cues on cognitive resources allocation characterized by electroencephalographic (EEG) features during human gait initiation.

Rhythmic visual cues are beneficial in gait initiation (GI) in Parkinson's disease patients with freezing of gait (FOG), however, the underlying neurophysiological mechanism remains poorly understood. The cognitive control modulated by visual cues during GI has been investigated and considered as a potential factor influencing automatic motor actions, but it is unclear how rhythmic visual cues affect cognitive resources demands during GI. The purpose of this study was to explore the effect of rhythmic visual cues on cognitive resources allocation by recording the anticipatory cerebral cortex electroencephalographic (EEG) activity during GI. Twenty healthy participants initiated gait in response to the rhythmic and non-rhythmic visual cues of stimulus presentation. We assessed the contingent negative variation (CNV) of averaged EEG data over 32 electrode positions during GI preparation, the results of which showed that the CNV was induced over prefrontal, frontal, central, and parietal regions in both rhythmic conditions and non-rhythmic conditions. Overall, different visual cues modulated the amplitude of CNV in the early and late stages of the GI preparation. Compared with the non-rhythmic condition, the CNV amplitude was lower in rhythmic condition over displayed regions precede the GI onset. In the late stage of GI preparation, it showed significant differences between the two conditions in prefrontal, frontal, and central regions, and the amplitude of CNV was lower under rhythmic condition. More to the point, the differences were more obvious in the late stage of GI preparation between the two conditions, which was closely associated with the cognitive resources. Therefore, the results indicate that less cognitive resources allocation is required to trigger GI under rhythmic visual cues compared with non-rhythmic visual cues. This study may provide a new insight into why rhythmic visual cues are more effective in improving GI ability compared to non-rhythmic visual cues.

A Novel Time-Incremental End-to-End Shared Neural Network with Attention-Based Feature Fusion for Multiclass Motor Imagery Recognition.

In the research of motor imagery brain-computer interface (MI-BCI), traditional electroencephalogram (EEG) signal recognition algorithms appear to be inefficient in extracting EEG signal features and improving classification accuracy. In this paper, we discuss a solution to this problem based on a novel step-by-step method of feature extraction and pattern classification for multiclass MI-EEG signals. First, the training data from all subjects is merged and enlarged through autoencoder to meet the need for massive amounts of data while reducing the bad effect on signal recognition because of randomness, instability, and individual variability of EEG data. Second, an end-to-end sharing structure with attention-based time-incremental shallow convolution neural network is proposed. Shallow convolution neural network (SCNN) and bidirectional long short-term memory (BiLSTM) network are used to extract frequency-spatial domain features and time-series features of EEG signals, respectively. Then, the attention model is introduced into the feature fusion layer to dynamically weight these extracted temporal-frequency-spatial domain features, which greatly contributes to the reduction of feature redundancy and the improvement of classification accuracy. At last, validation tests using BCI Competition IV 2a data sets show that classification accuracy and kappa coefficient have reached 82.7 ± 5.57% and 0.78 ± 0.074, which can strongly prove its advantages in improving classification accuracy and reducing individual difference among different subjects from the same network.

[Value of MRI in the assessment of invasion depth by endometrial carcinoma].

OBJECTIVE: To evaluate the value of MR imaging in the assessment of invasion depth by endometrial carcinoma. METHODS: Data of 122 patients with endometrial carcinoma proved by postoperative pathology were retrospectively reviewed. Preoperatively, all patients underwent conventional and contrast-enhanced MR scan. Compared with the results of pathology, the sensitivity, specificity and accuracy of different invasion depth determined by MRI were analyzed with SPSS software based on whether the junctional zone was involved or not as the criterion of myometrial invasion. RESULTS: (1) Based on MRI image, the tumor was confined to the endometrium in 17 patients, causing superficial myometrial invasion 60, deep-myometrial invasion 40 and having penetrated the serosa 5. Compared with postoperative pathology results, the incidence of sensitivity, specficity and accuracy of MRI assessment for tumor confined to endometrium was 64.7%, 94.3%, 90.2%, respectively; to superficial myometrial invasion: 64.6%, 82.5%, 70.5%, respectively; to deep-myometrial invasion: 94.4%, 77.9%, 80.3%, respectively; to tumor having penetrated the serosa: 80.0%, 99.1%, 98.4%, respectively. (2) Based on intact junctional zone as the criterion of tumor confined to endometrium, the sensitivity, specficity, accuracy, positive and negative predictive value was 92.9%, 67.9%, 73.1%, 43.3%, 97.3%, respectively. Based on the interruption of junctional zone as the criterion of tumor having involved the myometrium, the sensitivity, specficity, accuracy, positive and negative predictive value was 67.9%, 92.9%, 73.1%, 97.3%, 43.3%, respectively. CONCLUSION: MRI is valuable in the assessment of the invasion depth by endometrial carcinoma, and the dose plays an important role for the clinician in selecting proper way of therapy.