The brain-computer interface (BCI) based on motor imagery electroencephalography (MI-EEG) enables direct information interaction between the human brain and external devices. In this paper, a multi-scale EEG feature extraction convolutional neural network model based on time series data enhancement is proposed for decoding MI-EEG signals. First, an EEG signals augmentation method was proposed that could increase the information content of training samples without changing the length of the time series, while retaining its original features completely. Then, multiple holistic and detailed features of the EEG data were adaptively extracted by multi-scale convolution module, and the features were fused and filtered by parallel residual module and channel attention. Finally, classification results were output by a fully connected network. The application experimental results on the BCI Competition IV 2a and 2b datasets showed that the proposed model achieved an average classification accuracy of 91.87% and 87.85% for the motor imagery task, respectively, which had high accuracy and strong robustness compared with existing baseline models. The proposed model does not require complex signals pre-processing operations and has the advantage of multi-scale feature extraction, which has high practical application value.
Motor imagery electroencephalogram (EEG) signals are non-stationary time series with a low signal-to-noise ratio. Therefore, the single-channel EEG analysis method is difficult to effectively describe the interaction characteristics between multi-channel signals. This paper proposed a deep learning network model based on the multi-channel attention mechanism. First, we performed time-frequency sparse decomposition on the pre-processed data, which enhanced the difference of time-frequency characteristics of EEG signals. Then we used the attention module to map the data in time and space so that the model could make full use of the data characteristics of different channels of EEG signals. Finally, the improved time-convolution network (TCN) was used for feature fusion and classification. The BCI competition IV-2a data set was used to verify the proposed algorithm. The experimental results showed that the proposed algorithm could effectively improve the classification accuracy of motor imagination EEG signals, which achieved an average accuracy of 83.03% for 9 subjects. Compared with the existing methods, the classification accuracy of EEG signals was improved. With the enhanced difference features between different motor imagery EEG data, the proposed method is important for the study of improving classifier performance.
ObjectiveTo compare the prognosis of neoadjuvant chemotherapy (NAC) and adjuvant chemotherapy (AC) in patients with T1-2N1-2M0 luminal breast cancer, and to analyze the factors affecting the prognosis. MethodsPatients with luminal breast cancer who met the inclusion criteria and had complete follow-up data from January 2014 to December 2019 were retrospectively collected. Patients received either neoadjuvant chemotherapy (NAC) or adjuvant chemotherapy (AC), both based on anthracycline-containing regimens. Kaplan-Meier analysis was performed to estimate overall survival, and Cox proportional hazards regression was used to identify risk factors affecting 5-year cumulative overall survival rate. Statistical significance was defined as a P=0.05.ResultsA total of 206 patients (99 receiving NAC and 107 receiving AC) meeting the inclusion criteria were enrolled. The cohort comprised 101 patients with luminal A (57 AC, 44 NAC) and 105 with luminal B (50 AC, 55 NAC). At a median follow-up of 72.5 months, no significant differences in the 5-year cumulative overall survival rates were observed between AC and NAC patients (89.7% vs. 88.9%, P=0.571). However, the 5-year cumulative disease-free survival rate was significantly higher in the AC group as compared with the NAC group (85.0% vs. 73.5%, P<0.001). Subgroup analysis demonstrated no significant differences in the 5-year cumulative overall survival rates between AC and NAC patients within either luminal A (94.7% vs. 86.4%, P=0.727) or luminal B (84.0% vs. 89.3%, P=0.864). However, for patients with luminal A, the 5-year cumulative disease-free survival rate was significantly higher in the AC subgroup than in the NAC subgroup (93.0% vs. 77.3%, P<0.001). In contrast, no significant difference in the 5-year cumulative disease-free survival rate between AC and NAC was observed among patients with luminal B (74.0% vs. 71.4%, P=0.201). Multivariate analysis using the Cox proportional hazards model identified the following independent risk factors for lower 5-year cumulative overall survival rate in patients with T1-2N1-2M0 luminal breast cancer: N2 stage [HR (95%CI)=2.290 (1.249, 4.196)], lymphovascular invasion [HR (95%CI)=2.181 (1.182, 4.026)], omission of endocrine therapy [HR (95%CI)=6.013 (2.590, 13.965)], and absence of pathological complete response (pCR) after NAC [HR (95%CI)=2.403 (1.284, 4.496)]. ConclusionThe results of this study suggest that patients with T1-2N1-2M0 luminal breast cancer can achieve higher disease-free survival from AC, but it is still necessary to comprehensively consider the patient’s condition such as lymph node metastasis, vascular cancer thrombus, and other factors to formulate an individualized treatment plan to increase the overall survival rate of patients.
The aging population and the increasing prevalence of chronic diseases in the elderly have brought a significant economic burden to families and society. The non-invasive wearable sensing system can continuously and real-time monitor important physiological signs of the human body and evaluate health status. In addition, it can provide efficient and convenient information feedback, thereby reducing the health risks caused by chronic diseases in the elderly. A wearable system for detecting physiological and behavioral signals was developed in this study. We explored the design of flexible wearable sensing technology and its application in sensing systems. The wearable system included smart hats, smart clothes, smart gloves, and smart insoles, achieving long-term continuous monitoring of physiological and motion signals. The performance of the system was verified, and the new sensing system was compared with commercial equipment. The evaluation results demonstrated that the proposed system presented a comparable performance with the existing system. In summary, the proposed flexible sensor system provides an accurate, detachable, expandable, user-friendly and comfortable solution for physiological and motion signal monitoring. It is expected to be used in remote healthcare monitoring and provide personalized information monitoring, disease prediction, and diagnosis for doctors/patients.