Objective. The combination of the motor imagery (MI) electroencephalography (EEG) signals and deep learning-based methods is an effective way to improve MI classification accuracy. However, deep learning-based methods often need too many trainable parameters. As a result, the trade-off between the network decoding performance and computational cost has always been an important challenge in the MI classification research. Approach. In the present study, we proposed a new end-to-end convolutional neural network (CNN) model called the EEG-circular dilated convolution (CDIL) network, which takes into account both the lightweight model and the classification accuracy. Specifically, the depth-separable convolution was used to reduce the number of network parameters and extract the temporal and spatial features from the EEG signals. CDIL was used to extract the time-varying deep features that were generated in the previous stage. Finally, we combined the features extracted from the two stages and used the global average pooling to further reduce the number of parameters, in order to achieve an accurate MI classification. The performance of the proposed model was verified using three publicly available datasets. Main results. The proposed model achieved an average classification accuracy of 79.63% and 94.53% for the BCIIV2a and HGD four-classification task, respectively, and 87.82% for the BCIIV2b two-classification task. In particular, by comparing the number of parameters, computation and classification accuracy with other lightweight models, it was confirmed that the proposed model achieved a better balance between the decoding performance and computational cost. Furthermore, the structural feasibility of the proposed model was confirmed by ablation experiments and feature visualization. Significance. The results indicated that the proposed CNN model presented high classification accuracy with less computing resources, and can be applied in the MI classification research.

Respiratory detection plays a significant role in the assessment of human health status and in aiding the diagnosis of diseases. This requires the development of fast, economical and non-invasive detection technologies. This research employed tunable diode laser absorption spectroscopy (TDLAS) to analyze exhaled methane (CH ₄ ) levels in patients with gastrointestinal diseases. The system achieving a minimum detection limit (MDL) of 42.12 ppbv, and an accuracy of 8.04 ppbv. A total of 451 samples of exhaled CH ₄ concentration from gastrointestinal surgery patients were detected, and several gastrointestinal diseases with high average CH ₄ concentration were screened out. Support vector machine (SVM) algorithms were used to investigate the relationship between CH ₄ levels and specific gastrointestinal diseases. The study achieved a classification accuracy, recall, and precision rate of 100% for ileus, ascending colon cancer (COAD), and cancers with liver metastasis. Meantime, the rate of CH ₄ positivity for patients undergone postoperative chemotherapy was observed to be significantly lower than patients with active cancer. However, the presence of CH ₄ was not significantly associated with gastrointestinal stromal tumors of the stomach (GIST), stomach cancer (SC), or rectal cancer (READ), indicating the need for a multifaceted diagnostic approach incorporating additional gas markers. In conclusion, CH ₄ analysis holds promise as a novel non-invasive diagnostic tool for gastrointestinal diseases, with significant potential for clinical application.

Structural color (SC) has enormous potential for improving the visualization and identification of functional micro/nano structures for information encryption and intelligent sensing. Nevertheless, achieving the direct writing of SCs at the micro/nano scale and the change of color in response to external stimuli simultaneously is rather challenging. To this end, we directly printed woodpile structures (WSs) utilizing femtosecond laser two-photon polymerization (fs-TPP), which demonstrated obvious SCs under an optical microscope. After that, we achieved the change of SCs by transferring WSs between different mediums. Furthermore, the influence of laser power, structural parameters, and mediums on the SCs was systematically investigated, and the mechanism of SCs using the finite-difference time-domain (FDTD) method was further explored. Finally, we realized the reversible encryption and decryption of certain information. This finding holds broad application prospects in smart sensing, anti-counterfeiting tags, and advanced photonic devices.

Objective. Standing balance forms the basis of daily activities that require the integration of multi-sensory information and coordination of multi-muscle activation. Previous studies have confirmed that the cortex is directly involved in balance control, but little is known about the neural mechanisms of cortical integration and muscle coordination in maintaining standing balance. Approach. We used a direct directed transfer function (dDTF) to analyze the changes in the cortex and muscle connections of healthy subjects (15 subjects: 13 male and 2 female) corresponding to different standing balance tasks. Main results. The results show that the topology of the EEG brain network and muscle network changes significantly as the difficulty of the balancing tasks increases. For muscle networks, the connection analysis shows that the connection of antagonistic muscle pairs plays a major role in the task. For EEG brain networks, graph theory-based analysis shows that the clustering coefficient increases significantly, and the characteristic path length decreases significantly with increasing task difficulty. We also found that cortex-to-muscle connections increased with the difficulty of the task and were significantly stronger than the muscle-to-cortex connections. Significance. These results show that changes in the difficulty of balancing tasks alter EEG brain networks and muscle networks, and an analysis based on the directed network can provide rich information for exploring the neural mechanisms of balance control.

Light-induced thermoelastic spectroscopy (LITES), boasting an extended effective optical path length, stands as a potent approach for trace gas detection. The indispensable factor for heightened measurement sensitivity lies in the quartz tuning fork (QTF) probe, possessing a robust thermoelastic conversion efficiency. In this study, we introduce a novel laser micro-machining technique for fabricating sensitive miniature QTFs (mQTFs) tailored to suit LITES applications. This innovation effectively addresses the issues of subpar sensitivity and the elevated laser response threshold commonly observed in commercial QTFs. As a proof of concept, we demonstrate its capabilities through the measurement of CO ₂ . Capitalizing on the utilization of mQTFs, the sensing sensitivity was increased by an impressive factor of ~5, outperforming the conventional QTF-LITES approach. The minimum detection limits (MDLs) for CO ₂ are also notably improved, measuring 351.34 ppm for the mQTF and 968.29 ppm for the standard commercial QTF. This cost-effective technological advancement paves the way for the development of highly responsive trace gas LITES sensing systems.