EMPT: a sparsity Transformer for EEG-based motor imagery recognition.

Introduction: Transformer network is widely emphasized and studied relying on its excellent performance. The self-attention mechanism finds a good solution for feature coding among multiple channels of electroencephalography (EEG) signals. However, using the self-attention mechanism to construct models on EEG data suffers from the problem of the large amount of data required and the complexity of the algorithm. Methods: We propose a Transformer neural network combined with the addition of Mixture of Experts (MoE) layer and ProbSparse Self-attention mechanism for decoding the time-frequency-spatial domain features from motor imagery (MI) EEG of spinal cord injury patients. The model is named as EEG MoE-Prob-Transformer (EMPT). The common spatial pattern and the modified s-transform method are employed for achieving the time-frequency-spatial features, which are used as feature embeddings to input the improved transformer neural network for feature reconstruction, and then rely on the expert model in the MoE layer for sparsity mapping, and finally output the results through the fully connected layer. Results: EMPT achieves an accuracy of 95.24% on the MI EEG dataset for patients with spinal cord injury. EMPT has also achieved excellent results in comparative experiments with other state-of-the-art methods. Discussion: The MoE layer and ProbSparse Self-attention inside the EMPT are subjected to visualisation experiments. The experiments prove that sparsity can be introduced to the Transformer neural network by introducing MoE and kullback-leibler divergence attention pooling mechanism, thereby enhancing its applicability on EEG datasets. A novel deep learning approach is presented for decoding EEG data based on MI.

Predicting Two-Photon Absorption Spectra of Octupolar Molecules: A Deep-Learning Approach Based Exclusively on Molecular Structures.

Octupolar molecules possessing a strong two-photon response are vital for numerous advanced applications. However, accurately predicting their two-photon absorption (TPA) spectra requires high-precision quantum chemical calculations, which are computationally expensive due to repeated simulations of molecular excited-state properties. To address this challenge, we introduce a deep learning approach capable of rapidly and accurately forecasting TPA spectra for octupolar molecules. By leveraging the geometric structure as an initial descriptor, we employ a graph neural network to predict the maximum two-photon transition wavelength and cross-section. Our model demonstrates a mean absolute percentage error of less than 4% compared to time-dependent density-functional theory calculations, effectively reproducing experimental observations. Notably, this deep learning technique is nearly 100 000 times faster than comparable quantum calculations, making it an efficient and cost-effective tool for simulating TPA properties of octupolar molecules. Furthermore, this method holds great promise for the high-throughput screening of exceptional TPA materials.

Multilevel Laser-Induced Pain Measurement with Wasserstein Generative Adversarial Network - Gradient Penalty Model.

Pain is an experience of unpleasant sensations and emotions associated with actual or potential tissue damage. In the global context, billions of people are affected by pain disorders. There are particular challenges in the measurement and assessment of pain, and the commonly used pain measuring tools include traditional subjective scoring methods and biomarker-based measures. The main tools for biomarker-based analysis are electroencephalography (EEG), electrocardiography and functional magnetic resonance. The EEG-based quantitative pain measurements are of immense value in clinical pain management and can provide objective assessments of pain intensity. The assessment of pain is now primarily limited to the identification of the presence or absence of pain, with less research on multilevel pain. High power laser stimulation pain experimental paradigm and five pain level classification methods based on EEG data augmentation are presented. First, the EEG features are extracted using modified S-transform, and the time-frequency information of the features is retained. Based on the pain recognition effect, the 20-40[Formula: see text]Hz frequency band features are optimized. Afterwards the Wasserstein generative adversarial network with gradient penalty is used for feature data augmentation. It can be inferred from the good classification performance of features in the parietal region of the brain that the sensory function of the parietal lobe region is effectively activated during the occurrence of pain. By comparing the latest data augmentation methods and classification algorithms, the proposed method has significant advantages for the five-level pain dataset. This research provides new ways of thinking and research methods related to pain recognition, which is essential for the study of neural mechanisms and regulatory mechanisms of pain.

A Light Vehicle License-Plate-Recognition System Based on Hybrid Edge-Cloud Computing.

With the world moving towards low-carbon and environmentally friendly development, the rapid growth of new-energy vehicles is evident. The utilization of deep-learning-based license-plate-recognition (LPR) algorithms has become widespread. However, existing LPR systems have difficulty achieving timely, effective, and energy-saving recognition due to their inherent limitations such as high latency and energy consumption. An innovative Edge-LPR system that leverages edge computing and lightweight network models is proposed in this paper. With the help of this technology, the excessive reliance on the computational capacity and the uneven implementation of resources of cloud computing can be successfully mitigated. The system is specifically a simple LPR. Channel pruning was used to reconstruct the backbone layer, reduce the network model parameters, and effectively reduce the GPU resource consumption. By utilizing the computing resources of the Intel second-generation computing stick, the network models were deployed on edge gateways to detect license plates directly. The reliability and effectiveness of the Edge-LPR system were validated through the experimental analysis of the CCPD standard dataset and real-time monitoring dataset from charging stations. The experimental results from the CCPD common dataset demonstrated that the network's total number of parameters was only 0.606 MB, with an impressive accuracy rate of 97%.

Self-Supervised EEG Representation Learning with Contrastive Predictive Coding for Post-Stroke Patients.

Stroke patients are prone to fatigue during the EEG acquisition procedure, and experiments have high requirements on cognition and physical limitations of subjects. Therefore, how to learn effective feature representation is very important. Deep learning networks have been widely used in motor imagery (MI) based brain-computer interface (BCI). This paper proposes a contrast predictive coding (CPC) framework based on the modified s-transform (MST) to generate MST-CPC feature representations. MST is used to acquire the temporal-frequency feature to improve the decoding performance for MI task recognition. EEG2Image is used to convert multi-channel one-dimensional EEG into two-dimensional EEG topography. High-level feature representations are generated by CPC which consists of an encoder and autoregressive model. Finally, the effectiveness of generated features is verified by the k-means clustering algorithm. It can be found that our model generates features with high efficiency and a good clustering effect. After classification performance evaluation, the average classification accuracy of MI tasks is 89% based on 40 subjects. The proposed method can obtain effective feature representations and improve the performance of MI-BCI systems. By comparing several self-supervised methods on the public dataset, it can be concluded that the MST-CPC model has the highest average accuracy. This is a breakthrough in the combination of self-supervised learning and image processing of EEG signals. It is helpful to provide effective rehabilitation training for stroke patients to promote motor function recovery.

Revealing the influence of doping elements (Ge, Sn, Ca, and Sr) on the properties of α-CsPbI3: a DFT investigation.

In this study, using density functional theory, we calculated the band structure and photoelectric properties in a series of 12.5% B-doped (B = Ge, Sn, Ca, and Sr) CsPbI3 perovskite systems. It is found that Ge doping can improve the structural stability and is more conducive to applications under high-pressure or by applying stress via calculating the bond length, formation energy, elastic properties, and electronic local function. In addition, the optimal direction for applying stress is achieved according to the elastic properties. Furthermore, in terms of electronic properties, the reason of energy band variation and the influence of chemical bond on the structural stability of doped α-CsPbI3 are investigated. The possibility of the applications of the CsPb0.875B0.125I3 perovskite is explored based on the optical properties. Thus, the theoretical study of the CsPb0.875B0.125I3 perovskite provides novel insights into the design of next-generation photoelectric and photovoltaic materials.

One-Dimensional Local Binary Pattern and Common Spatial Pattern Feature Fusion Brain Network for Central Neuropathic Pain.

Central neuropathic pain (CNP) after spinal cord injury (SCI) is related to the plasticity of cerebral cortex. The plasticity of cortex recorded by electroencephalogram (EEG) signal can be used as a biomarker of CNP. To analyze changes in the brain network mechanism under the combined effect of injury and pain or under the effect of pain, this paper mainly studies the changes of brain network functional connectivity in patients with neuropathic pain and without neuropathic pain after SCI. This paper has recorded the EEG with the CNP group after SCI, without the CNP group after SCI, and a healthy control group. Phase-locking value has been used to construct brain network topological connectivity maps. By comparing the brain networks of the two groups of SCI with the healthy group, it has been found that in the [Formula: see text] and [Formula: see text] frequency bands, the injury increases the functional connectivity between the frontal lobe and occipital lobes, temporal, and parietal of the patients. Furthermore, the comparison of brain networks between the group with CNP and the group without CNP after SCI has found that pain has a greater effect on the increased connectivity within the patients' frontal lobes. Motor imagery (MI) data of CNP patients have been used to extract one-dimensional local binary pattern (1D-LBP) and common spatial pattern (CSP) features, the left and right hand movements of the patients' MI have been classified. The proposed LBP-CSP feature method has achieved the highest accuracy of 98.6% and the average accuracy of 91.5%. The results of this study have great clinical significance for the neural rehabilitation and brain-computer interface of CNP patients.

Theoretical Insights on the Sensing Performance for Newly-synthesized Two-photon Fluorescent N2H4 Probes Based on Spirobifluorence.

The development of fluorescent probe for hydrazine (N2H4) detection has attracted much attention due to the important role of N2H4 plays in the fields of medicine, agriculture, biology and environments. In this paper, the optical properties and water solubility of two novel two-photon fluorescent molecular probes (Probe1 and Probe2) before and after the reaction with N2H4 are studied by using the density function theory. The results show that electronic distribution and transition dipole moment of the probes are obviously changed after the reaction with N2H4, thus the optical properties of the molecules are influenced and the detection of N2H4 are realized. In addition, photoinduced electron transfer processes for Probe1 and Probe2 in the presence of N2H4 are theoretically characterized, which explains the experimental observations from the microscopic mechanism. Special attention has been paid on the analysis of the two-photon absorption for the probes with the absence and presence of N2H4 by the response theory method. Both probes with good water solubility show large variation on the two-photon absorption cross section when reacts with N2H4. In particular, the two-photon absorption response of Probe2 is more obvious, so it possesses preferable two-photon fluorescence microscopic imaging ability. More importantly, the receptor effect on the sensing performances of the probes are demonstrated, providing a theoretical reference for the design and synthesis on more efficient two-photon fluorescence N2H4 probes. Our study provides necessary information on the response mechanism of the studied chemosensors and helps to establish the relationship between the structure and optical properties of two-photon fluorescence N2H4 probes.

Influence of Sr doping on the photoelectronic properties of CsPbX3 (X = Cl, Br, or I): a DFT investigation.

To broaden the application of cesium lead halide perovskites, doping technology has been widely proposed. In this study, we calculated a 12.5% concentration of a Sr-doped CsPbX3 (X = Cl, Br, or I) perovskite via density functional theory. The results showed that the bandgap energy of the perovskite increased by 0.2-0.3 eV. The high symmetry points of the energy band changed from R to Γ after Sr doping because the Sr doping affected the initial distribution of atomic orbital hybridization. In addition, optical absorption spectra after doping showed an obvious blueshift, whereas the absorption coefficient of CsPb0.875Sr0.125X3 had the same magnitude as undoped CsPbX3. Moreover, the effective masses of electrons and holes changed within a small range (0.01-0.03 m0) after Sr doping. According to the findings of this study, the CsPb0.875Sr0.125X3 perovskite is expected to become an ideal candidate material for designing photovoltaic and photoelectric devices.

A deep learning model for predicting selected organic molecular spectra.

Accurate and efficient molecular spectra simulations are crucial for substance discovery and structure identification. However, the conventional approach of relying on the quantum chemistry is cost intensive, which hampers efficiency. Here we develop DetaNet, a deep-learning model combining E(3)-equivariance group and self-attention mechanism to predict molecular spectra with improved efficiency and accuracy. By passing high-order geometric tensorial messages, DetaNet is able to generate a wide variety of molecular properties, including scalars, vectors, and second- and third-order tensors-all at the accuracy of quantum chemistry calculations. Based on this we developed generalized modules to predict four important types of molecular spectra, namely infrared, Raman, ultraviolet-visible, and 1H and 13C nuclear magnetic resonance, taking the QM9S dataset containing 130,000 molecular species as an example. By speeding up the prediction of molecular spectra at quantum chemical accuracy, DetaNet could help progress toward real-time structural identification using spectroscopic measurements.

Direct generation of continuous-wave and passively Q-switched visible vortex beams from a doughnut-shaped diode-pumped Pr: YLF laser.

We demonstrate the direct generation of visible vortex beams (LG01 mode) from a doughnut-shaped diode-pumped Pr:YLF laser. In continuous-wave mode, the maximum vortex output power was 36 mW at 523 nm, 354 mW at 607 nm, 838 mW at 639 nm, 722 mW at 721 nm, respectively. Moreover, based on this operation, the orange and red passively Q-switched vortex lasers were also achieved by inserting a Co:MgAl2O4 crystal into the laser cavity as a saturable absorber. The shortest pulse width of Q-switched vortex laser was 58 ns for 607 nm, and 34 ns for 639 nm, respectively. Our work provides a reliable and efficient method for the direct generation of visible vortex lasers for potential applications.

EEG decoding method based on multi-feature information fusion for spinal cord injury.

To develop an efficient brain-computer interface (BCI) system, electroencephalography (EEG) measures neuronal activities in different brain regions through electrodes. Many EEG-based motor imagery (MI) studies do not make full use of brain network topology. In this paper, a deep learning framework based on a modified graph convolution neural network (M-GCN) is proposed, in which temporal-frequency processing is performed on the data through modified S-transform (MST) to improve the decoding performance of original EEG signals in different types of MI recognition. MST can be matched with the spatial position relationship of the electrodes. This method fusions multiple features in the temporal-frequency-spatial domain to further improve the recognition performance. By detecting the brain function characteristics of each specific rhythm, EEG generated by imaginary movement can be effectively analyzed to obtain the subjects' intention. Finally, the EEG signals of patients with spinal cord injury (SCI) are used to establish a correlation matrix containing EEG channel information, the M-GCN is employed to decode relation features. The proposed M-GCN framework has better performance than other existing methods. The accuracy of classifying and identifying MI tasks through the M-GCN method can reach 87.456%. After 10-fold cross-validation, the average accuracy rate is 87.442%, which verifies the reliability and stability of the proposed algorithm. Furthermore, the method provides effective rehabilitation training for patients with SCI to partially restore motor function.

Electronic Structures and Photoelectric Properties in Cs3Sb2X9 (X = Cl, Br, or I) under High Pressure: A First Principles Study.

Lead-free perovskites of Cs3Sb2X9 (X = Cl, Br, or I) have attracted wide attention owing to their low toxicity. High pressure is an effective and reversible method to tune bandgap without changing the chemical composition. Here, the structural and photoelectric properties of Cs3Sb2X9 under high pressure were theoretically studied by using the density functional theory. The results showed that the ideal bandgap for Cs3Sb2X9 can be achieved by applying high pressure. Moreover, it was found that the change of the bandgap is caused by the shrinkage of the Sb-X long bond in the [Sb2X9]3- polyhedra. Partial density of states indicated that Sb-5s and X-p orbitals contribute to the top of the valence band, while Sb-5p and X-p orbitals dominate the bottom of the conduction band. Moreover, the band structure and density of states showed significant metallicity at 38.75, 24.05 GPa for Cs3Sb2Br9 and Cs3Sb2I9, respectively. Moreover, the absorption spectra showed the absorption edge redshifted, and the absorption coefficient of the Cs3Sb2X9 increased under high pressure. According to our calculated results, the narrow bandgap and enhanced absorption ability under high pressure provide a new idea for the design of the photovoltaic and photoelectric devices.

Theoretical study on a series of naphthalimide-contained two-photon fluorescent hypochlorite probe targeting endoplasmic reticulum: response mechanism and receptor effect.

The development of detecting hypochlorous acid (HClO) in living endoplasmic reticulum has attracted much attention in the fields of biology, medicine, and pharmacy. In the present work, the one-photon absorption (OPA), one-photon emission (OPE), and two-photon absorption (TPA) properties of a series newly synthesized chemosensors with naphthalimide as the fluorophore were systematically investigated using time-dependent density functional theory in combination with response theory. Special emphasis is placed on evolution of the probes' optical properties in the presence of HClO. These compounds show drastic changes in their photoabsorption and photoemission properties when they react with HClO, indicating them to be excellent candidates as fluorescent chemosensors. To further understand the mechanisms of the two probes, we have employed the hole and electron analysis to investigate the charge transfer process for the photoemission of the molecules. The receptor effect is found to play a dominant role in the sensing performance of these probes. Specifically, two-photon absorption properties of the molecules are calculated. We have found that all probes show significant two-photon responses in the near-infrared light region. And the maximum two-photon absorption cross section of probe 2 is greatly enhanced with the presence of HClO, indicating that probe 2 can act as a potential two-photon excited fluorescent HClO probe. The theoretical investigations would be helpful to build the structure-property relationships for the naphthalimide-contained probes, providing information on the design of efficient two-photon fluorescent sensors that can be used for biological imaging of HClO in endoplasmic reticulum.

Versatile Dicyanomethylene-Based Fluorescent Probes for the Detection of ß-Amyloid in Alzheimer's Disease: A Theoretical Perspective.

Motivated by the growing demand for target chemosensors designed with diagnostic or therapeutic capability for fibrils related to amyloidosis diseases, we investigated in the present work the response mechanism of dicyanomethylene-based fluorescent probes for amyloid fibril using a combined approach, including molecular docking, quantum mechanics/molecular mechanics (QM/MM), and the quantum chemical method. Various binding modes for the probes in ß-amyloid (Aß) are discussed, and the fibril environment-induced molecular optical changes at the most stable site are compared to the fibril-free situation in aqueous environments. The results reveal that the fluorescence enhancement for the probes in Aß observed experimentally is an average consequence over multiple binding sites. In particular, the conformational difference, including conjugation length and donor effect, significantly contributes to the optical property of the studied probes both in water and fibril. To further estimate the transition nature of the molecular photoabsorption and photoemission processes, the hole-electron distribution and the structural variation on the first excited state of the probes are investigated in detail. On the basis of the calculations, structure-property relationships for the studied chemosensors are established. Our computational approach with the ability to elucidate the available experimental results can be used for designing novel molecular probes with applications to Aß imaging and the early diagnosis of Alzheimer's disease.

Ultrafast and Polarization-Sensitive ReS2/ReSe2 Heterostructure Photodetectors with Ambipolar Photoresponse.

Recently, two-dimensional (2D) van der Waals (vdWs) heterostructures provided excellent and fascinating platforms for advanced engineering in high-performance optoelectronic devices. Herein, novel ReS2/ReSe2 heterojunction phototransistors are constructed and explored systematically that display high responsivity, wavelength-dependent ambipolar photoresponse (negative and positive), ultrafast and polarization-sensitive detection capability. This photodetector exhibits a positive photoresponse from UV to visible spectrum (760 nm) with high photoresponsivities about 126.56 and 16.24 A/W under 350 and 638 nm light illumination, respectively, with a negative photoresponse over 760 nm, which is mainly ascribed to the ambipolar photoresponse modulated by gate voltage. In addition, profound linear polarization sensitivity is demonstrated with a dichroic ratio of about ∼1.2 at 638 nm and up to ∼2.0 at 980 nm, primarily owing to the wavelength-dependent absorption anisotropy and the stagger alignment of the crystal. Beyond static photodetection, the dynamic photoresponse of this vdWs device presents an ultrafast and repeatable photoswitching performance with a cutoff frequency (f3dB) exceeding 100 kHz. Overall, this study reveals the great potential of 2D ReX2-based vdWs heterostructures for high-performance, ultrafast, and polarization-sensitive broadband photodetectors.

Deep Convolution Generative Adversarial Network-Based Electroencephalogram Data Augmentation for Post-Stroke Rehabilitation with Motor Imagery.

The motor imagery brain-computer interface (MI-BCI) system is currently one of the most advanced rehabilitation technologies, and it can be used to restore the motor function of stroke patients. The deep learning algorithms in the MI-BCI system require lots of training samples, but the electroencephalogram (EEG) data of stroke patients is quite scarce. Therefore, the expansion of EEG data has become an important part of stroke clinical rehabilitation research. In this paper, a deep convolution generative adversarial network (DCGAN) model is proposed to generate artificial EEG data and further expand the scale of the stroke dataset. First, multichannel one-dimensional EEG data is converted into a two-dimensional EEG spectrogram using EEG2Image based on the modified S-transform. Then, DCGAN is used to artificially generate EEG data based on MI. Finally, the validity of the generated artificial EEG data is proved. This paper preliminarily indicates that generating artificial stroke data is a promising strategy, which contributes to the further development of stroke clinical rehabilitation.

Theoretical Investigation of Switch Effect on the Efficiency and Adaptivity of Molecular Optoelectronic Conversion Devices.

Molecular photoswitch can effectively regulate charge separation (CS) and charge recombination (CR) in donor-acceptor (D-A) systems. However, deformation of the donor-switch-acceptor (D-S-A) systems caused by the switch isomerization will destroy the geometrical stability of the battery. Here we take the planar platinum(II) terpyridyl complex of [Pt(t Bu3 tpy)(-C≡C-Ph)n ]+ as the typical D-A model, designed six D-S-A systems using different photoswitches (dimethyldihydropyrene, fulgimide, arylazopyrazole, N-salicylideneaniline, spiropyran, and dithienylethene, denoted as D-S-A 1-6 hereafter). Our investigations show that the D-S-A 1-6 can absorb visible light of 799 nm, 673 nm, 527 nm, 568 nm, 616 nm, and 629 nm, facilitating electrons transfer from the donor and the switch to the acceptor through the Switch-on channel. Then cationic character of the photoswitch can undergo much more rapid isomerization than the neutral form due to the lower energy barrier. The Switch-off isomer breaks the conjugation of the D-S-A system, effectively turning off the CT channel and forming the CS state. Based on the evaluated conjugated backbone twist (CBT) angle, we found that D-S-A 1, 2, 4, 6 exhibit little configurational change and can be good candidates as the organic solar cell. The proposed D-S-A design controlled by the molecular switch may help to develop a solution for solar-harvesting practical applications.

Time-Varying Effective Connectivity for Describing the Dynamic Brain Networks of Post-stroke Rehabilitation.

Hemiplegia is a common motor dysfunction caused by a stroke. However, the dynamic network mechanism of brain processing information in post-stroke hemiplegic patients has not been revealed when performing motor imagery (MI) tasks. We acquire electroencephalography (EEG) data from healthy subjects and post-stroke hemiplegic patients and use the Fugl-Meyer assessment (FMA) to assess the degree of motor function damage in stroke patients. Time-varying MI networks are constructed using the adaptive directed transfer function (ADTF) method to explore the dynamic network mechanism of MI in post-stroke hemiplegic patients. Finally, correlation analysis has been conducted to study potential relationships between global efficiency and FMA scores. The performance of our proposed method has shown that the brain network pattern of stroke patients does not significantly change from laterality to bilateral symmetry when performing MI recognition. The main change is that the contralateral motor areas of the brain damage and the effective connection between the frontal lobe and the non-motor areas are enhanced, to compensate for motor dysfunction in stroke patients. We also find that there is a correlation between FMA scores and global efficiency. These findings help us better understand the dynamic brain network of patients with post-stroke when processing MI information. The network properties may provide a reliable biomarker for the objective evaluation of the functional rehabilitation diagnosis of stroke patients.

Designing Self-Adaptive Donor-Switch-Acceptor Systems for Molecular Opto-Electronic Conversion Based on Dimethyldihydropyrene/Cyclophanediene.

The introduction of a self-adaptive molecular switch is an appealing strategy to achieve complete charge separation (CS) in donor-acceptor (D-A) systems. Here, we designed donor-switch-acceptor (D-S-A) systems using a platinum(II) terpyridyl complex as the acceptor, dimethyldihydropyrene/cyclophanediene (DHP/CPD) as the bridge, and methoxybenzene, thieno[3,2-b]thiophene, 2,2'-bifuran, and 4,8-dimethoxybenzo[1,2-b:4,5-b']difuran as donors, respectively. We then systematically studied the whole opto-electronic conversion process of the donor-DHP/CPD-acceptor (D-DHP/CPD-A) systems based on time-dependent density functional theory, time-dependent ultrafast electron evolution, and electron transport property calculations. We first found that the substitution of -CH3 by -H and -CN groups in DHP/CPD can enlarge the range of the adsorption wavenumber in opto-electric conversion. Then the light absorption induces the cationization of DHP switch, largely accelerating the forth-isomerization to CPD form. Once the D-CPD-A molecule is formed, the poor conjugation can realize the complete CS state by inhibiting the radiative and nonradiative charge recombinations. Finally, the repeatable and complete CS can be achieved through the automatic back-isomerization of CPD to DHP. The present work provides valuable insights into design of D-S-A systems for practical utilization of molecule-based solar harvesting.