Early warning of cardiovascular disease relies on long-term continuous monitoring of peripheral arterial hemodynamics. Existing noninvasive methods, such as photoplethysmography (PPG), are limited by optical penetration depth and therefore have limited capability in characterizing transient hemodynamic features of deep arteries. Near-field resonant sensing has potential for this application because of its sensitivity to the dielectric properties of deep tissues. However, conventional rigid sensing interfaces are poorly matched to curved human surfaces and are therefore susceptible to unstable coupling and waveform distortion caused by air gaps and motion disturbances. To address this issue, this paper proposes a flexible conformal resonant sensing interface. A planar spiral single-coil sensor based on a flexible printed circuit (FPC) was designed and combined with a continuous tracking-based readout architecture to enable conformal attachment to the radial artery region and continuous acquisition of resonant frequency variations. Comparative experiments were conducted to evaluate the differences in signal quality and interference resistance between the flexible FPC interface and a rigid FR-4 interface. The results show that the FPC interface reduced baseline drift and improved cardiac-cycle repeatability. Its resonance-derived waveform showed relatively high consistency with the synchronized PPG reference signal (Pearson r = 0.909, p < 0.001), and short-term morphological features such as the dicrotic notch were preserved in a subset of recordings. These findings suggest that the conformability of the sensing interface may be an important factor associated with signal fidelity and artifact robustness in near-field physiological sensing, and may help inform physical front-end design for continuous and noninvasive monitoring of arterial hemodynamics.

Post-stroke neuromuscular impairment arises from heterogeneous structural and electrophysiological alterations, including disrupted membrane integrity, fiber atrophy, and increased intramuscular fat. These changes substantially alter the resistive-capacitive behavior of muscle tissue, yet conventional electrical impedance myography (EIM) metrics remain limited in capturing subtle or spatially heterogeneous abnormalities. Building upon the bio-tissue equivalent circuit model, this study proposes a geometry-based framework termed Impedance Myography Circle Mapping (IMCM), which reformulates multi-frequency EIM spectra in the complex plane to extract two physiologically interpretable parameters: the circle center (IMCM-C), reflecting global conductivity balance and the short axis (IMCM-S), associated with membrane capacitance and microstructural integrity. A cross-sectional experiment was conducted on three upper-limb muscles-the biceps brachii, pronator teres, and flexor digitorum profundus-in 33 healthy controls and 50 stroke patients using multi-frequency ( 3 kHz-1 MHz) impedance measurements. Ten supervised machine learning models were employed to compare IMCM features with conventional single-frequency and Cole-model parameters. IMCM consistently achieved superior classification across all algorithms, reaching a maximum accuracy of 88.6% in binary (healthy vs. stroke) and 93.9% in multi-muscle classification tasks. The IMCM-S feature exhibited a significant negative correlation with the Modified Ashworth Scale (MAS) ($p< 0.001$), demonstrating its potential as a quantitative biomarker of spasticity severity. These findings establish IMCM as a physiologically grounded, model-consistent extension of EIM that integrates biophysical interpretability with data-driven robustness. By linking impedance geometry to muscle pathology, IMCM offers an objective and scalable framework for quantitative neuromuscular assessment and spasticity monitoring in stroke rehabilitation.

Wearable sweat sensor systems, which monitor target biomarkers in sweat, play a crucial role in sports and health monitoring. However, there are two major challenges: first, the difficulty of simultaneously monitoring multiple biomarkers; and second, the inability to dynamically adjust sensitivity, which limits their detection range. To address the complexity of sweat composition and the wide variation in concentration ranges, we propose an adaptive amplification circuit for extremely weak current signals, combined with a time-division detection strategy tailored for sweat sensor systems. This system enables dynamic adjustment of current sensitivity and simultaneous detection of multiple biomarkers in sweat, such as glucose, lactate, Na+, and K+. We utilize an integrated, numerically controlled, switchable feedback resistor network transimpedance amplifier (TIA) circuit architecture to achieve dynamic adjustment of current sensitivity over a broad range from 10 nA to 2 mA and to enhance the linear correlation coefficients for glucose and lactate detection, reaching 0.998 and 0.995, respectively. To further realize multi-channel synchronous acquisition, the system employs a numerically controlled switch circuit combined with a time-division multiplexing strategy and dual TIA channels to simultaneously monitor 4 types of sweat biomarkers. During in vivo cycling experiments, the sensor system successfully demonstrated in vivo monitoring of 4 biomarkers in human sweat. This study provides an effective technical solution for multi-biomarker detection and health monitoring with a wide detection range.

Implantable Medical Devices (IMDs) are evolving into collaborative networks, demanding robust and ultra-low power communication in dynamic and often pathological biolog ical environments. Multi-chamber leadless pacemakers (LCPs) represent a particularly challenging case, because the intrac ardiac channel exhibits not only periodic fluctuations during healthy rhythms but also abrupt and aperiodic fades under pathological conditions such as atrioventricular (AV) block. To address this challenge, an adaptive receiver based on a Super-Regenerative Receiver (SRR) and a hybrid analog-digital Automatic Gain Control (AGC) system is proposed. At its core is a beat-synchronous two-degree-of-freedom (2-DOF) control strategy that uses peak RSSI as an electrocardiogram-free rhythm surrogate, enabling active signal stabilization through predictive feedforward and incremental Proportional-Integral Derivative (PID) control. The proposed architecture was vali dated through a three-layer framework comprising Hardware in-the-Loop (HIL) transient characterization, programmable ex vivo porcine-heart testing under a representative Mobitz Type II AV block model, and complementary ATP-provoked acute in-vivo porcine validation. Experimental results show that the proposed 2-DOF controller stabilizes the input signal to within 1% of its target under channel variations exceeding 15dB, reducing the Bit Error Rate (BER) by one to two orders of magnitude and achieving an improvement factor of more than 60 at 5kbps. In the living heart, the adaptive loop maintained stable gain control and preserved digital demodulation during ATP-provoked transient AV-block-like intervals, while supplementary in-vivo BER measurements provided additional quantitative support for end-to-end communication robustness. These results support the proposed architecture as a promising solution for robust intrac ardiac communication in future multi-chamber LCP systems.

Wireless body area networks (WBANs) based on galvanic coupled communication are a promising physical-layer paradigm for the next-generation smart healthcare systems. However, high-frequency, fine-grained data exchange among multiple nodes is prone to concurrent collisions, which can severely degrade system stability and sensing accuracy. To elucidate the physical-layer collision mechanism and quantify receiver performance limits, this work adopts a multilayer human channel model as an enabling foundation to establish a bit-level concurrent transmission analysis system for galvanic coupled communication. This framework integrates the human body channel, asynchronous interferers, and differential correlation detection to formulate a quantitative mapping from the multidimensional concurrence parameter space to the conditional bit error rate (BER). Within this framework, capture boundaries are explicitly derived along the topology, carrier-frequency offset, and time/phase offset dimensions. The results show that successful reception is governed by structured capture regions rather than by a single signal-to-interference-ratio threshold. Furthermore, direct-sequence spread spectrum (DSSS) is shown to not only enhance interference resilience but also significantly shrink the effective interference radius, thereby enabling higher achievable capacity through spatial reuse. Validated by on-body channel measurements and benchtop software-defined radio (SDR) experiments, the predicted capture boundaries are confirmed on real galvanic coupled communication links. Overall, this work provides quantitative performance bounds and practical guidelines for protocol design and cross-layer optimization in body area networks based on galvanic coupled communication.

A machine learning (ML) based hybrid method integrating XGBoost, convolutional neural network (CNN) and NSGA-Ⅱ optimization algorithm is proposed for design of a wideband circularly polarized antenna. To reduce the dimensionality of antenna structure parameters in the dataset, XGBoost is first employed to select parameters with higher sensitivity to reflection coefficient S11 and 3-dB Axial ratio (AR). The CNN then provides a fast prediction on electromagnetic responses. To leverage the CNN's inherent strength in processing image-type data, Gramian Angular Field (GAF) transformation is adopted to convert the structure parameters into images. Additionally, the CNN integrated with Visual Geometry Group (VGG) and Residual Network (ResNet) blocks enhances the prediction accuracy and model stability. While the NSGA-Ⅱ algorithm optimizes the S11 and 3-dB AR bandwidths by using the trained CNN, which significantly saves the iteration time. A circularly polarized monopole antenna is designed to validate the proposed method. The results demonstrate that the antenna achieves both wide S11 bandwidth of 122.55% (3.04∼12.66 GHz) and a wide 3-dB AR bandwidth of 76.51% (3.01∼6.74 GHz). This method provides an efficient approach for the fast design of antennas with high efficiency.

Conductive Intracardiac Communication (CIC) uses cardiac tissue as a transmission medium for short-range wireless communication and is a potential method for enabling leadless multi-chamber pacing. However, the characterization of the intracardiac channel is significantly influenced by the experimental setup and conditions. The reported results in the literature vary depending on the measurement methods used, posing challenges in obtaining reliable channel characterization for CIC. In this paper, we aim to investigate the effects of different measurement devices and conditions on the intracardiac channel. By clarifying how impedance imbalance affects the gain measurement results, we design a weak-signal measurement circuit with high common-mode rejection. This new circuit provides a more accurate and effective gain measurement scheme for the CIC channel. An equivalent circuit model simulating cardiac biomechanical impedance is constructed to analyze how capacitive and resistive imbalances affect the gain measurement results. The effects of these imbalances are verified by intracardiac channel impedance imbalance experiments. A high common-mode rejection-high-resistance differential measurement circuit that can reduce the effects of capacitive and resistive imbalances simultaneously, is then designed to suppress the interference in the experiments. The results show that changes in the measurement equipment and isolation method lead to variations in the coupling circuit characteristics, causing differences of up to 16.65 dB in the measurement results. Experiments using the designed measurement circuits effectively mitigate interference from impedance imbalance on the measurement results. This study identifies the reasons behind the discrepancies in the experimental results of previous studies and provides a more reliable gain measurement scheme for CIC research.

Transcutaneous auricular vagus nerve stimulation (taVNS) has been shown to significantly enhance motor function in stroke rehabilitation. This study investigates the synergistic effects of combined taVNS and transcranial magnetic stimulation (TMS) on corticospinal excitability. A controlled experiment was conducted with five healthy individuals and ten stroke patients. Under optimized stimulation parameters, participants underwent TMS followed by five minutes of continuous taVNS. Resting-state electroencephalography(EEG) was recorded for three minutes before and after stimulation, while motor-evoked potential (MEP) parameters were collected from the contralateral muscles of the stimulated cortical region. Small-world network analysis was applied to EEG data, and cortical excitability was quantitatively assessed using TMS. Results revealed a significant increase in clustering coefficient in the total and α frequency bands, along with a reduction in characteristic path length (P < 0.05), indicating enhanced neural network efficiency. Combined stimulation led to increased MEP amplitude and reduced latency, with significant improvements observed in the experimental group. These findings suggest that taVNS-TMS intervention enhances cortical excitability, accelerates neural conduction, and promotes functional brain reorganization, offering promising therapeutic potential for stroke rehabilitation.

Significance This study developed and validated a combined system for simultaneous detection of photosensitizer-mediated singlet oxygen (1O2) luminescence and fluorescence. The system might provide a unique tool for studying 1O2 dosimetry. Approach Superconducting strip photon detector (SSPD) and time-resolved system were used for detecting 1O2 luminescence of 1270 nm. Spectral analyzer and photodiode were used for detecting photosensitizer fluorescence. Rose Bengal (RB) was used as model photosensitizer and Q-switch laser of 532 nm as excitation light source. Results The combined system could simultaneously detect 1O2 luminescence and fluorescence. 1O2 lifetimes and specific fluorescence of RB dissolved in various solutions were obtained and analyzed. Results were consistent with published data. Linear relationship between RB concentration and 1O2 concentration as well as linear correlation between fluorescence intensity and RB concentration were demonstrated under RB concentration < 10 μM. Conclusions The custom-built system can be used for simultaneous measurement of photosensitizer-mediated production of 1O2 and fluorescence.

Person re-identification (Re-ID) plays a crucial role in the domains of security surveillance and pedestrian behavior analysis, as it aims to retrieve specific individuals captured by different cameras. However, the task of Re-ID remains immensely challenging in the field of computer vision, primarily due to the extensive intra-class variations exhibited by individuals across cameras. These variations include occlusions, illuminations, viewpoints, and poses. In this paper, we present a novel Re-ID framework that addresses the inherent issues related to intra-class variations. Our proposed approach incorporates both auxiliary-domain classification (ADC) and layered semi-second-order information bottleneck (LyrS2IB) techniques. By incorporating ADC as an auxiliary task, we leverage coarse-grained essential features that effectively distinguish individuals from the background. This enables the development of both coarse- and fine-grained feature representations for Re-ID. Furthermore, our framework integrates LyrS2IB to handle redundancy, irrelevance, and noise present in Re-ID features resulting from intra-class variations. This integration allows us to compress and optimize these features without incurring additional computation overhead during inference. Extensive experiments validate the efficacy of our proposed method, demonstrating a significant reduction in the neural network output variance of intra-class person images, firmly establishing the superior performance of our approach in the field of Re-ID.

Monitoring vital signs is essential for assessing individuals' health status and supporting various medical interventions; however, conventional methods depend on expensive and invasive hospital-based or wearable devices. This article presents a novel approach to contactless heart rate monitoring that leverages an Antenna-on-Package Pulse Coherent Radar (AoP PCR) system. To address the inherently low sampling rates associated with the pulse repetition frequency of the PCR during remote monitoring, a signal enhancement algorithm is presented. This algorithm leverages the quasi-periodic nature of chest displacement signals, leading to significantly improved temporal resolution and enabling reliable heart rate monitoring using a cost-effective PCR system. Furthermore, extracting heartbeat signals faces a significant challenge in optimally tuning the parameters of Variational Modal Decomposition (VMD) due to variations in distance and angle. To tackle this, an enhanced method called VMD based on the Whale Optimization Algorithm with Quasi-Reflection Learning (QRWOA-VMD) has been devised to enhance the precision of parameter optimization in VMD, thereby improving the decomposition accuracy of heartbeat signals across diverse angles and distances, leading to more reliable and robust heartbeat signal extraction. Comprehensive evaluation demonstrates that the proposed method achieves over 97% accuracy in heart rate monitoring under standard conditions, with the radar facing the chest within a 1.5-meter range. Even in challenging scenarios, such as a +/- 30 degrees azimuth angles and a 20 degrees elevation angle relative to the chest, accuracy remains above 93%.

Implantable intrabody communication (IBC) is a method that enables low-power, high-security communication between implanted in-body devices that could track biomedical signals and an on-body receiver by using the human body as a communication medium. As the human body consists of various tissues that each have different conductivity, this paper explores the effects of the conductivity of the communication medium on the channel gain over a wide frequency range from 10 MHz up to 300 MHz through the measurements and two models: an electrical circuit model and a FEM simulation model. Measurements are conducted using a liquid phantom with varying conductivity values from 0 S/m up to 1 S/m, covering most human tissues in the frequency range of interest. The circuit and FEM models are designed to mimic the measurement setup in order to verify the measurement results. Results show that the circuit model predicts the communication channel characteristics well at lower frequencies but cannot account for the influence of the measurement setup at higher frequencies. The influence of wire inductances, which can cause a resonant behavior when measuring at frequencies above 100 MHz, was observed using the FEM model. The results also show that the higher the conductivity of the tissue in which the device is implanted, the lower the gain of the signal, with the difference in gain being more prominent when capacitive termination with a high-impedance load is used instead of low-impedance termination. These findings provide valuable insight for selecting the appropriate interface (low-impedance vs. high-impedance termination) across specific frequency ranges for in-body to on-body (IB2OB) communication devices, while illustrating the effect of tissue conductivity on an IBC channel, thereby supporting the optimized design and implementation of reliable IB2OB communication systems.

This study aims to compare the effects of fascia gun intervention and natural recovery on gastrocnemius muscle fatigue relief and to validate the feasibility of bioimpedance (Resistance R) as a convenient indicator for recovery assessment. Five subjects were recruited to perform exhaustive fatigue exercises, followed by either natural recovery (control group) or fascia gun intervention (experimental group). During the recovery period, surface electromyography (sEMG) and electrical impedance myography (EIM) signals of the gastrocnemius muscle were collected synchronously. The effects of the two methods were compared by calculating the median frequency (MF) recovery rate, and the consistency between the resistance R recovery rate and the MF recovery rate was examined using Pearson correlation analysis. The results showed that the fascia gun intervention was significantly more effective: after 30 minutes of recovery, its average MF recovery rate (81.35%) was much higher than that of the control group (47.38%). Furthermore, the resistance R recovery rate and the MF recovery rate showed an extremely strong positive correlation under both conditions (mean r > 0.95). In conclusion, the fascia gun can effectively accelerate the functional recovery of the gastrocnemius muscle, and the bioimpedance resistance (R) is highly synchronized with the recognized sEMG recovery standard, proving that EIM is a viable and effective indicator for muscle recovery assessment.

This article presents a method for constructing compact multiple-input-multiple-output (MIMO) antennas based on subarray partition and decoupling. An efficient two-step approach is proposed to transform the N-dimensional scattering matrix into a zero matrix. First, the intra-subarray coupling of two colocated-fed antenna elements is counteracted using a section of coupled transmission line. Second, the topology-optimized pixel meta-structures (PMs), with flexibly controllable network response, are introduced between subarrays for inter-subarray decoupling. By partitioning the high-dimensional scattering matrix into subblocks describing intra-and intersubarray couplings, the multielement array decoupling problem can be decomposed into small-scale subproblems and solved step by step. To verify the proposed design method, a compact four-element planar MIMO array with an area of merely 0.603 lambda L x0.414 lambda L (at the lowest frequency of the 15-dB isolation band) is prototyped and experimentally characterized, which offers a 15-dB isolation bandwidth of 11.0% and total efficiency up to 92%. Finally, the extension of the design to a six-port MIMO array is demonstrated.

People's increasing demand for high-quality network services has prompted the continuous attention and development of network traffic classification (NTC). In recent years, deep flow inspection (DFI) is considered to be the most effective and promising method to solve the NTC. However, DFI still cannot effectively address the problem of changes in flow characteristics of complex packet flows and the discovery of new traffic categories. In this paper, we propose a metric learning based deep learning solution with feature compressor, named deep flow embedding (DFE). The feature compressor is used to compress the feature information transmitted layer by layer in DL backbone while maintaining the computational accuracy, so that the backbone can remove as much noise, redundancy, and other irrelevant information from the input data as possible, and achieve more robust feature extraction of network traffic flow. The deep learning (DL) backbone generates an embedding vector for each network packet flow. Then the embedding vector is compared with the vector template preset for each traffic type in the template library to determine the category of the packet flow. Experimental results verify that our method is more effective than the traditional DFI methods in overcoming the problems of flow characteristics variation and new category discovery.

Multi-channel impedance measurement technology enhances the monitoring of dynamic changes in complex biological tissues, such as the muscle system, by simultaneously acquiring and processing impedance signals from multiple spatial locations. This enables accurate muscle function evaluation and real-time human-machine interaction. However, existing systems face challenges, such as signal crosstalk, sampling clock drift, and limited applicability. To address these, this paper proposes an anti-interference dynamic muscle impedance measurement method based on the Data Acquisition Card (DAC) platform. The method uses interleaved excitation signals to suppress crosstalk and an adaptive segmentation algorithm to compensate for clock drift, ensuring high-precision impedance demodulation. System validation results show a high degree of consistency with commercial impedance analyzers, with a root mean square error (RMSE) below 1% and a time resolution of 250 Hz. Dual-channel consistency analysis shows all points clustered around the y=x line, with a maximum deviation of 0.534 O, confirming high channel consistency. The method successfully resolves crosstalk and drift issues, offering a reliable solution for high-precision dynamic impedance monitoring and potential wearable device applications.

Nanozymes, nanomaterials with enzyme-like characteristics which exhibit lower cost, easier synthesis and functionalization, and better stability compared with natural enzymes, have been widely developed for biosensing, disease therapy and environmental governance. However, the lack of catalytic efficiency of nanozymes compared to natural enzymes makes it difficult for them to completely replace natural enzymes to achieve higher sensitivity and lower detection limits in biosensing. Herein, magnetism-controlled technology was used to form a nanozyme array consisting of stacked Fe3O4/Au NPs at the bottom of the microchannel as a spatially confined microreactor for the catalytic reaction. By enhancing the mass transfer process of the substrate towards nanozymes mediated by the corresponding V-structure, a higher local concentration of the substrate and more efficient utilization of active sites of nanozymes were achieved to increase the catalytic efficiency (kcat/KM) of the nanozyme array consisting of Fe3O4/Au NPs by 95.2%, which was two orders of magnitude higher than that of the open reactor. Based on this, a colorimetric method on an integrated microfluidic platform was proposed for sensitive biosensing of Salmonella typhimurium. The entire detection could be completed within 30 minutes, yielding a linear range from 102 to 107 CFU mL-1 and a detection limit as low as 5.6 CFU mL-1.

Compared to conventional wired pacemakers, leadless pacemakers reduce the risk of infection and other risks associated with wires and enhance safety. Intracardiac conduction communication, with its low power consumption, high stability, and security, is more suitable than radio frequency transmission for communication between pacemakers in different heart chambers. Super-regenerative circuits, with their simple architecture and low power consumption characteristics, are well suited for wireless pacemaker receiver circuits. However, due to the effect of channel attenuation fluctuation, the intracardiac signal strength received by the super-regenerative circuit fluctuates greatly, and additional circuits are needed at the receiving end to stabilize the amplitude fluctuation of the signal. In this paper, a super-regenerative receiver circuit based on amplitude adaptive technique is designed. Its front-end circuit uses adaptive technology to adjust the output amplitude of the input signal, and the back-end circuit uses the super-regenerative technology to demodulation the baseband signal. Experimental results show that the output voltage of the adaptive circuit is stabilized at about 1 V when the amplitude of the transmitted signal is between 2 mV and 10mV. The adaptive receiver circuit maintains a lower bit error rate compared to the super-regenerative receiver circuit at all transmission rates. At the same time, under different conditions of input power and data transmission rate, the time delay of the adaptive receiver circuit can be controlled within 120 μs.

Force prediction is crucial for functional rehabilitation of the upper limb. Surface electromyography (sEMG) signals play a pivotal role in muscle force studies, but its non-stationarity challenges the reliability of sEMG-driven models. This problem may be alleviated by fusion with electrical impedance myography (EIM), an active sensing technique incorporating tissue morphology information. This study designed a wearable multimodal physiological measurement system to acquire sEMG and EIM signals simultaneously. The feature quantification indexes were defined for quantitative analysis of the efficacy of EIM and sEMG in static force prediction. We finally proposed Self-Attention Convolutional Long Short-Term Memory (SACLSTM) network to capture the spatio-temporal information among EIM and sEMG features for cross-modal feature fusion. The results indicated that EIM exhibited greater sensitivity to variations in static force compared to sEMG, especially at low muscle activation levels. Furthermore, the proposed SACLSTM network is significantly superior to LSTM, ConvLSTM, and several other baseline methods. Compared to the LSTM and ConvLSTM networks, the SACLSTM model exhibits an R2 improvement of 12.4% and 3%, respectively, and an root mean square error reduction of 63% and 29%. Especially for patients with upper limb dysfunction, the accuracy and stability of the multimodal model were significantly improved after feature fusion compared with using only EIM or sEMG unimodal features. This study emphasised the great potential of fusing EIM and sEMG features to improve performance in the muscle force prediction, opening up new practice paths in the field of functional motor rehabilitation.

Heart rate variability (HRV), indicating the variation in intervals between consecutive heartbeats, is a crucial physiological indicator of human health. However, detecting HRV using frequency-modulated continuous-wave (FMCW) radar is highly susceptible to interference from respiration, minor body movements, and environmental noise, especially in multitarget scenarios. To address these challenges, we propose the Health-Radar system, which comprises three functional modules. In the target detection module, the system accurately identifies the number and locations of targets. In the phase extraction module, the signal undergoes dc offset calibration to extract the chest displacement signals. In the heartbeat signal extraction module, we introduce Health-VMD, an adaptive parameter variational mode decomposition (VMD) method. This method optimizes the VMD parameters using an improved grasshopper optimization algorithm (GOA) and accurately extracts vital sign signals from chest displacement signals to estimate HRV. In addition, we propose a novel objective function, composed of permutation entropy, mutual information, and energy loss rate (PME), specifically designed for vital sign extraction. Experiments with multiple participants in various scenarios demonstrated that the designed system can accurately identify different targets and detect HRV with high precision. The root-mean-square error (RMSE) of the detected interbeat intervals (IBIs) is 29.72 ms, the RMSE of the standard deviation of NN intervals (SDNN) is 4.1 ms, and the RMSE of the root mean square of successive differences (RMSSD) is 18.61 ms, outperforming existing methods.