Wanqing Wu

Objective: Upper-limb electrocardiogram (ECG) acquired from a single-arm wearable device offers a practical approach for dynamic cardiac monitoring. This study addresses whether single-channel Arm-ECG, characterized by higher noise and distinct morphology compared to standard 12-lead ECG, can reliably detect arrhythmias. Methods: We propose CLMF-Net, a contrastive multitask framework featuring multiscale convolutional layers to capture temporal-morphological patterns. A fine-grained reconstruction branch preserves subtle clinical features, including P waves and ST segments, while a contrastive module ensures robustness against signal quality variations. The complete framework is trained end-to-end using a unified loss function that balances classification, reconstruction, and consistency objectives. Results: Validation with 132 elderly participants undergoing six-minute walk tests confirmed the feasibility of dynamic Arm-ECG acquisition. CLMF-Net, trained for arrhythmia identification from single-channel ECG, achieved 95.19% accuracy on the Arm-ECG dataset. It effectively detects arrhythmias, despite challenges with noise and morphology. Generalizability was evidenced by 95.55% and 84.23% accuracy on the Chapman and China Physiological Signal Challenge 2018 (CPSC2018) datasets, respectively, demonstrating the model's ability to classify multiple arrhythmia types across diverse datasets. Conclusion: Beyond precise identification of arrhythmias, deep feature and morphology analyses showed that the learned representations do not always align with clinically emphasized intervals. This discrepancy highlights both the promise of leveraging Arm-ECG signals and the caution required when translating model-derived interpretations into clinical workflows. Significance: This study validates the clinical utility of single-arm ECG monitoring, demonstrating feature reliability for real-world cardiac health assessment and supporting the translational potential through consistency with standard recordings.

Interactions linking the heart to changes in brain organization offer insights into how central and peripheral systems jointly contribute to affective processing. Conventional approaches to brain-heart interaction, which rely primarily on EEG rhythms and heart rate variability, fail to capture the heartbeat-resolved dynamics that characterize the bidirectional interaction between cortical and cardiac activity. We apply Convergent Cross Mapping (CCM), a nonlinear causality framework, to 32-channel EEG and single-lead ECG signals recorded during validated virtual reality (VR) videos designed to elicit positive or negative affect. By modeling heartbeat-evoked potentials (HEPs) and full cardiac waveforms as evolving dynamical systems, we observed pronounced directional asymmetry in emotion-related brain-heart interactions: descending coupling patterns were stronger, more sustained, and spatially focused than ascending effects. Positive and negative emotional conditions were associated with differential bidirectional coupling patterns, with preferential enhancement of descending cortical-to-cardiac interactions. Spatially, modulation was left-lateralized, with a concentration in frontal and central electrodes. These findings advance embodied models of affect by demonstrating that such interactions are both directionally asymmetric and hemispherically organized, positioning cardiac dynamics as an actively involved component of emotion-related brain-heart coordination.

Evidence linking cardiovascular health to cognitive function suggests that interactions within the central autonomic network (CAN) may contribute to the development and progression of post-stroke cognitive impairment (PSCI). Heart-brain coupling (HBC) can serve as a readout of coordinated CAN function and offers a novel perspective for understanding PSCI. In this cross-sectional case–control study, we enrolled 16 patients with PSCI and 19 age-matched healthy controls and acquired simultaneous electroencephalography (EEG) and electrocardiography (ECG) recordings during resting state and a cognitive load condition. At both the sensor and source levels, we used the maximal information coefficient to quantify nonlinear coupling between cortical rhythms and heart rate variability (HRV) indices derived from Poincaré plots, and examined associations between these coupling metrics and the severity of cognitive impairment. Compared with healthy controls, HBC in PSCI is characterized by altered task-related modulation patterns, with a potential reduction in the flexibility of coordination between cortical activity and autonomic regulation under cognitive load. Some coupling features partially reflect the severity of cognitive impairment. Spatially, coupling patterns tend to involve widely distributed cortical regions rather than being confined to focal alterations. This study characterizes the nonlinear coupling between cortical rhythms and HRV under both resting and cognitive load conditions in PSCI, which may support a role of CAN dysfunction in the pathophysiology of PSCI and proposes testable potential regulatory targets for PSCI risk stratification and rehabilitation monitoring.

In EEG-based affective computing, addressing individual differences for model generalizability is crucial. Psychological research has demonstrated that personality traits significantly influence emotional perception, thereby introducing individual response specificity. Existing public datasets employed the Big Five personality scale and few-channel EEG sensors. As a supplement, we developed a dataset employing the widely utilized Eysenck personality scale in clinical settings and 32-channel EEG sensors. Our investigation into personality traits and EEG topography reveals that variations in personality traits significantly influence emotional EEG response patterns. Building on these findings, we propose a novel Personality Cross-Attention Mechanism (PCAM) designed to enhance the generalization capability of mainstream emotion recognition models. This mechanism leverages insights from psychology and neuroscience to improve model performance. Empirical evaluations on AMIGOS and our proprietary dataset demonstrate that PCAM effectively enhances the generalization ability of mainstream emotion recognition models, including CNN-LSTM, EEGNet, ACRNN, and Conformer. Furthermore, visualization of the PCAM weights reveals its ability to discern differences in emotional EEG responses associated with different personality traits, thereby providing explainability for the observed improvements in model generalization ability.

The paper proposes a low-power Successive Approximation Register (SAR) Analog-to-Digital Conversion (ADC) with dual bypass windows based on non-binary split capacitors. To reduce the power consumption, the bypass windows constituted by the split capacitors can maximize the coverage of biological signals both in the resting state and excited state. When the signal falls within the designated window, unnecessary conversion cycles are skipped. This process is mainly judged and controlled by digital circuits, which is highly robust and does not require calibration. Meanwhile, a low-power dynamic CMOS comparator is proposed, which can effectively reduce the voltage variation of the latch node during the comparator's operation, further reducing power consumption. The proposed SAR ADC, based on a 180nm process, measures a power consumption of 9.68nW at a supply voltage of 0.6V and a sampling rate of 5.21kS/s. The signal-to-noise-and-distortion ratio (SNDR) and the spur-free dynamic range (SFDR) are measured at 57.51dB and 71.68dB, respectively. It also achieves an effective number of bits (ENOB) of 9.26 bits and a Walden figure-of-merit (FoM) of 2.9 fJ/conv.-step. The proposed SAR ADC is also verified by collected electromyogram (EMG), electrocardiogram (ECG), and electroencephalogram (EEG) signals. The average power consumption for quantifying EMG signals is 7.95 nW, providing an attractive solution for low-power SAR ADCs in biomedical applications.

The challenge of post-stroke rehabilitation lies in the difficulty of quantifying the dynamic process of neural remodeling using traditional assessment met