GYGuosheng Yi

Guosheng Yi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Guosheng Yi has a total of 73 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Physiology-aware latent shift adaptation for sEMG-based hand kinematics decoding across able-bodied and amputee populations

    by Wuyue Zhang, Tian Gao, Jiang Wang, Bin Deng, Tianshi Gao

    Objective . Decoding continuous hand kinematics from surface electromyography (sEMG) across able-bodied and amputee populations remains a significant challenge, as amputation induces systematic alterations in neural drive and muscle activation patterns that reshape the generative relationship between motor intentions and observed sEMG signals. Such physiology-driven changes are not explicitly modeled by purely statistical domain alignment approaches, limiting their ability to achieve robust and interpretable cross-population decoding. Approach . We develop an integrated approach, termed physiology-aware latent shift adaptation (PALSA), which combines biophysical modeling with latent-space domain adaptation to characterize cross-population discrepancies at the level of latent neuromuscular representation. PALSA first learns a multidimensional, physiologically interpretable synergy latent space from able-bodied subjects (Ninapro DB2) by integrating a differentiable, motor-unit-based physiology layer. For amputee adaptation (Ninapro DB3), PALSA freezes both the neural encoder and kinematic regressor while introducing a compact, learnable latent shift vector to compensate for population-specific deviations within the learned neuromuscular manifold. Main results . Since Ninapro DB3 does not provide paired kinematics, we evaluate population-level plausibility rather than subject-specific accuracy, using DB9-derived calibrated joint angles as a standardized kinematic reference. PALSA outperforms the classic UDA algorithms CORAL and DANN in transfer learning for motor intention recognition of amputees while effectively characterizing the physiological differences between the two population groups. Across 40 gestures, PALSA achieves a mean cosine similarity of 0.6238 and a mean joint-angle root mean square error of 22.60 ∘ with respect to this reference, consistently demonstrating higher performance for extrinsic-dominant (macro) movements than for intrinsic-dependent (micro) gestures. Significance . By integrating neuromuscular modeling priors with deep transfer learning algorithms, this paper proposes a biologically meaningful and statistically robust solution for hand motion intention decoding targeted at amputees in cross-population scenarios.

  • Open Access

    Imbalance-Adaptive Dynamic Confidence Recurrent Neural Network for Fire Recognition in Multi-Sensor Environments

    by Shice Zhao, Yunhan Qin, Jixuan Wang, Yuqiao Li, Guosheng Yi, Bin Deng, Jiang Wang, Zhenxi Song, Tianshi Gao

    Accurate and timely fire recognition can effectively reduce the severe losses caused by fires. However, constrained by the imbalanced class distribution between fire and non-fire samples, traditional fire recognition models suffer from a performance bottleneck in terms of accuracy. Here, we propose an Imbalance-Adaptive Dynamic Confidence Recurrent Neural Network (IADC-RNN) for fire recognition in multi-sensor scenarios. To alleviate the problem of unbalanced numbers of fire and non-fire samples, we introduce an inter-class imbalance ratio into the loss function and assign different weights to samples according to the distribution of category numbers. We also incorporate the dynamic confidence difference to adjust the sample weights in real-time based on the current prediction confidence, strengthening the focus on hard-to-classify samples. To achieve effective fusion of multi-source data, we introduce a lightweight sensor-gated attention mechanism into IADC-RNN. Experimental results obtained on the National Institute of Standards and Technology (NIST) Report of Test FR 4016 dataset demonstrate that the IADC-RNN achieves an accuracy of 98.22%, with its true positive rate (TPR) and true negative rate (TNR) reaching 95.81% and 99.08%, respectively. Ablation experiments demonstrate that each modification contributes to performance improvement, and there is a good synergy among the three modifications. Comparative experiments indicate that the IADC-RNN outperforms state-of-the-art fire recognition models, achieving the highest accuracy and optimal comprehensive detection performance. In summary, the IADC-RNN achieves favorable fire recognition performance with low false and missed alarms, and makes a meaningful contribution to the accurate recognition of indoor fires.

  • EEG-Based Validation of Neural Response Similarity Between Vibration and Electrical Stimulation for Virtual Tactile Feedback

    by Tianshi Gao, Jiahui Jiang, Xiaoyun Sun, Guosheng Yi, Zhenxi Song, Jixuan Wang

    Electrical stimulation virtual vibration is widely used in fields like virtual reality and medical rehabilitation. However, its parameter optimization still relies on subjective psychological evaluation. This approach lacks objective quantitative criteria. The purpose of this paper is to investigate the neural response relationship between low-frequency vibration stimulation and electrical stimulation using EEG technology, providing quantitative theoretical support for optimizing electrical stimulation parameters. This paper employs a custom-built electrical stimulation system (incorporating a flexible electrode array) to conduct tactile EEG experiments with 42 participants under both electrical stimulation and vibration stimulation conditions. We extract beta-band power spectral density (PSD) and regional permutation entropy (PE) features for the two types of stimulation. Results demonstrate that under vibration stimulation, PSD at C6 and T8 channels exhibit positive correlation with frequency, while PE in the right central region and right temporal-frontal-parietal region shows positive correlation with amplitude. During electrical stimulation, corresponding neural features follow analogous patterns. For both modalities, PSD-frequency correlations (Pearson's r > 0.84) and PE-amplitude correlations (r > 0.68) achieve statistically significant levels. Finally, we conducted classification experiments using a k-nearest neighbor (kNN) classifier, with EEG features from vibratory stimulation as the training set and EEG features from electrical stimulation as the test set. The results show that the accuracy reached 66.7% for the frequency discrimination task, while the average accuracy for the amplitude discrimination task was 67.9%. These findings demonstrate significant similarity in neural signatures elicited by low-frequency vibration stimulation (1-15 Hz) and electrical stimulation. Our study provides new insights for quantifying the refinement of electrical stimulation parameters.

  • Effects of transcranial alternating current stimulation on Spike train correlation in two-compartment model neurons

    by Xuelin Huang, Xile Wei, Jiang Wang, Guosheng Yi

    Correlated spiking has been widely found in large population of neurons and been linked to neural coding. Transcranial alternating current stimulation (tACS) is a promising non-invasive brain stimulation technique that can modulate the spiking activity of neurons. Despite its growing application, the tACS effects on the temporal correlation between spike trains are still not fully understood. In this study, we use a pair of unconnected two-compartment model neurons of the integrate-and-fire (IF) type to simulate the correlated spike trains driven by shared fluctuating dendritic inputs and exposed to weak alternating electric fields. Our results show that the output correlation increases with field intensity, but increases and then decreases with field frequency, displaying thus a frequency resonance. Through varying somatic and dendritic morphologies, we demonstrate that morphological differences between the soma and dendrites fundamentally shape the correlation-frequency resonance, with more pronounced differences yielding stronger resonance effects. Moreover, the anti-phase sinusoidal modulations induced by tACS at the soma and dendrite promote this correlation-frequency resonance, particularly when dendritic fluctuations exhibit a large mean value. We further examine the tACS effects on output correlation in biophysically and morphologically realistic pyramidal model neurons, revealing similar patterns to those observed in the two-compartment models. Our findings provide new insights into how tACS modulates the correlated spike trains and highlight the critical role of morphological differences between the soma and dendrites in determining the frequency-dependent output correlation. These predictions should be taken into consideration when understanding the tACS effects on population correlation and population coding.


  • Multi-scale model of neural entrainment by transcranial alternating current stimulation in realistic cortical anatomy

    by Xuelin Huang, Xile Wei, Jiang Wang, Guosheng Yi

    Transcranial alternating current stimulation (tACS) enables non-invasive modulation of brain activity, holding promise for cognitive research and clinical applications. However, it remains unclear how the spiking activity of cortical neurons is modulated by specific electric field (E-field) distributions. Here, we use a multi-scale computational framework that integrates an anatomically accurate head model with morphologically realistic neuron models to simulate the responses of layer 5 pyramidal cells (L5 PCs) to the E-fields generated by conventional M1-SO tACS. Neural entrainment is quantified by calculating the phase-locking value (PLV) and preferred phase (PPh). We find that the tACS-induced E-field distributions across the L5 surface of interest (SOI) are heterogeneous, resulting in diverse neural entrainment of L5 PCs due to their sensitivities to the direction and intensity of the E-fields. Both PLV and PPh follow a smooth cosine dependency on the E-field polar angle, with minimal sensitivity to the azimuthal angle. PLV exhibits a positive linear dependence on the E-field intensity. However, PPh either increases or decreases logarithmically with E-field intensity that depends on the E-field direction. Correlation analysis reveals that neural entrainment can be largely explained by the normal component of the E-field or by somatic polarization, especially for E-field directed outward relative to the cortical surface. Moreover, cell morphology plays a crucial role in shaping the diverse neural entrainment to tACS. Although the uniform E-field extracted at the soma provides a good approximation for modeling tACS at the cellular level, the non-uniform E-field distribution should be considered for investigating more accurate cellular mechanisms of tACS. These findings highlight the crucial roles of heterogeneous E-field distributions, cell morphology, and E-field non-uniformity in modulating neuronal spiking activity by tACS in realistic neuroanatomy, deepening our understanding of the cellular mechanism underlying tACS. Our work bridges macroscopic brain stimulation with microscopic neural activity, which benefits the development of brain models and derived clinical applications relying on model-driven brain stimulation with tACS-induced weak E-fields.


  • Effects of dendritic Ca2+ spike on the modulation of spike timing with transcranial direct current stimulation in cortical pyramidal neurons

    by Xuelin Huang, Xile Wei, Jiang Wang, Guosheng Yi

    Transcranial direct current stimulation (tDCS) generates a weak electric field (EF) within the brain, which induces opposite polarization in the soma and distal dendrite of cortical pyramidal neurons. The somatic polarization directly affects the spike timing, and dendritic polarization modulates the synaptically evoked dendritic activities. Ca2+ spike, the most dramatic dendritic activity, is crucial for synaptic integration and top-down signal transmission, thereby indirectly influencing the output spikes of pyramidal cells. Nevertheless, the role of dendritic Ca2+ spike in the modulation of neural spike timing with tDCS remains largely unclear. In this study, we use morphologically and biophysically realistic models of layer 5 pyramidal cells (L5 PCs) to simulate the dendritic Ca2+ spike and somatic Na+ spike in response to distal dendritic synaptic inputs under weak EF stimulation. Our results show that weak EFs modulate the spike timing through the modulation of dendritic Ca2+ spike and somatic polarization, and such field effects are dependent on synaptic inputs. At weak synaptic inputs, the spike timing is advanced due to the facilitation of dendritic Ca2+ spike by field-induced dendritic depolarization. Conversely, it is delayed by field-induced dendritic hyperpolarization. In this context, the Ca2+ spike exhibits heightened sensitivity to weak EFs, thereby governing the changes in spike timing. At strong synaptic inputs, somatic polarization dominates the changes in spike timing due to the decreased sensitivity of Ca2+ spike to EFs. Consequently, the spike timing is advanced/delayed by field-induced somatic depolarization/hyperpolarization. Moreover, EFs have significant effects on the changes in the timing of somatic spike and Ca2+ spike when synaptic current injection coincides with the onset of EFs. Field effects on spike timing follow a cosine dependency on the field polar angle, with maximum effects in the field direction parallel to the somato-dendritic axis. Furthermore, our results are robust to morphological and biological diversity. These findings clarify the modulation of spike timing with weak EFs and highlight the crucial role of dendritic Ca2+ spike. These predictions shed light on the neural basis of tDCS and should be considered when understanding the effect of tDCS on population dynamics and cognitive behavior.