Huang Zheng
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
Research Subjects
Huang Zheng has a total of 42 co-authors. Here are the top co-authors by number of shared publications:
- Shihui Han
School of Psychological and Cognitive Sciences, PKU-IDG/McGovern Institute for Brain Research, Peking University
2 shared publications Almira Kustubayeva
University of Cincinnati
1 shared publicationBin Zhang
Medical Imaging Center, First Affiliated Hospital of Jinan University, Guangzhou, People’s Republic of China
1 shared publicationErik Nelson
University of Cincinnati
1 shared publicationFei Wang
Medical Imaging Center, First Affiliated Hospital of Jinan University, Guangzhou, People’s Republic of China
1 shared publication- Guangjun Ji
Department of Early Intervention, Mental Health and Artificial Intelligence Research Center, The Second Affiliated Hospital of Xinxiang Medical University, Henan Mental Hospital, 453002 Xinxiang, Henan, China
1 shared publication
Latest Publications
- Open Access
- 18 December 2025
- Published by IMR Press in ALPHA PSYCHIATRY
- Vol. 26 (6)
Machine Learning Based Identification of Depressive Symptoms Among Students in a Chinese University Using Functional Near-Infrared Spectroscopy
Background: Individuals suffer from depression at a high rate on university campuses and current assessment methods primarily rely on subjective questionnaires. Therefore, there is a pressing need to develop objective measures for the automatic detection of depression. This study aimed to investigate the functional near-infrared spectroscopy (fNIRS) changes associated with depression and assess the potential of fNIRS signals in detecting depression among university students. Methods: A total of 192 participants were recruited for psychological assessment. A 48-channel fNIRS system was employed to measure cerebral blood oxygenation signals during the verbal fluency task (VFT). Two-sample t-tests were used to detect group differences. The association between fNIRS data and depression was identified using Pearson correlation analysis. We applied five machine learning classifiers to differentiate depression using fNIRS signals. Model performance was evaluated using receiver operating characteristic (ROC) curves, area under the curve (AUC), precision, accuracy, recall, and F1 score. A ten-fold cross-validation incorporating the recursive feature elimination algorithm was utilized. Results: Significant hemodynamic alterations were observed in the depression group at channels 4, 16, 21, 26, 32, 43, 44, and 47, in comparison with the control group. The bilateral medial prefrontal cortices (MPFC), left dorsolateral prefrontal cortex, and left temporal lobe, represented by channels 4, 16, 43, and 44, were associated with depression. Among the five machine learning algorithms, K-Nearest-Neighbors (KNN) exhibited superior classification performance (AUC = 66.51%). The left MPFC was the most significant contributor to the classification efficacy of the KNN model. Conclusion: fNIRS-VFT may serve as an objective tool for evaluating depressive symptoms in university students. The findings underscore the central role of the left MPFC in the neural mechanisms underlying depression. This work developed an fNIRS-based identification system for depression in university students.
- 18 December 2025
- Published by IMR Press in ALPHA PSYCHIATRY
- Vol. 26 (6)
Open AccessMachine Learning Based Identification of Depressive Symptoms Among Students in a Chinese University Using Functional Near-Infrared Spectroscopy
Background: Individuals suffer from depression at a high rate on university campuses and current assessment methods primarily rely on subjective questionnaires. Therefore, there is a pressing need to develop objective measures for the automatic detection of depression. This study aimed to investigate the functional near-infrared spectroscopy (fNIRS) changes associated with depression and assess the potential of fNIRS signals in detecting depression among university students. Methods: A total of 192 participants were recruited for psychological assessment. A 48-channel fNIRS system was employed to measure cerebral blood oxygenation signals during the verbal fluency task (VFT). Two-sample t-tests were used to detect group differences. The association between fNIRS data and depression was identified using Pearson correlation analysis. We applied five machine learning classifiers to differentiate depression using fNIRS signals. Model performance was evaluated using receiver operating characteristic (ROC) curves, area under the curve (AUC), precision, accuracy, recall, and F1 score. A ten-fold cross-validation incorporating the recursive feature elimination algorithm was utilized. Results: Significant hemodynamic alterations were observed in the depression group at channels 4, 16, 21, 26, 32, 43, 44, and 47, in comparison with the control group. The bilateral medial prefrontal cortices (MPFC), left dorsolateral prefrontal cortex, and left temporal lobe, represented by channels 4, 16, 43, and 44, were associated with depression. Among the five machine learning algorithms, K-Nearest-Neighbors (KNN) exhibited superior classification performance (AUC = 66.51%). The left MPFC was the most significant contributor to the classification efficacy of the KNN model. Conclusion: fNIRS-VFT may serve as an objective tool for evaluating depressive symptoms in university students. The findings underscore the central role of the left MPFC in the neural mechanisms underlying depression. This work developed an fNIRS-based identification system for depression in university students.
- Open Access
- 22 April 2025
- Published by Oxford University Press (OUP) in Social Cognitive and Affective Neuroscience
- Vol. 20 (1)
Priority of spontaneous gender categorization of same-sex faces in young adults
To cluster others into male and female groups during face perception is pivotal for appropriate social behaviours. What remains unclear is whether gender categorization of faces is mediated by the same pattern of cognitive and neural processes in women and men. The perception bias hypothesis predicts earlier gender categorization of female (vs. male) faces regardless of an observer’s gender. In contrast, the social task demand hypothesis predicts earlier gender categorization of faces that are of the same (vs. different) sex of an observer. We tested these predictions by recording electroencephalography signals to faces of one gender presented in a repetition condition and to both female and male faces in an alternating condition. The neural processes underlying gender categorization were assessed by quantifying repetition suppression of brain activities to faces in the repetition relative to alternating conditions. We found significant repetition suppression of a positive frontal-central activity at 170–210 ms after face onset (the P2 component) to female (but not to male) faces in women. However, repetition suppression of the P2 amplitude occurred to male (but not to female) faces in men. Our findings suggest that observers’ genders are pivotal for prioritization of gender categorization of male or female faces in young adults.
- 22 April 2025
- Published by Oxford University Press (OUP) in Social Cognitive and Affective Neuroscience
- Vol. 20 (1)
Open AccessPriority of spontaneous gender categorization of same-sex faces in young adults
To cluster others into male and female groups during face perception is pivotal for appropriate social behaviours. What remains unclear is whether gender categorization of faces is mediated by the same pattern of cognitive and neural processes in women and men. The perception bias hypothesis predicts earlier gender categorization of female (vs. male) faces regardless of an observer’s gender. In contrast, the social task demand hypothesis predicts earlier gender categorization of faces that are of the same (vs. different) sex of an observer. We tested these predictions by recording electroencephalography signals to faces of one gender presented in a repetition condition and to both female and male faces in an alternating condition. The neural processes underlying gender categorization were assessed by quantifying repetition suppression of brain activities to faces in the repetition relative to alternating conditions. We found significant repetition suppression of a positive frontal-central activity at 170–210 ms after face onset (the P2 component) to female (but not to male) faces in women. However, repetition suppression of the P2 amplitude occurred to male (but not to female) faces in men. Our findings suggest that observers’ genders are pivotal for prioritization of gender categorization of male or female faces in young adults.
- 12 December 2024
- Published by Elsevier in Clinical Neurophysiology
- Vol. 170, 110-119
Time-Frequency functional connectivity alterations in Alzheimer’s disease and frontotemporal dementia: An EEG analysis using machine learning
- 12 December 2024
- Published by Elsevier in Clinical Neurophysiology
- Vol. 170, 110-119
Time-Frequency functional connectivity alterations in Alzheimer’s disease and frontotemporal dementia: An EEG analysis using machine learning
- 4 June 2024
- Published by Oxford University Press (OUP) in Cerebral Cortex
- Vol. 34 (6)
Do we empathize humanoid robots and humans in the same way? Behavioral and multimodal brain imaging investigations
Humanoid robots have been designed to look more and more like humans to meet social demands. How do people empathize humanoid robots who look the same as but are essentially different from humans? We addressed this issue by examining subjective feelings, electrophysiological activities, and functional magnetic resonance imaging signals during perception of pain and neutral expressions of faces that were recognized as patients or humanoid robots. We found that healthy adults reported deceased feelings of understanding and sharing of humanoid robots’ compared to patients’ pain. Moreover, humanoid robot (vs. patient) identities reduced long-latency electrophysiological responses and blood oxygenation level–dependent signals in the left temporoparietal junction in response to pain (vs. neutral) expressions. Furthermore, we showed evidence that humanoid robot identities inhibited a causal input from the right ventral lateral prefrontal cortex to the left temporoparietal junction, contrasting the opposite effect produced by patient identities. These results suggest a neural model of modulations of empathy by humanoid robot identity through interactions between the cognitive and affective empathy networks, which provides a neurocognitive basis for understanding human–robot interactions.
- 4 June 2024
- Published by Oxford University Press (OUP) in Cerebral Cortex
- Vol. 34 (6)
Do we empathize humanoid robots and humans in the same way? Behavioral and multimodal brain imaging investigations
Humanoid robots have been designed to look more and more like humans to meet social demands. How do people empathize humanoid robots who look the same as but are essentially different from humans? We addressed this issue by examining subjective feelings, electrophysiological activities, and functional magnetic resonance imaging signals during perception of pain and neutral expressions of faces that were recognized as patients or humanoid robots. We found that healthy adults reported deceased feelings of understanding and sharing of humanoid robots’ compared to patients’ pain. Moreover, humanoid robot (vs. patient) identities reduced long-latency electrophysiological responses and blood oxygenation level–dependent signals in the left temporoparietal junction in response to pain (vs. neutral) expressions. Furthermore, we showed evidence that humanoid robot identities inhibited a causal input from the right ventral lateral prefrontal cortex to the left temporoparietal junction, contrasting the opposite effect produced by patient identities. These results suggest a neural model of modulations of empathy by humanoid robot identity through interactions between the cognitive and affective empathy networks, which provides a neurocognitive basis for understanding human–robot interactions.
- Open Access
- 1 June 2024
- Published by MDPI AG in Brain Sciences
- Vol. 14 (6) , 565
Multi-Threshold Recurrence Rate Plot: A Novel Methodology for EEG Analysis in Alzheimer’s Disease and Frontotemporal Dementia
This study introduces Multi-Threshold Recurrence Rate Plots (MTRRP), a novel methodology for analyzing dynamic patterns in complex systems, such as those influenced by neurodegenerative diseases in brain activity. MTRRP characterizes how recurrence rates evolve with increasing recurrence thresholds. A key innovation of our approach, Recurrence Complexity, captures structural complexity by integrating local randomness and global structural features through the product of Recurrence Rate Gradient and Recurrence Hurst, both derived from MTRRP. We applied this technique to resting-state EEG data from patients diagnosed with Alzheimer’s Disease (AD), Frontotemporal Dementia (FTD), and age-matched healthy controls. The results revealed significantly higher recurrence complexity in the occipital areas of AD and FTD patients, particularly pronounced in the Alpha and Beta frequency bands. Furthermore, EEG features derived from MTRRP were evaluated using a Support Vector Machine with leave-one-out cross-validation, achieving a classification accuracy of 87.7%. These findings not only underscore the utility of MTRRP in detecting distinct neurophysiological patterns associated with neurodegenerative diseases but also highlight its broader applicability in time series analysis, providing a substantial tool for advancing medical diagnostics and research.
- 1 June 2024
- Published by MDPI AG in Brain Sciences
- Vol. 14 (6) , 565
Open AccessMulti-Threshold Recurrence Rate Plot: A Novel Methodology for EEG Analysis in Alzheimer’s Disease and Frontotemporal Dementia
This study introduces Multi-Threshold Recurrence Rate Plots (MTRRP), a novel methodology for analyzing dynamic patterns in complex systems, such as those influenced by neurodegenerative diseases in brain activity. MTRRP characterizes how recurrence rates evolve with increasing recurrence thresholds. A key innovation of our approach, Recurrence Complexity, captures structural complexity by integrating local randomness and global structural features through the product of Recurrence Rate Gradient and Recurrence Hurst, both derived from MTRRP. We applied this technique to resting-state EEG data from patients diagnosed with Alzheimer’s Disease (AD), Frontotemporal Dementia (FTD), and age-matched healthy controls. The results revealed significantly higher recurrence complexity in the occipital areas of AD and FTD patients, particularly pronounced in the Alpha and Beta frequency bands. Furthermore, EEG features derived from MTRRP were evaluated using a Support Vector Machine with leave-one-out cross-validation, achieving a classification accuracy of 87.7%. These findings not only underscore the utility of MTRRP in detecting distinct neurophysiological patterns associated with neurodegenerative diseases but also highlight its broader applicability in time series analysis, providing a substantial tool for advancing medical diagnostics and research.
- 31 July 2020
- Published by Springer Nature in Brain Imaging and Behavior
- Vol. 15 (2) , 1058-1067
Changes in the topological organization of the default mode network in autism spectrum disorder
Neuroimaging studies have demonstrated that autism spectrum disorder (ASD) is accompanied by abnormal functional and structural features in specific brain
- 31 July 2020
- Published by Springer Nature in Brain Imaging and Behavior
- Vol. 15 (2) , 1058-1067
Changes in the topological organization of the default mode network in autism spectrum disorder
Neuroimaging studies have demonstrated that autism spectrum disorder (ASD) is accompanied by abnormal functional and structural features in specific brain
- 1 December 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2891-2894
A novel oversampling method based on SeqGAN for imbalanced text classification
Nowadays, investors prefer to learn the latest trends of stock market through the Internet, grasp the best time to buy and sell stocks, and simultaneously release their comments during the trading process through various online media. Therefore, many researchers utilize NLP approaches to extract sentiment tendency of investors from online comments. However, since there exists noise in corpora related to stock, when extracting emotional tendency from these corpora by using text classification, it is inevitable to face the imbalanced dataset problem, which is harmful to traditional machine learning classifiers. Our research innovatively proposes a preprocessing method based on SeqGAN algorithm, in order to solve the imbalanced classification problem. We carry out experiments on different corpora, and results show that compared with other four oversampling methods, the oversampling method based on SeqGAN can improve the performance of text classifier at best both in binary class and multiclass text classification, which proves to be an effective method for preprocessing text data.
- 1 December 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2891-2894
A novel oversampling method based on SeqGAN for imbalanced text classification
Nowadays, investors prefer to learn the latest trends of stock market through the Internet, grasp the best time to buy and sell stocks, and simultaneously release their comments during the trading process through various online media. Therefore, many researchers utilize NLP approaches to extract sentiment tendency of investors from online comments. However, since there exists noise in corpora related to stock, when extracting emotional tendency from these corpora by using text classification, it is inevitable to face the imbalanced dataset problem, which is harmful to traditional machine learning classifiers. Our research innovatively proposes a preprocessing method based on SeqGAN algorithm, in order to solve the imbalanced classification problem. We carry out experiments on different corpora, and results show that compared with other four oversampling methods, the oversampling method based on SeqGAN can improve the performance of text classifier at best both in binary class and multiclass text classification, which proves to be an effective method for preprocessing text data.
- 1 May 2017
- Published by Elsevier in Biological Psychiatry
- Vol. 81 (10) , S195
479. Brain Complexity Changes during Stress Task in Patients with Major Depressive Disorder
- 1 May 2017
- Published by Elsevier in Biological Psychiatry
- Vol. 81 (10) , S195
479. Brain Complexity Changes during Stress Task in Patients with Major Depressive Disorder
- 1 November 2011
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 240-243
Lightning current models' characteristics based on continuous wavelet transform
The analysis and simulation of two typical lightning current models by continuous wavelet transform (CWT) were given in the paper. The results compared with traditional Fourier transformation showed that the two-dimensional gray image obtained by CWT could get the information of time and frequency of the lightning current models' signals. Further more, the corresponding spectral characteristics of the lightning current models' signals were investigated. And theoretical basics for further analysis of lightning current models' characteristics were provided.
- 1 November 2011
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 240-243
Lightning current models' characteristics based on continuous wavelet transform
The analysis and simulation of two typical lightning current models by continuous wavelet transform (CWT) were given in the paper. The results compared with traditional Fourier transformation showed that the two-dimensional gray image obtained by CWT could get the information of time and frequency of the lightning current models' signals. Further more, the corresponding spectral characteristics of the lightning current models' signals were investigated. And theoretical basics for further analysis of lightning current models' characteristics were provided.