Tengjun Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Tengjun Liu has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
- Shaomin Zhang
Qiushi Academy for Advanced Studies, Zhejiang University, Hangzhou, China
4 shared publications - Tuoru Li
Qiushi Academy for Advanced Studies, Zhejiang University, China and Department of Biomedical Engineering, Zhejiang University, China
2 shared publications - Chaoming Fang
Center of Excellence in Biomedical Research on Advanced Integrated-on-chips Neurotechnologies (CenBRAIN Neurotech), School of Engineering, Westlake University, Hangzhou, China
1 shared publication Guihua Wan
1 shared publication- Guihua Wan
Qiushi Academy for Advanced Studies, Zhejiang University, Hangzhou 310027, China
1 shared publication - Haiyan Zhang
Zhejiang University
1 shared publication
Latest Publications
- 20 November 2025
- Published by Springer Nature
- p. 84-98
Neural-CAT-SNN: Spike-Driven Neural Decoder with Multi-scale Temporal Dynamic and Reparameterized Linear Attention
- 20 November 2025
- Published by Springer Nature
- p. 84-98
Neural-CAT-SNN: Spike-Driven Neural Decoder with Multi-scale Temporal Dynamic and Reparameterized Linear Attention
- Open Access
- 26 May 2023
- Published by Association for Computing Machinery (ACM)
- p. 143-148
Transfer Learning based Efficient Schizophrenia Classification on Electroencephalogram (EEG) Signals: A Cross-Dataset Study
Although EEG classification for schizophrenia has shown promising results on individual datasets, the cross-dataset generalizability of such classification remains unknown. This study aimed to assess this generalizability through transfer learning at segment and individual levels, by employing the spectral convolutional neural network (CNN-S) on two distinct EEG datasets for schizophrenia classification. While direct cross-decoding only obtained baseline transfer accuracies of 54.72% ± 2.77% and 47.78% ± 3.62% at the segment and individual levels, the fine-tuned CNN-S achieved average cross-dataset transfer accuracies of 76.46% ± 1.17% and 56.71% ± 5.76, respectively. To improve the limited generalizability at the individual level, we applied transfer component analysis (TCA), a domain adaptation approach, to the two datasets, leading to an average cross-dataset transfer accuracy of 62.12% ± 4.86%. By combining fine-tuning and TCA, the study obtained an average cross-dataset transfer accuracy of 69.15% ± 4.09% at the individual level. Overall, transfer learning proves useful for cross-dataset EEG schizophrenia classification.
- 26 May 2023
- Published by Association for Computing Machinery (ACM)
- p. 143-148
Open AccessTransfer Learning based Efficient Schizophrenia Classification on Electroencephalogram (EEG) Signals: A Cross-Dataset Study
Although EEG classification for schizophrenia has shown promising results on individual datasets, the cross-dataset generalizability of such classification remains unknown. This study aimed to assess this generalizability through transfer learning at segment and individual levels, by employing the spectral convolutional neural network (CNN-S) on two distinct EEG datasets for schizophrenia classification. While direct cross-decoding only obtained baseline transfer accuracies of 54.72% ± 2.77% and 47.78% ± 3.62% at the segment and individual levels, the fine-tuned CNN-S achieved average cross-dataset transfer accuracies of 76.46% ± 1.17% and 56.71% ± 5.76, respectively. To improve the limited generalizability at the individual level, we applied transfer component analysis (TCA), a domain adaptation approach, to the two datasets, leading to an average cross-dataset transfer accuracy of 62.12% ± 4.86%. By combining fine-tuning and TCA, the study obtained an average cross-dataset transfer accuracy of 69.15% ± 4.09% at the individual level. Overall, transfer learning proves useful for cross-dataset EEG schizophrenia classification.
- 19 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Circuits and Systems
- Vol. 17 (3) , 598-609
A 0.99-to-4.38 uJ/class Event-Driven Hybrid Neural Network Processor for Full-Spectrum Neural Signal Analyses
Versatile and energy-efficient neural signal processors are in high demand in brain-machine interfaces and closed-loop neuromodulation applications. In this paper, we propose an energy-efficient processor for neural signal analyses. The proposed processor utilizes three key techniques to efficiently improve versatility and energy efficiency. 1) Hybrid neural network design: The processor supports artificial neural network (ANN)- and spiking neural network (SNN)-based neuromorphic processing where ANN is used to support the processing of ExG signals and SNN is used for handling neural spike signals. 2) Event-driven processing: The processor can perform always-on binary neural network (BNN)-based event detection with low-energy consumption, and it only switches to the high-accuracy convolutional neural network (CNN)-based recognition mode when events are detected. 3) Reconfigurable architecture: By exploiting the computational similarity of different neural networks, the processor supports critical BNN, CNN, and SNN operations with the same processing elements, achieving significant area reduction and energy efficiency improvement over those of a naive implementation. It achieves 90.05% accuracy and 4.38 uJ/class in a center-out reaching task with an SNN and 99.4% sensitivity, 98.6% specificity, and 1.93 uJ/class in an EEG-based seizure prediction task with dual neural network-based event-driven processing. Moreover, it achieves a classification accuracy of 99.92%, 99.38%, and 86.39% and energy consumption of 1.73, 0.99, and 1.31 uJ/class for EEG-based epileptic seizure detection, ECG-based arrhythmia detection, and EMG-based gesture recognition, respectively.
- 19 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Circuits and Systems
- Vol. 17 (3) , 598-609
A 0.99-to-4.38 uJ/class Event-Driven Hybrid Neural Network Processor for Full-Spectrum Neural Signal Analyses
Versatile and energy-efficient neural signal processors are in high demand in brain-machine interfaces and closed-loop neuromodulation applications. In this paper, we propose an energy-efficient processor for neural signal analyses. The proposed processor utilizes three key techniques to efficiently improve versatility and energy efficiency. 1) Hybrid neural network design: The processor supports artificial neural network (ANN)- and spiking neural network (SNN)-based neuromorphic processing where ANN is used to support the processing of ExG signals and SNN is used for handling neural spike signals. 2) Event-driven processing: The processor can perform always-on binary neural network (BNN)-based event detection with low-energy consumption, and it only switches to the high-accuracy convolutional neural network (CNN)-based recognition mode when events are detected. 3) Reconfigurable architecture: By exploiting the computational similarity of different neural networks, the processor supports critical BNN, CNN, and SNN operations with the same processing elements, achieving significant area reduction and energy efficiency improvement over those of a naive implementation. It achieves 90.05% accuracy and 4.38 uJ/class in a center-out reaching task with an SNN and 99.4% sensitivity, 98.6% specificity, and 1.93 uJ/class in an EEG-based seizure prediction task with dual neural network-based event-driven processing. Moreover, it achieves a classification accuracy of 99.92%, 99.38%, and 86.39% and energy consumption of 1.73, 0.99, and 1.31 uJ/class for EEG-based epileptic seizure detection, ECG-based arrhythmia detection, and EMG-based gesture recognition, respectively.
- Open Access
- 20 September 2022
- Published by MDPI AG in Brain Sciences
- Vol. 12 (10) , 1269
Volitional Generation of Reproducible, Efficient Temporal Patterns
One of the extraordinary characteristics of the biological brain is the low energy expense it requires to implement a variety of biological functions and intelligence as compared to the modern artificial intelligence (AI). Spike-based energy-efficient temporal codes have long been suggested as a contributor for the brain to run on low energy expense. Despite this code having been largely reported in the sensory cortex, whether this code can be implemented in other brain areas to serve broader functions and how it evolves throughout learning have remained unaddressed. In this study, we designed a novel brain–machine interface (BMI) paradigm. Two macaques could volitionally generate reproducible energy-efficient temporal patterns in the primary motor cortex (M1) by learning the BMI paradigm. Moreover, most neurons that were not directly assigned to control the BMI did not boost their excitability, and they demonstrated an overall energy-efficient manner in performing the task. Over the course of learning, we found that the firing rates and temporal precision of selected neurons co-evolved to generate the energy-efficient temporal patterns, suggesting that a cohesive rather than dissociable processing underlies the refinement of energy-efficient temporal patterns.
- 20 September 2022
- Published by MDPI AG in Brain Sciences
- Vol. 12 (10) , 1269
Open AccessVolitional Generation of Reproducible, Efficient Temporal Patterns
One of the extraordinary characteristics of the biological brain is the low energy expense it requires to implement a variety of biological functions and intelligence as compared to the modern artificial intelligence (AI). Spike-based energy-efficient temporal codes have long been suggested as a contributor for the brain to run on low energy expense. Despite this code having been largely reported in the sensory cortex, whether this code can be implemented in other brain areas to serve broader functions and how it evolves throughout learning have remained unaddressed. In this study, we designed a novel brain–machine interface (BMI) paradigm. Two macaques could volitionally generate reproducible energy-efficient temporal patterns in the primary motor cortex (M1) by learning the BMI paradigm. Moreover, most neurons that were not directly assigned to control the BMI did not boost their excitability, and they demonstrated an overall energy-efficient manner in performing the task. Over the course of learning, we found that the firing rates and temporal precision of selected neurons co-evolved to generate the energy-efficient temporal patterns, suggesting that a cohesive rather than dissociable processing underlies the refinement of energy-efficient temporal patterns.