special issue on brain-inspired neural networks for biomedical signal processing submission date: 2024-04-15 biomedical signals, such as the electrocardiogram (ecg), electroencephalogram (eeg), and electromyogram (emg), encode physiological information about the nature and condition of the human body. the study of such biomedical signals not only provides insight into physiological processes but also enables novel human–computer interfaces. recent advances in artificial neural networks (anns) have contributed to biomedical signal processing for clinical decision-making and the development of brain-computer interfaces. despite much progress, there are still gaps between technical solutions and real-world applications. the main challenges include:
• interpretability and explainability: as neural models become more complex and layered, understanding the reasoning behind their decisions becomes difficult. developing techniques and methodologies to interpret and explain the output of computational models in the context of biomedical signal processing is essential for gaining trust and acceptance in the medical community.
• limited data and noisy signals: biomedical data are typically limited, for example, small sample sizes, class imbalance, and noisy signals. data scarcity hinders the training and run-time inference of neural networks. effective strategies for limited training data and noisy signals are crucial to ensure robust and accurate signal processing results.
• generalization and adaptability: biomedical signals can exhibit significant variations across individuals, populations, and clinical conditions. neural networks need to generalize well and adapt to different signal characteristics, ensuring their effectiveness in diverse medical scenarios. developing techniques for transfer learning, domain adaptation, and model personalization can enhance the generalization and adaptability of computational models in biomedical signal processing.
• model efficiency and real-time implementation: the computational complexity of conventional neural networks and their requirement for specialized hardware, such as gpus, pose challenges to their applicability. real-time processing of biomedical signals with limited computational resources calls for efficient algorithms, hardware acceleration techniques, and optimized architectures.
overcoming these challenges, researchers have turned to the natural world for inspiration, drawing insights from biological systems to develop brain-inspired neural networks. by mimicking the principles governing the human brain’s information processing, these brain-inspired neural networks exhibit unique advantages, offering novel solutions to the aforementioned issues. this special issue invites original and unpublished research contributions on related topics, but not limited to,
• bio-inspired neural network architectures for biomedical signal processing• deep learning approaches for classification and analysis of biomedical signals• theory, learning algorithms, and computational models of brain-inspired neural networks• interpretable and explainable bio-inspired models for biomedical signal analysis• real-time implementation and hardware acceleration of bio-inspired neural networks• brain-computer interfaces for perception, learning, and motor control• applications of bio-inspired neural networks in clinical decision support systems
guest editors:
dr. siqi cai (executive guest editor)
national university of singapore, singapore
email: elesiqi@nus.edu.sg
areas of expertise: neuroscience, neural network, brain-computer interface (bci), machine learning
prof. dr.-ing. tanja schultz
university of bremen, germany
email: tanja.schultz@uni-bremen.de
areas of expertise: speech recognition, biosignals, silent speech, human-machine interfaces, brain-computer interfaces
prof. haizhou li
the chinese university of hong kong, shenzhen (cuhk-shenzhen), china
email: haizhouli@cuhk.edu.cn
areas of expertise: automatic speech recognition, speaker recognition, language recognition, voice conversion, machine translation
manuscript submission information:
important dates:
• manuscript submission due: april 15, 2024
• first review completed: july 15, 2024
• revised manuscript due: october 15, 2024
• second review completed: december 15, 2024
• final decision notification: february 15, 2025
• publication: april 2025