Control of assistive devices such as prostheses, exoskeletons, and neuroprostheses via Bio-signals (electromyography (EMG), tactile myography (TMG), electroencephalography (EEG), etc.) is a key field of research to improve the quality of life of people with disabilities, and also a promising research avenue in the field of function restoration for individuals affected by movement disorders. As well, movement analysis based on similar signals as well as kinematics signals plays an important role to understand how diseases progress and also to improve the diagnosis and treatment of people with movement disorders. In addition, disease detection via bio-signals and/or the diagnosis of movement disorders (abnormality) have been investigated in the literature.
Deep learning methods have been proposed and rapidly explored in healthcare to perform a variety of tasks. In particular, they can play an important role as an assistive diagnostic tool or in the control of devices, due to their potential to achieve high performance in classification, regression, and prediction tasks. Deep learning models such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Stacked Autoencoders (SAE) or deep reinforcement learning, offer a mean of performing prostheses control with bio-signals and EEG signals, and movement analysis. Deep generative models can help to augment data in prosthesis control applications. However, widespread and practical implementation is still limited and needs to be further investigated.
The main goal of this Research Topic is to collect articles with the latest advances of deep learning methods in the field of movement analysis and prosthetic control. More specifically, research is needed in areas such as fusion of multimodal signals, interpretability of deep models and neuroprosthetics, and motor prostheses device control, as the technology providing high-resolution and high-speed bio-signals is rapidly progressing.
Submissions of both original research and review manuscripts are welcomed to the current Research Topic on “Current trends in Deep Learning for Movement analysis and Prostheses Control” relevant to the following topics including, but not limited to:
1- Deep control of upper-limb prostheses: Current trends in feature extraction and learning.
2- Biosignals analysis – e.g., TMG, Ultrasound, EEG, MEG, High density (HD) EMG, etc. – using deep learning methods.
3- Motor prosthesis and neuroprosthetics control with deep learning methods, e.g. CNN, or with novel simplified structures, e.g. compactCNN, convLSTM, or deep reinforcement learning, etc.
4- Exploration of the effects of external clinical and physiological factors on deep learning models, and the clinical and user acceptance of these techniques.
5- Fusion of multimodal signals.
6- Interpretability of deep models.
7- Movement kinematics analysis with deep learning methods.
8- Diagnosis and prediction of neuromuscular diseases and movement disorders from biosignals using deep learning.
9- Developing DL (deep learning) approaches suited to embedded systems.
10- Open science, open data, reproducible research, and standardized benchmarking solutions in deep learning for movement analysis and control.
Control of assistive devices such as prostheses, exoskeletons, and neuroprostheses via Bio-signals (electromyography (EMG), tactile myography (TMG), electroencephalography (EEG), etc.) is a key field of research to improve the quality of life of people with disabilities, and also a promising research avenue in the field of function restoration for individuals affected by movement disorders. As well, movement analysis based on similar signals as well as kinematics signals plays an important role to understand how diseases progress and also to improve the diagnosis and treatment of people with movement disorders. In addition, disease detection via bio-signals and/or the diagnosis of movement disorders (abnormality) have been investigated in the literature.
Deep learning methods have been proposed and rapidly explored in healthcare to perform a variety of tasks. In particular, they can play an important role as an assistive diagnostic tool or in the control of devices, due to their potential to achieve high performance in classification, regression, and prediction tasks. Deep learning models such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Stacked Autoencoders (SAE) or deep reinforcement learning, offer a mean of performing prostheses control with bio-signals and EEG signals, and movement analysis. Deep generative models can help to augment data in prosthesis control applications. However, widespread and practical implementation is still limited and needs to be further investigated.
The main goal of this Research Topic is to collect articles with the latest advances of deep learning methods in the field of movement analysis and prosthetic control. More specifically, research is needed in areas such as fusion of multimodal signals, interpretability of deep models and neuroprosthetics, and motor prostheses device control, as the technology providing high-resolution and high-speed bio-signals is rapidly progressing.
Submissions of both original research and review manuscripts are welcomed to the current Research Topic on “Current trends in Deep Learning for Movement analysis and Prostheses Control” relevant to the following topics including, but not limited to:
1- Deep control of upper-limb prostheses: Current trends in feature extraction and learning.
2- Biosignals analysis – e.g., TMG, Ultrasound, EEG, MEG, High density (HD) EMG, etc. – using deep learning methods.
3- Motor prosthesis and neuroprosthetics control with deep learning methods, e.g. CNN, or with novel simplified structures, e.g. compactCNN, convLSTM, or deep reinforcement learning, etc.
4- Exploration of the effects of external clinical and physiological factors on deep learning models, and the clinical and user acceptance of these techniques.
5- Fusion of multimodal signals.
6- Interpretability of deep models.
7- Movement kinematics analysis with deep learning methods.
8- Diagnosis and prediction of neuromuscular diseases and movement disorders from biosignals using deep learning.
9- Developing DL (deep learning) approaches suited to embedded systems.
10- Open science, open data, reproducible research, and standardized benchmarking solutions in deep learning for movement analysis and control.
