Clinically translatable applications of machine intelligence (MI) in neurology rely on a complex orchestration of medical procedures and technologies (a combination of computational methods and translational medical devices) to enable actionable insights and informed decision-making in diagnosis and therapeutic interventions. Leveraging the advancement of neuroimaging and computational technologies, recent MI research has demonstrated its promising potential in neurology (e.g., detecting, localizing, segmenting, and assessing brain lesions, tumors, hemorrhagic expansion, traumatic brain injuries, and disease-specific brain network), but has also presented unique challenges in clinical practice that has limited its wider adoption. The challenges include incomprehensible machine decisions due to the increasing complexity of MI, in addition to uncertainty, bias, reliability, and privacy issues in MI models and data. Moreover, few studies examine the ergonomic issues of using MI in assistive medical devices, limiting the efficient use of MI in practice. Other challenges include portability, cost, and energy efficiency in current MI and neuroimaging systems, which hinder clinical access for vulnerable communities.
This Research Topic calls for original research that contributes to a foundational and practical roadmap of machine intelligence and technology in clinical neurology. The research includes computational methods and translational medical devices in neuroimaging acquisition, neural signal processing, neural data analytics using MI, data science, computational intelligence, and artificial intelligence (AI). The areas of research can be experimental, clinical, computational, and algorithmic studies, as well as the innovative development of medical devices and computational hardware. By bridging interdisciplinary research, this Research Topic envisages a translatable, comprehensive, and wider adoption of MI in clinical neurology for diagnosis and treatment.
Research areas of interest include but are not limited to the following:
A) Translatable MI and Computational Methods in Neuroimaging and Neural Signal Processing with Features, such as:
- MI Methods/Frameworks tailored for Neurological Data/Signal
- MI Explainability, Uncertainty, Bias, and Trust
- MI for Precision and Personalized Prevention and Treatment
- Neurological Data Privacy and Security in MI
- Time Series Modelling of Neuroimaging data
- Functional Brain Network Analysis and Applications
B) Innovation in Assistive Medical and Computational Device, to be Used Standalone or with fMRI, fNIRS, EEG, MEG, PET, etc., for Clinical Decision Support. It includes:
- MI System/Interface in Clinical Settings and Environments
- Neuroimaging Analysis Toolbox for Clinicians
- Portable Neuroimaging Devices for Inequities in Remote and Resource-limited Settings.
- Cost and Energy Efficient Neuroimaging Devices for Improving Access to the Vulnerable Communities
- Cost and Energy Efficient Computational Hardware and Architecture for MI-based Neurology (FPGA, GPU, ARM, Custom Chip, Edge, Distributed MI, and others.)
Clinically translatable applications of machine intelligence (MI) in neurology rely on a complex orchestration of medical procedures and technologies (a combination of computational methods and translational medical devices) to enable actionable insights and informed decision-making in diagnosis and therapeutic interventions. Leveraging the advancement of neuroimaging and computational technologies, recent MI research has demonstrated its promising potential in neurology (e.g., detecting, localizing, segmenting, and assessing brain lesions, tumors, hemorrhagic expansion, traumatic brain injuries, and disease-specific brain network), but has also presented unique challenges in clinical practice that has limited its wider adoption. The challenges include incomprehensible machine decisions due to the increasing complexity of MI, in addition to uncertainty, bias, reliability, and privacy issues in MI models and data. Moreover, few studies examine the ergonomic issues of using MI in assistive medical devices, limiting the efficient use of MI in practice. Other challenges include portability, cost, and energy efficiency in current MI and neuroimaging systems, which hinder clinical access for vulnerable communities.
This Research Topic calls for original research that contributes to a foundational and practical roadmap of machine intelligence and technology in clinical neurology. The research includes computational methods and translational medical devices in neuroimaging acquisition, neural signal processing, neural data analytics using MI, data science, computational intelligence, and artificial intelligence (AI). The areas of research can be experimental, clinical, computational, and algorithmic studies, as well as the innovative development of medical devices and computational hardware. By bridging interdisciplinary research, this Research Topic envisages a translatable, comprehensive, and wider adoption of MI in clinical neurology for diagnosis and treatment.
Research areas of interest include but are not limited to the following:
A) Translatable MI and Computational Methods in Neuroimaging and Neural Signal Processing with Features, such as:
- MI Methods/Frameworks tailored for Neurological Data/Signal
- MI Explainability, Uncertainty, Bias, and Trust
- MI for Precision and Personalized Prevention and Treatment
- Neurological Data Privacy and Security in MI
- Time Series Modelling of Neuroimaging data
- Functional Brain Network Analysis and Applications
B) Innovation in Assistive Medical and Computational Device, to be Used Standalone or with fMRI, fNIRS, EEG, MEG, PET, etc., for Clinical Decision Support. It includes:
- MI System/Interface in Clinical Settings and Environments
- Neuroimaging Analysis Toolbox for Clinicians
- Portable Neuroimaging Devices for Inequities in Remote and Resource-limited Settings.
- Cost and Energy Efficient Neuroimaging Devices for Improving Access to the Vulnerable Communities
- Cost and Energy Efficient Computational Hardware and Architecture for MI-based Neurology (FPGA, GPU, ARM, Custom Chip, Edge, Distributed MI, and others.)
