Applications based on Explainable Artificial Intelligence (XAI) have proliferated exponentially in recent years across a wide range of diagnostic domains. In particular, Explainable Deep Learning (XDL) based algorithms have been improved to uncover reliable biomarkers for increasing brain illnesses, analyze the correlations between various factors, and extract patterns in high-dimensional clinical and diagnostic imaging datasets. Several powerful techniques have been proposed to map morphological, structural, and functional neuroimages to neurological diseases. XDL has recently been used to analyze neuroimages and is a powerful tool to improve our understanding of neurological diseases. However, for these algorithms to be successfully applied in the diagnostic domain, several issues must be resolved, including the blending and harmonization of high-dimensional datasets, the robust validation of algorithms' performance, the generalization of methods created using various datasets, and the clinical interpretability of algorithmic decisions. XDL approaches enable non-experts to use machine learning effectively because, in contrast to typical machine learning algorithms, they can recognize abstract and complicated patterns and automatically discover informative representations without requiring domain expertise.
XDL has recently been applied to the analysis of neuroimages, such as structural magnetic resonance imaging (MRI), functional MRI, and positron emission tomography (PET). It has produced notable performance gains over conventional machine learning computer-aided diagnosis of brain disorders and diseases. We will compile recent developments in machine learning and deep learning methods for detecting neurodegenerative disorders using imaging, genomic, and clinical data in this Research Topic. We support initiatives and methods that attempt to improve the application of algorithms in diagnostic practice. This Research Topic aims to investigate how XDL methods are being developed to analyze neurological diseases based on neuroimaging. We are looking for studies spanning a range of specializations that apply and build XDL methods for neuroimage processing and explore the relationship between neuroimaging and neurological illnesses. Potential topics include but are not limited to the following:
• Explainable deep learning (XDL) algorithms for the classification and diagnosis of neurodegenerative diseases
• Machine learning/deep learning models for subtyping neurodegenerative diseases
• Harmonization techniques for machine/deep learning
• Multimodal data integration for machine/deep learning models
• XDL based causal modeling for neurological diseases
• XDL based multimodal data fusion to characterize neurological disorders
• Novel deep learning algorithms and pipeline design for neuroimage analysis
• XDL applications for the diagnosis of neurological disorders
• Brain biomarker discovery using deep learning
• Deep learning regression models for prognostic value prediction in neurology
Applications based on Explainable Artificial Intelligence (XAI) have proliferated exponentially in recent years across a wide range of diagnostic domains. In particular, Explainable Deep Learning (XDL) based algorithms have been improved to uncover reliable biomarkers for increasing brain illnesses, analyze the correlations between various factors, and extract patterns in high-dimensional clinical and diagnostic imaging datasets. Several powerful techniques have been proposed to map morphological, structural, and functional neuroimages to neurological diseases. XDL has recently been used to analyze neuroimages and is a powerful tool to improve our understanding of neurological diseases. However, for these algorithms to be successfully applied in the diagnostic domain, several issues must be resolved, including the blending and harmonization of high-dimensional datasets, the robust validation of algorithms' performance, the generalization of methods created using various datasets, and the clinical interpretability of algorithmic decisions. XDL approaches enable non-experts to use machine learning effectively because, in contrast to typical machine learning algorithms, they can recognize abstract and complicated patterns and automatically discover informative representations without requiring domain expertise.
XDL has recently been applied to the analysis of neuroimages, such as structural magnetic resonance imaging (MRI), functional MRI, and positron emission tomography (PET). It has produced notable performance gains over conventional machine learning computer-aided diagnosis of brain disorders and diseases. We will compile recent developments in machine learning and deep learning methods for detecting neurodegenerative disorders using imaging, genomic, and clinical data in this Research Topic. We support initiatives and methods that attempt to improve the application of algorithms in diagnostic practice. This Research Topic aims to investigate how XDL methods are being developed to analyze neurological diseases based on neuroimaging. We are looking for studies spanning a range of specializations that apply and build XDL methods for neuroimage processing and explore the relationship between neuroimaging and neurological illnesses. Potential topics include but are not limited to the following:
• Explainable deep learning (XDL) algorithms for the classification and diagnosis of neurodegenerative diseases
• Machine learning/deep learning models for subtyping neurodegenerative diseases
• Harmonization techniques for machine/deep learning
• Multimodal data integration for machine/deep learning models
• XDL based causal modeling for neurological diseases
• XDL based multimodal data fusion to characterize neurological disorders
• Novel deep learning algorithms and pipeline design for neuroimage analysis
• XDL applications for the diagnosis of neurological disorders
• Brain biomarker discovery using deep learning
• Deep learning regression models for prognostic value prediction in neurology
