Many medical practices can be regarded as decision-making in the medical field. Nowadays, computers have evolved as crucial components in biomedical decision-making. Still, the general perception of “computers in medicine” is often only of computer applications that assist physicians in the diagnosis of illnesses. To integrate computers more closely into the biomedical fields, researchers and practitioners are increasingly relying on knowledge-based deep learning systems (KDLS) in medicine and their underlying technologies in the field of medicine, particularly neuroimaging (NI).

The medicinal data can be obtained from multiple imaging modalities, such as Computed Tomography (CT), Magnetic Resonance (MR) Imaging, Ultrasound, Single Photon Emission Computed Tomography (SPECT), Positron Emission Tomography (PET), Magnetic Particle Imaging (MPI), EE/MEG, Optical Microscopy and Tomography, Photoacoustic Tomography, Electron Tomography, and Atomic Force Microscopy, etc. Analysis of such modalities has been traditionally conducted with classical statistics, either hypothesis testing or Bayesian inference, that relies on frequently violated assumptions. One promising solution is machine learning (ML) within knowledge-based deep learning systems, where high-dimensional relationships between datasets are empirically established.

This special issue aims to report the newest methodological developments of KDLS to assess functional connectivity, applications of the neurological disorder application fields, and clinical neuroscience, e.g., Alzheimer’s, Parkinson’s, strokes, brain tumors, epilepsy, multiple sclerosis, ALS, Autism, etc. Furthermore, the explanations for black-box predictions with ML methods to develop KDLS in brain diseases and conditions are the main goal of the special issue.

Topics of interest include, but are not limited to:
 Theory, models, frameworks, and tools of KDLS in medicine
 Advanced statistical inference from clinical trials based on KDLS: Symbolic inference based on KDLS models
 Validity and reproducibility of biomedical research methods: Trustworthy KDLS models against the replication crisis
 Advanced statistical paradigms for KDLS models
 KDLS with autonomous deep learning, automated reasoning, and reinforcement learning
 Deep graph networks for KDLS
 Attention mechanism and explainability in KDLS models
 KDLS-enabled natural language processing and understanding
 Security and privacy-related KDLS systems
 KDLS in robotics, social science, and healthcare
 Big data in KDLS systems and their applications
 KDLS-based question answering systems and recommendation systems
 KDLS for human-machine interaction and collaborations