Usability and Clinical Applicability of Non-Rigid Registration Methods In Neurodegenerative and Neurological Disease Imaging

Neurodegenerative diseases such as Alzheimer’s and Parkinson’s disease, frontotemporal dementia, and related disorders are characterized by progressive and heterogeneous structural and functional changes in the brain. Neuroimaging techniques, including monomodal and multimodal acquisitions from Magnetic Resonance Imaging (MRI), Positron Emission Tomography (PET), etc., as well as longitudinal acquisitions, play a central role in detecting, monitoring, and quantifying these changes across disease stages. Beyond neurodegeneration, similar imaging strategies are essential for the study of other neurological disorders, including cerebrovascular diseases, demyelinating and inflammatory conditions, brain tumors, etc., where anatomical variability and disease progression pose comparable challenges. In these contexts, accurate image alignment is not only a fundamental prerequisite for atlas-based quantification, population-based studies, biomarker extraction, and longitudinal analysis, but also a key enabling tool that makes such analyses feasible and interpretable by establishing meaningful spatial correspondence across subjects, imaging modalities, and temporal stages.Deformable image registration has become an essential tool for quantifying anatomical variability and disease-related morphological changes in neuroimaging studies. In particular, diffeomorphic registration methods have been widely adopted due to their mathematically correct framework, theoretical guarantees of topology preservation, and ability to model complex, large deformations. However, neurodegenerative processes introduce specific challenges for image registration, including pronounced inter-subject heterogeneity, severe cortical and subcortical atrophy, ventricular enlargement, and localized tissue growth or loss. These pathological alterations can lead to biologically implausible deformations, local misregistrations, or unstable optimization behavior. While a wide range of sophisticated registration algorithms has been proposed in the literature, their usability, robustness, interpretability, and clinical reliability in neurodegenerative and neurological disease imaging contexts remain insufficiently explored and only systematically validated.In many cases, validation is performed under highly controlled or idealized experimental settings, relying on limited datasets, surrogate ground truths, or narrow performance metrics that do not fully capture the complexity, variability, and failure modes encountered in real-world clinical and translational imaging workflows. Despite the potential impact, the relationship between registration performance, biomarker robustness, and clinically meaningful outcomes remains poorly characterized, underscoring the need for systematic evaluations that link technical behavior with real-world clinical relevance. The gap between methodological advances and their practical adaptation is dramatically growing.This Research Topic aims to critically examine the usability and clinical applicability of nonrigid registration methods in neurodegenerative and neurological disease imaging. We seek contributions that go beyond algorithmic novelty to address how registration methods perform in real-world datasets, how they affect clinical interpretation, and how their limitations can be systematically evaluated and mitigated. By highlighting practical challenges, realistic validation frameworks, and best practices, this topic aims to support more transparent and reliable use of non-rigid registration techniques in both research and clinical settings.We welcome submissions addressing, but not limited to, the following themes:• Evaluation of non-rigid registration methods across neurodegenerative and neurological diseases commonly studied with neuroimaging• Usability, robustness, and failure analysis of deformable registration in the presence of pathological anatomical variability (e.g., atrophy, lesions, tissue loss, or abnormal growth)• Impact of registration accuracy on downstream Computational Anatomy and quantitative imaging analyses (e.g., morphometry, cortical thickness, deformation-based metrics, PET quantification)• Longitudinal consistency, bias, and error propagation in registration-based studies of disease progression and treatment recovery• Practical and clinically usable validation strategies, including ground-truth surrogates, uncertainty and plausibility estimates• Comparison of classical, biomechanical, and learning-based registration approaches under real-world imaging conditions• Interpretability, reliability, and clinical trustworthiness of registration outputs for research and clinical decision support• Registration challenges in multimodal neuroimaging (e.g., T1-T2 MRI, MRI–PET, MRI–DTI, structural–functional alignment) across neurological disorders• Practical guidelines, benchmarks, and open datasets for deformable registration in neurodegenerative and neurological imaging Topic editors: Alessia Sarica, Jose Tomás Alcalá Nalvaiz, Sara Castel FEced, Mónica Hernández Source: https://www.frontiersin.org/research-topics/78656/usability-and-clinical-applicability-of-non-rigid-registration-methods-in-neurodegenerative-and-neurological-disease-imaging