Open-source tools for neuroimaging analysis, hippocampal segmentation, and disease progression modelling in Alzheimer’s disease. All repositories are freely available under open-source licences.
Disease Progression Models
preAD_DPM (2025)
Benchmarking suite for parametric Disease Progression Models (DPMs) in pre-clinical and early Alzheimer’s disease. Implements and compares three model families — GRACE, Leaspy and RPDPM — for staging cognitive and biomarker trajectories from ADNI longitudinal data. Outputs: subject-level disease staging, biomarker ordering, model-fit metrics (AIC/BIC, cross-validated RMSE). Languages: Python · R · MATLAB. Licence: MIT.
Repository: github.com/cplatero/preAD_DPM
Related paper: Platero C., Bengoa J., Computer Methods and Programs in Biomedicine, 2025. DOI: 10.1016/j.cmpb.2025.109162 · OA: oa.upm.es/92018, PMID: 41237536
Permanent archive: 10.5281/zenodo.19695261 · RRID: [pending]
Keywords: disease progression model, Alzheimer’s disease, pre-clinical AD, biomarker ordering, GRACE, Leaspy, RPDPM, longitudinal MRI, ADNI, Python, R, MATLAB, MCI, cognitive decline
RPDPM_MCItoDementia (2025)
MATLAB implementation of the Robust Parametric Disease Progression Model (RPDPM) applied to the MCI-to-dementia transition. Models the temporal ordering of neuropsychological and imaging biomarker trajectories using a optimization framework with mixed-effects parametric curves. Outputs: individual disease stage estimates, biomarker progression curves, classification of stable vs. progressive MCI. Language: MATLAB R2023b. Licence: MIT.
Repository: https://github.com/cplatero/RPDPM_MCItoDementia
Related paper: Molina, N., Platero C., et al, International Psychogeriatrics, 2025. DOI:https://doi.org/10.1016/j.inpsyc.2025.100129 · OA: https://oa.upm.es/92640/, PMID: 40803947
Permanent archive: 10.5281/zenodo.19698332· RRID: SCR_028391
Keywords: disease progression model, MCI to dementia, Alzheimer’s disease, biomarker trajectories, mixed-effects model, ADNI, MATLAB, longitudinal, cognitive decline
Predictive Models
twogrsurvana (2020) — 295 downloads
Toolbox comparing two longitudinal approaches for predicting MCI-to-Alzheimer’s conversion using ADNI data: (a) two-group comparison — LDA trained on linear mixed-effects (LME) trajectory residues — and (b) survival analysis — extended Cox proportional hazards model with LME random effects. Input: longitudinal MRI markers (cortical thickness, subcortical volumes via FreeSurfer) and neuropsychological measurements with demographic covariates. Outputs: cross-validated AUC, accuracy, sensitivity and specificity at baseline, 12, 24 and 36 months. Language: MATLAB R2018b. Licence: MIT.
Repository: nitrc.org/projects/twogrsurvana
Related papers: Platero C., Tobar M.C., Journal of Neuroscience Methods, JUL 2020. DOI: 10.1016/j.jneumeth.2020.108698 · Platero C., Journal of Neuroscience Methods, MAR 2022. DOI: 10.1016/j.jneumeth.2022.109581 · Platero C., Journal of Alzheimer’s Disease Reports, 2025. DOI: 10.1177/25424823241306097
Permanent archive: 10.5281/zenodo.19755827 · RRID: SCR_028330
Keywords: MCI to AD prediction, survival analysis, Cox model, LDA, linear mixed-effects, FreeSurfer, cortical thickness, longitudinal MRI, ADNI, Alzheimer’s disease, MATLAB, neuropsychological tests, AUC
predict_mci2ad (2018) — 603 downloads
Predictive modelling framework for MCI-to-Alzheimer’s conversion combining MRI-based markers (cortical thickness and subcortical volumes from FreeSurfer) with neuropsychological test scores. Extracts a compact longitudinal feature set discriminating progressive from stable MCI over a 3-year horizon, trained and evaluated on ADNI. Outputs: feature ranking, classification performance (AUC, accuracy, sensitivity, specificity) at multiple time points. Language: MATLAB R2017a. Licence: MIT.
Related paper: Platero C., Tobar M.C., Brain Imaging and Behavior, NOV 2020. DOI: 10.1007/s11682-020-00366-8 · OA: oa.upm.es/91864 · PMID: 33169305
Repository: nitrc.org/projects/predict_mci2ad
Permanent archive: 10.5281/zenodo.20034115 · RRID: SCR_028391
Keywords: MCI to AD conversion, predictive model, multimodal biomarkers, FreeSurfer, cortical thickness, neuropsychological tests, feature selection, longitudinal MRI, ADNI, Alzheimer’s disease, MATLAB
Hippocampal Segmentation
longhippsegm (2019) — 353 downloads
Longitudinal framework for joint hippocampal segmentation and marker extraction from brain MRI at multiple time points. Combines 4D graph cuts with longitudinal atrophy constraints, discriminative appearance modelling (k-NN + filter-bank features), and non-rigid atlas registration (15 atlases, Elastix). Outputs: hippocampal volume and surface roughness per time point. Language: MATLAB R2017b (MEX win64). Licence: MIT.
Repository: nitrc.org/projects/longhippsegm
Related paper: Platero C., Lin L., Tobar M.C., Neuroinformatics, JAN 2019. DOI: 10.1007/s12021-018-9380-2 · OA: oa.upm.es/92042
Permanent archive: https://doi.org/10.5281/zenodo.19819265 · RRID: SCR_016282
Keywords: longitudinal MRI, hippocampal segmentation, graph cuts, label fusion, atlas registration, discriminative model, surface roughness, Elastix, Alzheimer’s disease, ADNI, MATLAB, neuroimaging, MCI, disease progression
lf_patches (2017) — 1,069 downloads
Patch-based label fusion method for hippocampal segmentation cooperating with non-rigid registration. Selected patches and their weights are computed from a combination of intensity-based and labeling-based similarity distances, where a prior labeling of the target image is inferred via non-rigid registration label fusion. Validated on public MR brain image databases against multiple state-of-the-art label fusion methods. Language: MATLAB R2014b+ (MEX win64). Licence: MIT.
Repository: nitrc.org/projects/lf_patches
Related paper: Platero, C. J Neurosci Methods, 2016. https://doi.org/10.1016/j.jneumeth.2016.06.013, https://oa.upm.es/95372/ PMID: 27328371 · Platero C., Tobar M.C., Neuroinformatics, APR 2017. DOI: 10.1007/s12021-017-9323-3 · PMID: 28132187 · OA: oa.upm.es/92041· Platero C., et. al., Human Brain Mapping, 2019. DOI: https://doi.org/10.1002/hbm.24478, https://oa.upm.es/63172/, PMID: 30451343
Permanent archive: https://doi.org/10.5281/zenodo.19881708 · RRID: SCR_028349 (activación pendiente)
Keywords: hippocampal segmentation, patch-based, label fusion, non-rigid registration, Alzheimer’s disease, MRI, ADNI, MATLAB, neuroimaging, similarity measures, CRF
lf_crf (2015) — 374 downloads
Label fusion method based on minimisation of a pseudo-Boolean energy function via graph-cut techniques. Uses a Conditional Random Field (CRF) model incorporating shape, appearance and context information through unary, pairwise and higher-order potentials. Evaluated on two public T1-weighted MR brain image databases for hippocampal segmentation. Languages: MATLAB + C++ (MEX win64). Licence: MIT.
Related paper: Platero C., Tobar M.C., Artificial Intelligence in Medicine, JUN 2015. DOI: 10.1016/j.artmed.2015.04.005 · OA: oa.upm.es/92040
Repository: nitrc.org/projects/lf_crf
Permanent archive: [Zenodo deposit pending] · RRID: [pending]
Keywords: hippocampal segmentation, label fusion, conditional random field, graph cuts, higher-order potentials, MRI, Alzheimer’s disease, MATLAB, C++, shape model, appearance model
Image Pre-processing
NonlinearDiffusion (2013)
Pre-processing toolbox for medical image segmentation using nonlinear (anisotropic) diffusion filters. Implements piecewise-smooth image filtering to enhance boundary contrast while suppressing intra-region noise, serving as a pre-processing step for atlas-based and graph-cut segmentation pipelines. Includes a demo with manual segmentation ground-truth data. Authors: J. Sanguino, C. Platero, O. Velasco. Language: MATLAB. Licence: MIT.
Repository: github.com/cplatero/NonlinearDiffusion
Related paper: Sanguino, J., Platero C., Velasco O., Pre-process for segmentation task with nonlinear diffusion filters, DOI: https://doi.org/10.48550/arXiv.2604.21422. · OA: https://oa.upm.es/95763/
Related thesis: J. Sanguino, Estudio y desarrollo de un preproceso basado en la difusión no lineal para la segmentación en imágenes, EUTI-UPM, 2013. OA: oa.upm.es/14616
Permanent archive: https://doi.org/10.5281/zenodo.19724922· RRID: [pending]
Keywords: nonlinear diffusion, anisotropic filtering, image pre-processing, segmentation, piecewise-smooth, MRI, MATLAB, boundary detection
Liver Segmentation
LiverSegm (2010) — Registered software · Exploitation licence, 1269 downloads
Automatic segmentation framework for liver delineation in abdominal CT and MRI volumes. Combines probabilistic atlas-based segmentation with affine moment descriptors and graph-cut optimisation for multi-atlas label fusion. Validated on public and clinical image datasets; released as a standalone cross-platform tool integrating ITK libraries via MEX objects for MATLAB. Served as the foundation for five undergraduate final projects (PFC/TFG), including the ETSII-UPM Best PFC Award (V. Rodrigo, 2008). Authors: C. Platero, V. Rodrigo, Language: C++, (MEX, ITK). Licence: exploitation rights transferred by formal agreement.
Registered software: M-5643/2010 (Spanish Intellectual Property Registry) · Exploitation licence held
Repository: sourceforge.net/projects/liversegm · Institutional reference: upm.es/recursosidi · LiverSegm
Related papers: Platero C., Tobar M.C., Computational and Mathematical Methods in Medicine, 2014. DOI: 10.1155/2014/182909 · OA: oa.upm.es/94740 · Platero C. et al., CAIP 2011, Springer LNCS. DOI: 10.1007/978-3-642-23672-3_18 · OA: oa.upm.es/13028 · Platero C. et al., ISBI 2008, IEEE. DOI: 10.1109/ISBI.2008.4540920 · OA: oa.upm.es/3326
Permanent archive: oa.upm.es/94740 · RRID: [pending]
Keywords: liver segmentation, abdominal CT, MRI, multi-atlas, graph cuts, affine moment descriptors, ITK, MATLAB, MEX, medical image analysis, registered software
