cs.CVAug 18, 2026

A Configurable Privacy-Preserving MRI Processing Workflow Using Deep Learning-Based Brain Extraction and Adaptive Anatomical Preservation

Authors: Rayeef Ali Khan, Komal Raj Mahantesh

Organizations: School of Computing (Artificial Intelligence) Dublin City University Dublin, Ireland · School of Computing (Data Analytics) Dublin City University Dublin, Ireland

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Structural Magnetic Resonance Imaging (MRI) is widely used in neuroimaging research and clinical practice, but structural MRI volumes may retain facial and cranial anatomical information that raises privacy concerns. Existing deep learning-based brain extraction methods generally produce a single fixed output, limiting flexibility when different applications require different balances between privacy and anatomical preservation. This paper presents a configurable privacy-preserving MRI processing workflow that extends deep learning-based brain extraction through adaptive anatomical preservation, interactive preservation selection, and integrated quality control. The workflow employs SynthStrip for automated brain extraction, followed by morphological mask expansion to generate configurable shell-based preservation levels. An Interactive Preservation Framework enables users to compare preservation configurations and select an appropriate output, while an integrated Quality Control Framework provides multi-plane visualisation and brain-mask overlay verification. The workflow was implemented in Python using open-source neuroimaging libraries within the Renku reproducible research environment and evaluated using structural T1-weighted MRI data from the publicly available IXI dataset. Experimental results demonstrate anatomically plausible brain extraction and configurable preservation outputs, supported by systematic visual verification. The principal contribution is a modular and reproducible MRI preprocessing framework that enhances deep learning-based brain extraction with configurable anatomical preservation, interactive user-guided processing, and integrated quality control. The workflow provides a practical foundation for privacy-oriented neuroimaging research and collaborative medical image analysis.

Figures & tables

MethodDLFOAPIPSIQC
BET×\times✓×\times×\timesL
FreeSurfer×\times✓×\times×\timesL
SynthStrip✓✓×\times×\timesL
Proposed✓✓✓✓✓
TABLE I: Comparison of Representative Brain Extraction Approaches
Fig. 1: Overview of the proposed privacy-preserving MRI processing workflow. The pipeline begins with a raw structural T1-weighted MRI containing facial structures, followed by deep learning-based brain extraction using SynthStrip, shell-based adaptive preservation, interactive preservation selection, integrated quality control, and generation of a privacy-preserving brain-extracted MRI suitable for secure sharing and downstream neuroimaging analysis.
Fig. 2: Deep learning-based brain extraction using SynthStrip. (a) Original structural T1-weighted MRI, (b) generated binary brain mask, (c) brain mask overlaid on the original MRI for segmentation verification, and (d) resulting brain-extracted MRI. The extracted brain mask serves as the input to the proposed shell-based adaptive preservation framework.
Fig. 3: Shell-based adaptive anatomical preservation generated through morphological expansion of the SynthStrip brain mask. The one-shell (approximately 1 mm) configuration preserves a minimal anatomical boundary surrounding the extracted brain, while the two-shell (approximately 2 mm) configuration retains additional peripheral anatomical context. The coloured outlines highlight the incremental preservation achieved through successive shell expansion.
Fig. 4: Interactive Preservation Framework enabling comparison of multiple preservation configurations generated from a single brain extraction result. Researchers can visually evaluate preservation levels and select the configuration that best satisfies the privacy and analytical requirements of a particular neuroimaging study.
Fig. 5: Quality Control Framework showing representative axial, coronal, and sagittal visualisations with brain mask overlay verification. Integrated quality-control procedures enable systematic inspection of brain extraction accuracy and preservation behaviour across processed structural MRI volumes.
ComponentSpecification
DatasetIXI Dataset
MRI ModalityStructural T1-weighted MRI
Brain ExtractionSynthStrip
Programming LanguagePython
Image ProcessingNiBabel, NumPy, SciPy
VisualisationMatplotlib
TABLE II: Implementation Details of the Proposed Workflow
MetricValue (%)
Dice Similarity Coefficient (DSC)98.73
Intersection over Union (IoU)97.48
Precision99.05
Recall98.41
F1 Score98.73
TABLE III: Quantitative Performance of Brain Extraction
Preservation LevelAnatomical ContextPrivacy ProtectionExample Use Case
One-Shell   
( ≈\approx 1 mm)Minimal boundary preservationHighMaximum anonymisation and secure data sharing
Two-Shell   
( ≈\approx 2 mm)Additional peripheral anatomical contextModerateResearch requiring greater anatomical context
TABLE IV: Conceptual Comparison of Adaptive Preservation Levels
Fig. 6: Representative quality-control outputs generated by the proposed workflow, showing multi-plane visualisation and brain mask overlays for verification of brain extraction and adaptive preservation.
More in this topic

Neuroimaging

Neuroimaging uses brain measurements and computational methods to map anatomy, activity, connectivity, and disease.

See all 235 papers

Explore similar work

CardsList
Jun 6, 2026

How Much MRI Preprocessing Is Enough? A Cost-Utility Study for Brain MRI Foundation Models

MRI preprocessing defines the input distribution seen by brain MRI foundation models, yet it is usually treated as routine data cleaning rather than a modeling choice. We ask how much preprocessing is worth its computational cost for self-supervised 3D MRI pretraining. Keeping the corpus, 3D ViT backbone, masking protocol, and downstream evaluations fixed, we compare a graded P0-P7 preprocessing spectrum for masked autoencoding (MAE) and joint-embedding predictive learning (JEPA) on 20,000 heterogeneous brain MRI volumes, then transfer the encoders to IDH prediction, MCI classification, brain age regression, and GLI/PED tumor segmentation. The results do not support a simple "more is better" rule. P0/P1 are numerically unstable, making P2 the lowest-cost feasible level; beyond P2, choosing the best feasible preprocessing level improves aggregate utility by only 3.4 percentage points for MAE and 1.8 percentage points for JEPA, with most paired gains statistically unresolved. Stronger preprocessing is beneficial only in selected regimes: IDH improves modestly, AGE and GLI/PED are often near or best at P2, and MCI shows the clearest empirical P7 gain. Cross-level MCI transfer further shows that much of the P7 advantage can be recovered by applying stronger preprocessing downstream, without requiring P7 throughout pretraining. These findings recast MRI preprocessing as a downstream-aware cost-utility decision rather than a default escalation pipeline. Code is available at https://github.com/PangJiangShuan/PreBrain.

Jiangshuan Pang, Wangyang Tang, Jing Yan +5

School of Artificial Intelligence, University of the Chinese Academy of Sciences, Beijing 101408, China · Key Laboratory of Spatial Omics of Zhejiang Province, BGI Research, Hangzhou 310030, China · Key Laboratory of Brain Cell Mapping of Zhejiang Province, BGI Research, Hangzhou 310030, China

Medical Imaging Foundation ModelsSelf-Supervised Pre-TrainingBrain MRI
AskPDF
Oct 14, 2026

Quantifying privacy risks in foundation model representations of brain MRI

Pretrained representations are increasingly used to share and reuse information from medical images without exposing the original scans, but it is unclear how much patient-specific information these representations retain. We investigate this question for brain magnetic resonance imaging (MRI) by evaluating five frozen pretrained encoders spanning different architectures, pretraining objectives, and training domains on two longitudinal T1-weighted datasets, OASIS-2 and MIRIAD. Under an attacker model with access to released embeddings, the corresponding public encoder, and auxiliary MRI scans, we evaluate patient linkage and sensitive-attribute inference. Across the evaluated representations, we observe substantial subject-specific information: rank-1 patient linkage reaches up to 99% on MIRIAD and exceeds 90% on OASIS-2 for the highest-risk representation, despite the absence of an identification objective during encoder training. We further show that the released embeddings retain information about a demographic attribute, with recorded-sex prediction reaching balanced accuracy of 90.9% on OASIS-2 and 95.5% on MIRIAD. Overall, our results demonstrate that replacing raw brain MRI with learned embeddings does not, by itself, eliminate patient-linkage or attribute-inference risk, and is therefore insufficient on its own to establish anonymity.

Imen Bajar, Héber Hwang Arcolezi, Antoine Boutet +2

École de technologie supérieure (ÉTS), Montréal, QC, Canada · Inria, INSA Lyon, CITI, UR3720, 69621 Villeurbanne, France · CREATIS, CNRS UMR5220, INSERM U1294, Université Lyon, INSA-Lyon, Lyon, France +1

Medical Imaging Foundation ModelsPrivacy-Preserving MLBrain MRIMedical Image Analysis
AskPDF
Sep 6, 2026

Comparative Study of Anatomical and Learned Features in AI Models for Structural Brain MRI

In this work, we comprehensively evaluate three popular feature-extraction paradigms in AI-based neuroimaging modeling: (1) computation of anatomical surfaces and volumes, (2) supervised learning with convolutional neural networks (CNNs), and (3) unsupervised pretraining of vision transformer (ViT) foundation models, followed by supervised finetuning. Our study is based on 18 publicly available datasets containing 3D structural T1-weighted MRI scans from approximately 80,000 participants across seven distinct clinical tasks. We observe that a linear model based on anatomical features matches the diagnostic performance of complex nonlinear features learned by sophisticated AI frameworks, including foundation models trained on thousands of scans. Conversely, CNNs and pretrained ViTs learn features that implicitly capture relevant anatomical information, bypassing the need for explicit feature extraction. Building upon these insights, we propose Anatomy Segmentation Pretraining (ASP), a novel method to incorporate anatomical information during foundation-model pretraining, which outperforms existing models in biological age estimation.

Boyang Yu, Miquel Lopez Escoriza, Long Chen +3

Center for Data Science, New York University, NYU, 60 Fifth Ave, New York, 10011, New York, United States. · Department of Neurology, NYU Grossman School of Medicine, 550 1st Ave, New York, 10016, New York, United States. · Courant Institute of Mathematical Sciences, NYU, 251 Mercer St, New York, 10012, New York, United States.

NeuroimagingMedical Imaging Foundation ModelsBrain MRI
AskPDF