arXiv subjects

Elodie Germani

Publications and source records attributed to Elodie Germani.

11 recordsLinked to original sources

arXiv2026

Foundational values for foundation models

Research values, properties with a distinctive normative dimension, often affect how technological research is performed in both direct and indirect ways by influencing how technical decisions are made. In machine learning for medical imaging, understanding these values can be important for understanding why particular researchers justify the decisions made in their publications and explain why certain technologies become ubiquitous (or not) in the scientific literature and in the clinic. This article explores one of these technologies, foundation models, finding detailed justifications both for their use and abstention from their use. By taking a Socratic approach to research values arising from this specific technical decision, this article aims to better illustrate how foundation models fit into the philosophy of machine learning in medicine.

cs.CY↗

arXiv2026

Multi-scale radiomics in pelvic MRI for endometriosis subtyping: highlighting data heterogeneity constraints

Analyzing female pelvic MRIs is challenging, especially for evaluating endometriosis, where visual features are influenced by several factors, including anatomical complexity, technical variability, and inter-reader variability. Here, we evaluate a radiomics-based pipeline for patient-level endometriosis subtyping using the publicly available UT-EndoMRI dataset. We extract radiomics features from manually segmented uterine and ovarian regions and compare several multi-scale feature representations and feature-selection strategies. We train supervised classifiers to distinguish patients with at least one endometrioma from those without, and perform an unsupervised perturbation analysis to assess whether radiomics profiles reveal reproducible patient subgroups. The best supervised performance is achieved using raw Wavelet-derived features and a Gradient Boosting classifier, yielding an AUC of 0.80. However, this model produces several false positives, resulting in low specificity. ComBat harmonization does not consistently improve performance, suggesting that post hoc harmonization is insufficient in a small, multi-site cohort in which acquisition groups contained very few patients. Using an unsupervised clustering analysis, we identify reproducible but poorly separated partitions that remain associated with acquisition variables. Overall, these results suggest that pelvic MRI radiomics contain a preliminary signal for endometriosis subtyping, while highlighting the fragility of radiomics-based subtyping in small, multi-site datasets.

eess.IV↗

arXiv2026

Disentangling Prompt Dependence to Evaluate Segmentation Reliability in Gynecological MRI

Promptable segmentation models (e.g., the Segment Anything Models) enable generalizable, zero-shot segmentation across diverse domains. Although predictions are deterministic for a fixed image-prompt pair, the robustness of these models to variations in user prompts, referred to as prompt dependence, remains underexplored. In safety-critical workflows with substantial inter-user variability, interpretable and informative frameworks are needed to evaluate prompt dependence. In this work, we assess the reliability of promptable segmentation by analyzing and measuring its sensitivity to prompt variability. We introduce the first formulation of prompt dependence that explicitly disentangles prompt ambiguity (inter-user variability) from local sensitivity (interaction imprecision), offering an interpretable view of segmentation robustness. Experiments on two female pelvic MRI datasets for uterus and bladder segmentation reveal a strong negative correlation between both metrics and segmentation performance, highlighting the value of our framework for assessing robustness. The two metrics have low mutual correlation, supporting the disentangled design of our formulation, and provide meaningful indicators of prompt-related failure modes.

cs.CV↗

arXiv2025

MRIQT: Physics-Aware Diffusion Model for Image Quality Transfer in Neonatal Ultra-Low-Field MRI

Portable ultra-low-field MRI (uLF-MRI, 0.064 T) offers accessible neuroimaging for neonatal care but suffers from low signal-to-noise ratio and poor diagnostic quality compared to high-field (HF) MRI. We propose MRIQT, a 3D conditional diffusion framework for image quality transfer (IQT) from uLF to HF MRI. MRIQT combines realistic K-space degradation for physics-consistent uLF simulation, v-prediction with classifier-free guidance for stable image-to-image generation, and an SNR-weighted 3D perceptual loss for anatomical fidelity. The model denoises from a noised uLF input conditioned on the same scan, leveraging volumetric attention-UNet architecture for structure-preserving translation. Trained on a neonatal cohort with diverse pathologies, MRIQT surpasses recent GAN and CNN baselines in PSNR 15.3% with 1.78% over the state of the art, while physicians rated 85% of its outputs as good quality with clear pathology present. MRIQT enables high-fidelity, diffusion-based enhancement of portable ultra-low-field (uLF) MRI for deliable neonatal brain assessment.

cs.CV↗

arXiv2025

Bias and Generalizability of Foundation Models across Datasets in Breast Mammography

Over the past decades, computer-aided diagnosis tools for breast cancer have been developed to enhance screening procedures, yet their clinical adoption remains challenged by data variability and inherent biases. Although foundation models (FMs) have recently demonstrated impressive generalizability and transfer learning capabilities by leveraging vast and diverse datasets, their performance can be undermined by spurious correlations that arise from variations in image quality, labeling uncertainty, and sensitive patient attributes. In this work, we explore the fairness and bias of FMs for breast mammography classification by leveraging a large pool of datasets from diverse sources-including data from underrepresented regions and an in-house dataset. Our extensive experiments show that while modality-specific pre-training of FMs enhances performance, classifiers trained on features from individual datasets fail to generalize across domains. Aggregating datasets improves overall performance, yet does not fully mitigate biases, leading to significant disparities across under-represented subgroups such as extreme breast densities and age groups. Furthermore, while domain-adaptation strategies can reduce these disparities, they often incur a performance trade-off. In contrast, fairness-aware techniques yield more stable and equitable performance across subgroups. These findings underscore the necessity of incorporating rigorous fairness evaluations and mitigation strategies into FM-based models to foster inclusive and generalizable AI.

cs.CV↗

arXiv2025

Predicting Parkinson's disease trajectory using clinical and functional MRI features: a reproduction and replication study

Parkinson's disease (PD) is a common neurodegenerative disorder with a poorly understood physiopathology and no established biomarkers for the diagnosis of early stages and for prediction of disease progression. Several neuroimaging biomarkers have been studied recently, but these are susceptible to several sources of variability related for instance to cohort selection or image analysis. In this context, an evaluation of the robustness of such biomarkers to variations in the data processing workflow is essential. This study is part of a larger project investigating the replicability of potential neuroimaging biomarkers of PD. Here, we attempt to reproduce (re-implementing the experiments with the same data, same method) and replicate (different data and/or method) the models described in [1] to predict individual's PD current state and progression using demographic, clinical and neuroimaging features (fALFF and ReHo extracted from resting-state fMRI). We use the Parkinson's Progression Markers Initiative dataset (PPMI, ppmi-info.org), as in [1] and aim to reproduce the original cohort, imaging features and machine learning models as closely as possible using the information available in the paper and the code. We also investigated methodological variations in cohort selection, feature extraction pipelines and sets of input features. Different criteria were used to evaluate the reproduction and compare the reproduced results with the original ones. Notably, we obtained significantly better than chance performance using the analysis pipeline closest to that in the original study (R2 \> 0), which is consistent with its findings. Moreover, using derived data provided by the authors of the original study, we were able to make an exact reproduction and managed to obtain results that were close to the original ones. The challenges encountered while reproducing and replicating the original work are likely explained by the complexity of neuroimaging studies, in particular in clinical settings. We provide recommendations to further facilitate the reproducibility of such studies in the future.

q-bio.NC↗

arXiv2025

Mitigating analytical variability in fMRI results with style transfer

We propose a novel approach to improve the reproducibility of neuroimaging results by converting statistic maps across different functional MRI pipelines. We make the assumption that pipelines used to compute fMRI statistic maps can be considered as a style component and we propose to use different generative models, among which, Generative Adversarial Networks (GAN) and Diffusion Models (DM) to convert statistic maps across different pipelines. We explore the performance of multiple GAN frameworks, and design a new DM framework for unsupervised multi-domain styletransfer. We constrain the generation of 3D fMRI statistic maps using the latent space of an auxiliary classifier that distinguishes statistic maps from different pipelines and extend traditional sampling techniques used in DM to improve the transition performance. Our experiments demonstrate that our proposed methods aresuccessful: pipelines can indeed be transferred as a style component, providing animportant source of data augmentation for future medical studies.

eess.IV↗

arXiv2024

Uncovering communities of pipelines in the task-fMRI analytical space

Analytical workflows in functional magnetic resonance imaging are highly flexible with limited best practices as to how to choose a pipeline. While it has been shown that the use of different pipelines might lead to different results, there is still a lack of understanding of the factors that drive these differences and of the stability of these differences across contexts. We use community detection algorithms to explore the pipeline space and assess the stability of pipeline relationships across different contexts. We show that there are subsets of pipelines that give similar results, especially those sharing specific parameters (e.g. number of motion regressors, software packages, etc.). Those pipeline-to-pipeline patterns are stable across groups of participants but not across different tasks. By visualizing the differences between communities, we show that the pipeline space is mainly driven by the size of the activation area in the brain and the scale of statistic values in statistic maps.

cs.AI↗

arXiv2024

The HCP multi-pipeline dataset: an opportunity to investigate analytical variability in fMRI data analysis

Results of functional Magnetic Resonance Imaging (fMRI) studies can be impacted by many sources of variability including differences due to: the sampling of the participants, differences in acquisition protocols and material but also due to different analytical choices in the processing of the fMRI data. While variability across participants or across acquisition instruments have been extensively studied in the neuroimaging literature the root causes of analytical variability remain an open question. Here, we share the \textit{HCP multi-pipeline dataset}, including the resulting statistic maps for 24 typical fMRI pipelines on 1,080 participants of the HCP-Young Adults dataset. We share both individual and group results - for 1,000 groups of 50 participants - over 5 motor contrasts. We hope that this large dataset covering a wide range of analysis conditions will provide new opportunities to study analytical variability in fMRI.

q-bio.NC↗

arXiv2024

On the validity of fMRI studies with subject-level data processed through different pipelines

In recent years, the lack of reproducibility of research findings has become an important source of concerns in many scientific fields, including functional Magnetic Resonance Imaging (fMRI). The low statistical power often observed in fMRI studies was identified as one of the leading causes of irreproducibility. The development of data sharing opens up new opportunities to achieve larger sample sizes by reusing existing data. FMRI studies proceed by first preparing subject-level data using a given analysis pipeline and then combining those into a group analysis. Historically the subject-level analysis pipeline was identical for all subjects. As practices evolve towards more data reuse, researchers might want to directly combine subject-level data thata were processed using different pipelines. Here, we investigate the impact of combining subject-level data processed with different pipelines in between-group fMRI studies. We used the HCP Young-Adult dataset (N=1,080 participants) processed with 24 different pipelines. We then performed between-group analyses comparing subject data processed with different pipelines. We worked under the null hypothesis of no differences between subjects and compared the estimated false-positive rates obtained with the nominal rates. We showed that the analytical variability induced by the parameters explored in this dataset increases the false positive rates of studies combining data from different pipelines. We conclude that different processed subject data cannot be combined without taking into account the processing applied on these data.

q-bio.NC↗

arXiv2023

On the benefits of self-taught learning for brain decoding

Context. We study the benefits of using a large public neuroimaging database composed of fMRI statistic maps, in a self-taught learning framework, for improving brain decoding on new tasks. First, we leverage the NeuroVault database to train, on a selection of relevant statistic maps, a convolutional autoencoder to reconstruct these maps. Then, we use this trained encoder to initialize a supervised convolutional neural network to classify tasks or cognitive processes of unseen statistic maps from large collections of the NeuroVault database. Results. We show that such a self-taught learning process always improves the performance of the classifiers but the magnitude of the benefits strongly depends on the number of samples available both for pre-training and finetuning the models and on the complexity of the targeted downstream task. Conclusion. The pre-trained model improves the classification performance and displays more generalizable features, less sensitive to individual differences.

cs.NE↗