arXiv subjects

Martin Guillemaud

Publications and source records attributed to Martin Guillemaud.

7 recordsLinked to original sources

arXiv2026

Graph-based analysis of inflammatory profiles in New Onset Refractory Status Epilepticus (NORSE)

Background and Objectives: Cryptogenic new-onset refractory status epilepticus (cNORSE) represents one of the most severe forms of status epilepticus, occurring in patients without prior neurological disease, and remaining of unknown aetiology despite extensive diagnostic evaluation. Emerging evidence supports a role for immune dysregulation in cNORSE; however, marked heterogeneity in inflammatory signatures has been reported, complicating the selection of targeted immunotherapies. Therefore, a critical need for tools facilitating the interpretation of cytokine panels exists. Methods: Building on the identification of distinct inflammatory groups of cNORSE patients using a graph clustering approach applied to a cohort of 62 patients with serum profiling of 96 cytokines, we tailored new models to quantify attribution probability to biologically validated clusters. Statistical assessment of the most informative model involved Monte-Carlo simulations and custom-developed parametric tests. Ultimately, we applied our framework to the implementation of a clinician-friendly interface for inflammatory profiling. Results: Our approach enables quick processing of several cytokine profiles, providing the most likely inflammatory cluster, associated attribution probability, and statistical confidence. For longitudinal assessments, the proposed method may also allow tracking the evolution of inflammatory trajectories over time. Conclusion: Systematic statistical characterization of the inflammatory heterogeneity in cNORSE requires the development of clinically actionable support tools. Our study offers a framework that may support personalized immunomodulatory strategies in cNORSE patients through clustering-based cytokine profiling.

q-bio.NC↗

arXiv2025

Hyperbolic embedding of multilayer networks

Multilayer networks offer a powerful framework for modeling complex systems across diverse domains, effectively capturing multiple types of connections and interdependent subsystems commonly found in real world scenarios. To analyze these networks, embedding techniques that project nodes into a lower-dimensional geometric space are essential. This paper introduces a novel hyperbolic embedding framework that advances the state of the art in multilayer network analysis. Our method, which supports heterogeneous node sets across networks and inter-layer connections, generates layer-specific hyperbolic embeddings, enabling detailed intra-layer analysis and inter-layer comparisons, while simultaneously preserving the global multilayer structure within hyperbolic space, a capability that sets it apart from existing approaches, which typically rely on independent embedding of layers. Through experiments on synthetic multilayer stochastic block models, we demonstrate that our approach effectively preserves community structure, even when layers consist of different node sets. When applied to real brain networks, the method successfully clusters disease-related brain regions from different patients, outperforming layer-independent approaches and highlighting its relevance for comparative analysis. Overall, this work provides a robust tool for multilayer network analysis, enhancing interpretability and offering new insights into the structure and function of complex systems.

cs.SI↗

arXiv2025

Low-dimensional representation of brain networks for seizure risk forecasting

Identifying preictal states -- periods during which seizures are more likely to occur -- remains a central challenge in clinical computational neuroscience. In this study, we introduce a novel framework that embeds functional brain connectivity networks, derived from intracranial EEG (iEEG) recordings, into a low-dimensional Euclidean space. This compact representation captures essential topological features of brain dynamics and facilitates the detection of subtle connectivity changes preceding seizures. Using standard machine learning techniques, we define a dimensionless biomarker, $\mathcal{B}$, that discriminates between interictal (seizure-free) and preictal (within 24 hours of seizure) network states. Our method focuses on connectivity patterns among a subset of informative iEEG electrodes, identified through permutation-based testing, enabling robust classification of brain states across time. We validate our approach using a leave-one-out cross-validation scheme and a pseudo-prospective forecasting strategy, assessing performance with metrics such as F1-score and balanced accuracy. Results show that low-dimensional Euclidean embeddings of iEEG connectivity yield interpretable and predictive markers of preictal activity, offering promising implications for real-time seizure forecasting and individualized therapeutic interventions.

q-bio.NC↗

arXiv2025

Hyperbolic embedding of brain networks as a tool for epileptic seizures forecasting

The evidence indicates that intracranial EEG connectivity, as estimated from daily resting state recordings from epileptic patients, may be capable of identifying preictal states. In this study, we employed hyperbolic embedding of brain networks to capture non-trivial patterns that discriminate between connectivity networks from days with (preictal) and without (interictal) seizure. A statistical model was constructed by combining hyperbolic geometry and machine learning tools, which allowed for the estimation of the probability of an upcoming seizure. The results demonstrated that representing brain networks in a hyperbolic space enabled an accurate discrimination (85%) between interictal (no-seizure) and preictal (seizure within the next 24 hours) states. The proposed method also demonstrated excellent prediction performances, with an overall accuracy of 87% and an F1-score of 89% (mean Brier score and Brier skill score of 0.12 and 0.37, respectively). In conclusion, our findings indicate that representations of brain connectivity in a latent geometry space can reveal a daily and reliable signature of the upcoming seizure(s), thus providing a promising biomarker for seizure forecasting.

q-bio.NC↗

arXiv2025

Hyperbolic embedding of brain networks detects regions disrupted by neurodegeneration in Alzheimer's disease

Graph theoretical methods have proven valuable for investigating alterations in both anatomical and functional brain connectivity networks during Alzheimer's disease (AD). Recent studies suggest that representing brain networks in a suitable geometric space can better capture their connectivity structure. This study introduces a novel approach to characterize brain connectivity changes using low dimensional, informative representations of networks in a latent geometric space. Specifically, the networks are embedded in a polar representation of the hyperbolic plane, the hyperbolic disk. Here, we use a geometric score, entirely based on the computation of distances between nodes in the latent space, to measure the effect of a perturbation on the nodes. Precisely, the score is a local measure of distortion in the geometric neighborhood of a node following a perturbation. The method is applied to a brain network dataset of patients with AD and healthy participants, derived from diffusion weighted (DWI) and functional (fMRI) magnetic resonance imaging scans. We show that, compared with standard graph measures, our method more accurately identifies the brain regions most affected by neurodegeneration. Notably, the abnormalities detected in memory related and frontal areas are robust across multiple brain parcellation scales. Finally, our findings suggest that the geometric perturbation score could serve as a potential biomarker for characterizing the progression of the disease.

q-bio.NC↗

arXiv2024

Geometric representations of brain networks can predict the surgery outcome in temporal lobe epilepsy

Epilepsy surgery, particularly for temporal lobe epilepsy (TLE), remains a vital treatment option for patients with drug-resistant seizures. However, accurately predicting surgical outcomes remains a significant challenge. This study introduces a novel biomarker derived from brain connectivity, analyzed using non-Euclidean network geometry, to predict the surgery outcome in TLE. Using structural and diffusion magnetic resonance imaging (MRI) data from 51 patients, we examined differences in structural connectivity networks associated to surgical outcomes. Our approach uniquely utilized hyperbolic embeddings of pre- and post-surgery brain networks, successfully distinguishing patients with favorable outcomes from those with poor outcomes. Notably, the method identified regions in the contralateral hemisphere relative to the epileptogenic zone, whose connectivity patterns emerged as a potential biomarker for favorable surgical outcomes. The prediction model achieves an area under the curve (AUC) of 0.87 and a balanced accuracy of 0.81. These results underscore the predictive capability of our model and its effectiveness in individual outcome forecasting based on structural network changes. Our findings highlight the value of non-Euclidean representation of brain networks in gaining deeper insights into connectivity alterations in epilepsy, and advancing personalized prediction of surgical outcomes in TLE.

q-bio.NC↗

arXiv2024

Detecting local perturbations of networks in a latent hyperbolic embedding space

Graph theoretical approaches have been proven to be effective in the characterization of connected systems, as well as in quantifying their dysfunction due to perturbation. In this paper, we show the advantage of a non-Euclidean (hyperbolic) representation of networks to identify local connectivity perturbations and to characterize the induced effects on a large scale. We propose two perturbation scores based on representations of the networks in a latent geometric space, obtained through an embedding onto the hyperbolic Poincaré disk. We numerically demonstrate that these methods are able to localize perturbations in networks with homogeneous or heterogeneous degree connectivity. We apply this framework to identify the most perturbed brain areas in epileptic patients following surgery. This study is conceived in the effort of developing more powerful tools to represent and analyze brain networks, and it is the first to apply geometric network embedding techniques to the case of epilepsy.

q-bio.QM↗