arXiv subjects

Xavier Vasques

Publications and source records attributed to Xavier Vasques.

16 recordsLinked to original sources

arXiv2026

Artificial intelligence in deep brain stimulation for movement disorders: a systematic review and technology readiness assessment

Artificial intelligence (AI) is increasingly explored across deep brain stimulation (DBS) for movement disorders, yet whether current systems are approaching deployment remains unclear. To characterise their scope, validation maturity, and translational readiness, we systematically evaluated 239 peer-reviewed studies published between 2000 and 2025, assessing AI methods, validation practices, and barriers constraining clinical translation. Research was dominated by Parkinson's disease and subthalamic nucleus targeting, with limited coverage of other disorders and targets. Most studies reported encouraging internal performance; however, external validation was rare, evaluations remained predominantly retrospective and single-centre, and more than one-quarter involved small-sample, high-dimensional datasets with elevated overfitting risk. Technology readiness assessment revealed that most systems remain at early-to-intermediate translational stages, constrained more by limited validation than by algorithmic inadequacy, compounded by the biological heterogeneity and dynamic complexity inherent to DBS. Nevertheless, emerging external and prospective studies suggest a field moving toward clinical maturity, with promising applications in targeting, programming, outcome prediction, and adaptive therapy delivery.

q-bio.NC↗

arXiv2026

Subject-Level Heterogeneity in EEG Motor Imagery Decoding: A Large-Scale Benchmark and Portfolio-Based Reduction of the Search Space

Robust EEG motor imagery decoding remains limited by strong inter-individual variability, making it difficult to identify pipelines that generalize across users. We present a large-scale, standardized within-session benchmark of decoding pipelines across three public datasets: Cho2017 (52 subjects), PhysionetMI (109 subjects), and Zhou2016 (4 subjects). Using a common MOABB LeftRightImagery setting, two frequency bands (8-15 Hz and 8-30 Hz), and a broad combination of feature extraction, preprocessing, and classification steps, we analyzed 216,714 raw evaluation rows, which after structured aggregation yielded 44,928, 109,000, and 4,192 subject-level observations respectively. Covariance tangent-space projection (cov-tgsp) and Common Spatial Patterns (CSP) consistently defined the strongest methodological families, though their relative ordering was dataset-dependent. On Cho2017, the best family-level mean accuracy came from cov-tgsp in 8-30 Hz (0.712 +/- 0.140), whereas Zhou2016 favored CSP (0.832 +/- 0.121 in 8-15 Hz). These aggregate rankings concealed substantial subject-level heterogeneity: 42 distinct winning pipelines across 52 Cho2017 subjects, and 93 across 109 PhysionetMI subjects. We then used the benchmark as an empirical performance landscape for building compact portfolios of pipelines of size K. Several construction procedures were compared, including a ranking-based Top-K Mean heuristic and search-based strategies. Results were broadly consistent, with Top-K Mean giving the best trade-off. A single best global pipeline already retained 94.2% of the oracle in Cho2017 and 81.8% in PhysionetMI; at K = 12, oracle retention rose to 96.5% and 90.0%. The landscape is therefore subject-dependent, and this heterogeneity can be exploited through compact portfolios that make personalization more feasible.

q-bio.NC↗

arXiv2026

Average Rankings Mask Per-Subject Optimality: A Friedman-Nemenyi Benchmark of EEG Motor-Imagery BCI Decoders

Electroencephalography (EEG) is the dominant non-invasive modality for brain-computer interfaces (BCIs), yet reliable decoding of motor imagery is hampered by inter- and intra-individual variability. A recurring claim is that one decoding pipeline, most often a spatial or Riemannian method, is broadly preferable. We test the weakest version of that claim under the most favourable conditions. Using the Mother of All BCI Benchmarks (MOABB) framework, we evaluated 1,056 decoding configurations (feature extractor x scaler x classifier), >340,000 subject-level model fits, across three public left-versus-right motor-imagery datasets (PhysionetMI, 109 participants; Cho2017, 52; Zhou2016, 4) and two frequency bands (8-15 Hz, 8-30 Hz). Every model is fit and tested within a single session of a single participant, the easiest regime, giving every pipeline its best chance. We apply the statistics standard for multi-classifier comparison: Friedman omnibus tests, Nemenyi critical-difference analysis and Wilcoxon signed-rank tests with effect sizes. Covariance tangent-space projection (cov-tgsp) and Common Spatial Patterns (CSP) are the strongest families, but their ordering is dataset-dependent and, on the largest and most heterogeneous cohort (PhysionetMI), statistically indistinguishable (Nemenyi p = 0.27; Kendall's W = 0.11). At the individual level the single best pipeline is optimal for only 35% of PhysionetMI participants, and nonlinear descriptors are best for roughly one third; matching pipeline to participant adds about seven accuracy points over the best fixed choice. The ranking is not an artefact of dimensionality, and classifier and scaler choices are secondary to the feature representation. Even in the easiest regime, no single pipeline dominates: a lower bound on the personalization problem and a quantitative case for participant-aware model selection rather than a universal decoder.

cs.HC↗

arXiv2026

Simultaneous hyperkinetic movement disorders phenotyping: a cross-cohort pediatric transfer study using routine videos, markerless pose estimation and a tabular foundation model

Objective: To develop and externally test a video-based framework for simultaneous detection of hyperkinetic MDs phenomenologies: dystonia, tremor, myoclonus, chorea, athetosis, ballismus, stereotypies, and tics using routine clinical recordings, with explicit testing of external, cross-cohort transfer from adult to pediatric populations. Methods: In this proof-of-concept study, the framework combines markerless pose estimation, kinematic descriptors, and a pretrained fondation model. A shared predictive backbone was developed on 21 adults with confirmed hyperkinetic MDs and 4 healthy controls assessed under a standardized protocol. External validation was performed on an independent external cohort: a real-world pediatric sample (n=12, monogenic combined MDs). For the external dataset, the backbone was deployed without retraining; lightweight calibration adjusted only the final subject-level decision step using a small labeled subset of patients selected by clinicians as representative of the cohort's phenotypic range. Results: After local calibration of the decision layer on the clinician-selected subset, performance improved consistently on the held-out pediatric patients (n=7): Hamming accuracy rose from 0.804 to 0.839 and the Jaccard index from 0.548 to 0.633. This calibrated performance was preserved, and the Jaccard index further improved, when the evaluation was restricted to the phenomenologies with more definite clinician agreement (Hamming accuracy 0.9, Jaccard index 0.786), indicating that the gains did not rest on the least-reliable labels.

cs.CV↗

arXiv2026

Deep Learning Pose Estimation for Multi-Label Recognition of Combined Hyperkinetic Movement Disorders

Hyperkinetic movement disorders (HMDs) such as dystonia, tremor, chorea, myoclonus, and tics are disabling motor manifestations across childhood and adulthood. Their fluctuating, intermittent, and frequently co-occurring expressions hinder clinical recognition and longitudinal monitoring, which remain largely subjective and vulnerable to inter-rater variability. Objective and scalable methods to distinguish overlapping HMD phenotypes from routine clinical videos are still lacking. Here, we developed a pose-based machine-learning framework that converts standard outpatient videos into anatomically meaningful keypoint time series and computes kinematic descriptors spanning statistical, temporal, spectral, and higher-order irregularity-complexity features.

cs.CV↗

arXiv2025

Rethinking Generalized BCIs: Benchmarking 340,000+ Unique Algorithmic Configurations for EEG Mental Command Decoding

Robust decoding and classification of brain patterns measured with electroencephalography (EEG) remains a major challenge for real-world (i.e. outside scientific lab and medical facilities) brain-computer interface (BCI) applications due to well documented inter- and intra-participant variability. Here, we present a large-scale benchmark evaluating over 340,000+ unique combinations of spatial and nonlinear EEG classification. Our methodological pipeline consists in combinations of Common Spatial Patterns (CSP), Riemannian geometry, functional connectivity, and fractal- or entropy-based features across three open-access EEG datasets. Unlike prior studies, our analysis operates at the per-participant level and across multiple frequency bands (8-15 Hz and 8-30 Hz), enabling direct assessment of both group-level performance and individual variability. Covariance tangent space projection (cov-tgsp) and CSP consistently achieved the highest average classification accuracies. However, their effectiveness was strongly dataset-dependent, and marked participant-level differences persisted, particularly in the most heterogeneous of the datasets. Importantly, nonlinear methods outperformed spatial approaches for specific individuals, underscoring the need for personalized pipeline selection. Our findings highlight that no universal 'one-size-fits-all' method can optimally decode EEG motor imagery patterns across all users or datasets. Future work will require adaptive, multimodal, and possibly novel approaches to fully address neurophysiological variability in practical BCI applications where the system can automatically adapt to what makes each user unique.

q-bio.NC↗

arXiv2025

Synthetic Data Generation for Classifying Electrophysiological and Morpho-Electrophysiological Neurons from Mouse Visual Cortex

The accurate classification of neuronal cell types is central to decoding brain function, yet remains hindered by data scarcity and cellular heterogeneity. Here, we benchmarked classical and deep generative synthetic data augmentation strategies -- including SMOTE, GANs, VAEs, Normalizing Flows, and DDPMs -- for supervised classification of both electrophysiological (e-type) and morpho-electrophysiological (mee-type) neuron types from the mouse visual cortex. Using a curated dataset annotated with 48 electrophysiological and 24 morphological features, we established baseline classifiers and introduced synthetic data generated by each method. Our results demonstrate that SMOTE-based augmentation yields the highest classification accuracies (absolute gains of 0.16 for e-types, 0.12 for mee-types), outperforming deep generative models. GANs approached similar performance when hyperparameters and sample sizes were optimized, but were more sensitive to model specification. In addition, we benchmarked synthetic neuron fidelity by comparing mean absolute errors between synthetic and real class profiles against the natural phenotypic variability observed between real neuronal classes.

q-bio.NC↗

arXiv2025

A 3d particle visualization system for temperature management

This paper deals with a 3D visualization technique proposed to analyze and manage energy efficiency from a data center. Data are extracted from sensors located in the IBM Green Data Center in Montpellier France. These sensors measure different information such as hygrometry, pressure and temperature. We want to visualize in real-time the large among of data produced by these sensors. A visualization engine has been designed, based on particles system and a client server paradigm. In order to solve performance problems, a Level Of Detail solution has been developed. These methods are based on the earlier work introduced by J. Clark in 1976. In this paper we introduce a particle method used for this work and subsequently we explain different simplification methods we have applied to improve our solution.

cs.CE↗

arXiv2025

Analysis and Knowledge Discovery from Sensors Data to Improve Energy Efficiency

Increases in energy prices and the global goal of mitigating CO2 emissions necessitate the development of intelligent Building Management Systems (BMS) that operate on an energy-efficient basis. Data Centers, buildings and/or group of buildings are often responsible for huge energy consumption. One way to monitor and optimize energy consumption is to instrument buildings using sensors (temperature, pressure, humidity ...) in order to track and solve wrong usage of energy management systems. The majority of the BMS are processing the data dynamically without taking into account the data history due to their constraint problems (time, bandwidth and calculation capability) and data resources. The RIDER project brings together a consortium of research laboratories and enterprises including IBM, to share their expertise in research and development of smart Information Technology (IT) energy platforms. In this context, we aim to improve energy efficiency of buildings or group of building (including data centers) using IT. One of the objectives is to identify valid, potentially useful, and ultimately understandable patterns in data for improving energy efficiency. We propose in this paper an approach of using an integrated platform able to interconnect instrumented buildings and sites, and to provide a high-level point of view for increasing our knowledge from sensors. The expected results are to estimate physical parameters that influence energy consumption based on data set history. Different correlation could be found between different variables, for example, indoor air quality and energy consumption. These results could be applied at a location where no sensor is placed and predict energy consumption from different variables.

eess.SP↗

arXiv2025

Freezing of Gait as a Complication of Pallidal Deep Brain Stimulation in DYT- KMT2B Patients with Evidence of Striatonigral Degeneration

Background: Mutations in KMT2B are a recognized cause of early-onset complex dystonia, with deep brain stimulation (DBS) of the internal globus pallidus (GPi-DBS) being an effective treatment. However, gait impairment, particularly freezing of gait (FOG), remains a significant challenge in DYT-KMT2B patients post-DBS. Objectives: To characterize the emergence of FOG in DYT-KMT2B patients treated with GPi-DBS and explore potential underlying mechanisms, including striatonigral degeneration. Methods: Five patients (four females) with KMT2B-related dystonia and protein-truncating variants (PTVs) were retrospectively analyzed. Clinical progression, response to GPi-DBS, and the presence of FOG were documented. Dopaminergic function was assessed using DaTscan (SPECT for ^123I-ioflupane) in four patients. Results: FOG developed in all patients, with onset ranging from 1 to 15.5 years post-DBS. DaTscan abnormalities, indicative of bilateral striatal dopaminergic denervation, were observed in four cases. Prior to DBS, all patients exhibited dystonia unresponsive to L-dopa, and post-DBS, FOG remained refractory to dopaminergic treatment in most cases. Despite initial improvements in gait post-DBS, only one patient maintained independent ambulation at the last follow-up. Conclusions: FOG is an emerging complication in DYT-KMT2B patients with PTVs undergoing GPi-DBS, potentially linked to underlying striatonigral degeneration. The findings suggest a need for long-term motor surveillance and consideration of alternative therapeutic strategies, including dopaminergic trials, in this patient population. Further studies are required to elucidate the precise mechanisms driving DBS-related hypokinetic gait disturbances in DYT-KMT2B dystonia.

q-bio.NC↗

arXiv2025

Long-term follow-up of DYT1 dystonia patients treated by deep brain stimulation: an open-label study

Long-term efficacy of internal globus pallidus (GPi) deep-brain stimulation (DBS) in DYT1 dystonia and disease progression under DBS was studied. Twenty-six patients of this open-label study were divided into two groups: (A) with single bilateral GPi lead, (B) with a second bilateral GPi lead implanted owning to subsequent worsening of symptomatology. Dystonia was assessed with the Burke Scale. Appearance of new symptoms and distribution according to body region were recorded. In the whole cohort, significant decreases in motor and disability subscores (P < 0.0001) were observed at 1 year and maintained up to 10 years. Group B showed worsening of the symptoms. At 1 year, there were no significant differences between Groups A (without subsequent worsening) and B; at 5 years, a significant difference was found for motor and disability scores. Within Group B, four patients exhibited additional improvement after the second DBS surgery. In the 26 patients, significant difference (P = 0.001) was found between the number of body regions affected by dystonia preoperatively and over the whole follow-up. DBS efficacy in DYT1 dystonia can be maintained up to 10 years (two patients). New symptoms appear with long-term follow-up and may improve with additional leads in a subgroup of patients.

q-bio.NC↗

arXiv2025

Automatic target validation based on neuroscientific literature mining for tractography

Target identification for tractography studies requires solid anatomical knowledge validated by an extensive literature review across species for each seed structure to be studied. Manual literature review to identify targets for a given seed region is tedious and potentially subjective. Therefore, complementary approaches would be useful. We propose to use text-mining models to automatically suggest potential targets from the neuroscientific literature, full-text articles and abstracts, so that they can be used for anatomical connection studies and more specifically for tractography. We applied text-mining models to three structures: two well-studied structures, since validated deep brain stimulation targets, the internal globus pallidus and the subthalamic nucleus and, the nucleus accumbens, an exploratory target for treating psychiatric disorders. We performed a systematic review of the literature to document the projections of the three selected structures and compared it with the targets proposed by text-mining models, both in rat and primate (including human). We ran probabilistic tractography on the nucleus accumbens and compared the output with the results of the text-mining models and literature review. Overall, text-mining the literature could find three times as many targets as two man-weeks of curation could. The overall efficiency of the text-mining against literature review in our study was 98% recall (at 36% precision), meaning that over all the targets for the three selected seeds, only one target has been missed by text-mining. We demonstrate that connectivity for a structure of interest can be extracted from a very large amount of publications and abstracts. We believe this tool will be useful in helping the neuroscience community to facilitate connectivity studies of particular brain regions. The text mining tools used for the study are part of the HBP Neuroinformatics Platform, publicly available at http://connectivity-brainer.rhcloud.com

q-bio.NC↗

arXiv2025

Morphological Neuron Classification Using Machine Learning

Classification and quantitative characterization of neuronal morphologies from histological neuronal reconstruction is challenging since it is still unclear how to delineate a neuronal cell class and which are the best features to define them by. The morphological neuron characterization represents a primary source to address anatomical comparisons, morphometric analysis of cells, or brain modeling. The objectives of this paper are (i) to develop and integrate a pipeline that goes from morphological feature extraction to classification and (ii) to assess and compare the accuracy of machine learning algorithms to classify neuron morphologies. The algorithms were trained on 430 digitally reconstructed neurons subjectively classified into layers and/or m-types using young and/or adult development state population of the somatosensory cortex in rats. For supervised algorithms, linear discriminant analysis provided better classification results in comparison with others. For unsupervised algorithms, the affinity propagation and the Ward algorithms provided slightly better results.

q-bio.NC↗

arXiv2025

Application of quantum machine learning using quantum kernel algorithms on multiclass neuron M type classification

The functional characterization of different neuronal types has been a longstanding and crucial challenge. With the advent of physical quantum computers, it has become possible to apply quantum machine learning algorithms to translate theoretical research into practical solutions. Previous studies have shown the advantages of quantum algorithms on artificially generated datasets, and initial experiments with small binary classification problems have yielded comparable outcomes to classical algorithms. However, it is essential to investigate the potential quantum advantage using real-world data. To the best of our knowledge, this study is the first to propose the utilization of quantum systems to classify neuron morphologies, thereby enhancing our understanding of the performance of automatic multiclass neuron classification using quantum kernel methods. We examined the influence of feature engineering on classification accuracy and found that quantum kernel methods achieved similar performance to classical methods, with certain advantages observed in various configurations.

quant-ph↗

arXiv2021

A new step for computing

The data center of tomorrow is a data center made up of heterogeneous systems, which will run heterogeneous workloads. The systems will be located as close as possible to the data. Heterogeneous systems will be equipped with binary, biological inspired and quantum accelerators. These architectures will be the foundations to address challenges. Like an orchestra conductor, the hybrid cloud will make it possible to set these systems to music thanks to a layer of security and intelligent automation.

cs.DC↗

arXiv2020

The data center of tomorrow is made up of heterogeneous accelerators

physics.pop-ph↗