Biology subjects

Caron, B.

Publications and source records attributed to Caron, B..

9 recordsLinked to original sources

bioRxiv2026

AtlasOT - The Fused Unbalanced Gromov-Wasserstein for Multimodal Integration of Disease Atlases

Single-cell and spatial multiomics technologies are transforming disease atlas construction, but computational integration of unpaired modalities remains challenging, as existing methods overlook two biological constraints of data with matching samples. First, cells should only be mapped across modalities within the same donor or biospecimen, and second, cell recovery frequently differs substantially between modalities. To address these gaps, we present AtlasOT, an optimal-transport framework based on the fused unbalanced Gromov-Wasserstein (FUGW) formulation for multi-modal integration. AtlasOT jointly models a shared cross-modality feature space and modality-specific geometric structures while restricting transport to within-sample cell pairs. Moreover, AtlasOT relaxes strict mass conservation of the balanced optimal transport optimization to accommodate unbalanced cell numbers. We benchmark AtlasOT against state-of-the-art methods on scRNA-scATAC and scRNA-spatial transcriptomics mapping scenarios. Our results indicate that AtlasOT outperforms baselines and state-of-the-art methods in all considered scenarios. Moreover, we demonstrate that AtlasOT's transport plan can be used to improve several relevant tasks such as the detection of rare cell populations, spot-level cell-type deconvolution, spatial gene imputation, and reconstruction of transcription-factor-driven spatial regulatory dynamics. These results establish AtlasOT as a unified, biologically constrained framework for multimodal integration in disease atlas studies, with broad applicability to label transfer, imputation, and regulatory network analysis.

bioinformatics↗

bioRxiv2026

A retinotopic wiring principle of the human brain

Summary paragraphA central challenge in neuroscience is to understand how anatomical connectivity links neural representations to behavior. In sensory systems, cortical areas encode structured maps of the external world, yet it remains unclear whether and how this representational geometry constrains long-range connectivity in the living human brain. Here, we show that structural connectivity in the human visual cortex preserves retinotopic organization, revealing a retinotopic wiring principle; cortical regions representing the same locations in visual space preferentially connect. Across more than 1,700 participants spanning ages 2 to 88 years, we further show that well-known perceptual asymmetries correspond to systematic asymmetries in connectivity within early visual cortex, but not in connections linking visual cortex to the rest of the brain. We introduce a scalable approach to map retinotopic connectivity when tractography alone is unreliable. Together, these findings demonstrate that anatomical connectivity preserves the geometry of neural representations, providing a general principle linking brain structure, function, and behavior.

neuroscience↗

bioRxiv2022

Denoising diffusion weighted imaging data using convolutional neural networks

Diffusion weighted imaging (DWI) with multiple, high b-values is critical for extracting tissue microstructure measurements; however, high b-value DWI images contain high noise levels that can overwhelm the signal of interest and bias microstructural measurements. Here, we propose a simple denoising method that can be applied to any dataset, provided a low-noise, single-subject dataset is acquired using the same DWI sequence. The denoising method uses a one-dimensional convolutional neural network (1D-CNN) and deep learning to learn from a low-noise dataset, voxel-by-voxel. The trained model can then be applied to high-noise datasets from other subjects. We validated the 1D-CNN denoising method by first demonstrating that 1D-CNN denoising resulted in DWI images that were more similar to the noise-free ground truth than comparable denoising methods, e.g., MP-PCA, using simulated DWI data. Using the same DWI acquisition but reconstructed with two common reconstruction methods, i.e. SENSE1 and sum-of-square, to generate a pair of low-noise and high-noise datasets, we then demonstrated that 1D-CNN denoising of high-noise DWI data collected from human subjects showed promising results in three domains: DWI images, diffusion metrics, and tractography. In particular, the denoised images were very similar to a low-noise reference image of that subject, more than the similarity between repeated low-noise images (i.e. computational reproducibility). Finally, we demonstrated the use of the 1D-CNN method in two practical examples to reduce noise from parallel imaging and simultaneous multi-slice acquisition. We conclude that the 1D-CNN denoising method is a simple, effective denoising method for DWI images that overcomes some of the limitations of current state-of-the-art denoising methods, such as the need for a large number of subjects for training and accounting for the rectified noise floor.

neuroscience↗

bioRxiv2021

A massive 7T fMRI dataset to bridge cognitive and computational neuroscience

Extensive sampling of neural activity during rich cognitive phenomena is critical for robust understanding of brain function. We present the Natural Scenes Dataset (NSD), in which high-resolution fMRI responses to tens of thousands of richly annotated natural scenes are measured while participants perform a continuous recognition task. To optimize data quality, we develop and apply novel estimation and denoising techniques. Simple visual inspections of the NSD data reveal clear representational transformations along the ventral visual pathway. Further exemplifying the inferential power of the dataset, we use NSD to build and train deep neural network models that predict brain activity more accurately than state-of-the-art models from computer vision. NSD also includes substantial resting-state and diffusion data, enabling network neuroscience perspectives to constrain and enhance models of perception and memory. Given its unprecedented scale, quality, and breadth, NSD opens new avenues of inquiry in cognitive and computational neuroscience.

neuroscience↗

bioRxiv2021

Development of white matter tracts between and within the dorsal and ventral streams

The degree of interaction between the ventral and dorsal visual streams has been discussed in multiple scientific domains for decades. Recently, several white matter tracts that directly connect cortical regions associated with the dorsal and ventral streams have become possible to study due to advancements in automated and reproducible methods. The developmental trajectory of this set of tracts, here referred to as the posterior vertical pathway (PVP), has yet to be described. We propose an input-driven model of white matter development and provide evidence for the model by focusing on the development of the PVP. We used reproducible, cloud-computing methods and diffusion imaging from adults and children (ages 5-8 years) to compare PVP development to that of tracts within the ventral and dorsal pathways. PVP microstructure was more adult-like than dorsal stream microstructure, but less adult-like than ventral stream microstructure. Additionally, PVP microstructure was more similar to the microstructure of the ventral than the dorsal stream and was predicted by performance on a perceptual task in children. Overall, results suggest a potential role for the PVP in the development of the dorsal visual stream that may be related to its ability to facilitate interactions between ventral and dorsal streams during learning. Our results are consistent with the proposed model, suggesting that the microstructural development of major white matter pathways is related, at least in part, to the propagation of sensory information within the visual system. Significance StatementUnderstanding white matter development is important to building predictive models that can inform interventions and targeted educational methods. We propose and provide evidence for an input-driven model of white matter development. We tested an uncharacterized aspect of human brain development. Namely, how the recently described posterior vertical white matter tracts develop. Our results suggest a developmental progression along the known, direct anatomical connections from posterior visual areas to anterior ventral and dorsal areas. Our results suggest fundamental biological mechanisms that clarify the role of white matter in predicting human learning and behavior.

neuroscience↗

bioRxiv2020

White matter alterations in glaucoma and vision-deprived brains differ outside the visual system

The degree to which glaucoma has effects beyond the eye -in the brain- is unclear. We investigated white matter microstructure (WMM) alterations in 37 tracts of patients with glaucoma, monocular blindness and controls. We used reproducible methods and the advanced cloud computing platform brainlife.io. White matter tracts were subdivided into seven categories ranging from primarily involved in vision (the visual white matter) to primarily involved in cognition and motor control. WMM in both glaucoma and monocular blind subjects was lower than controls in the visual white matter, suggesting neurodegenerative mechanisms due to reduced sensory inputs. In glaucoma participants WMM differences from controls decreased outside the visual white matter. A test-retest validation approach was used to validate these results. The pattern of results was different in monocular blind participants, where WMM properties increased outside the visual white matter as compared to controls. The pattern of results suggests that whereas in the blind loss of visual input might promote white matter reorganization outside of the early visual system, such reorganization might be reduced or absent in glaucoma. The results provide indirect evidence that in glaucoma unknown factors might limit the brain plasticity effects that in other patient groups follow visual loss.

neuroscience↗

bioRxiv2020

Collegiate athlete brain data for white matter mapping and network neuroscience

We describe a dataset of processed data with associated reproducible preprocessing pipeline collected from two collegiate athlete groups and one non-athlete group. The dataset shares minimally processed diffusion-weighted magnetic resonance imaging (dMRI) data, three models of the diffusion signal in the voxel, full-brain tractograms, segmentation of the major white matter tracts as well as structural connectivity matrices. There is currently a paucity of similar datasets openly shared. Furthermore, major challenges are associated with collecting this type of data. The data and derivatives shared here can be used as a reference to study the effects of long-term exposure to collegiate athletics, such as the effects of repetitive head impacts. We use advanced anatomical and dMRI data processing methods publicly available as reproducible web services at brainlife.io.

neuroscience↗

bioRxiv2018

GenPipes: an open-source framework for distributed and scalable genomic analyses

With the decreasing cost of sequencing and the rapid developments in genomics technologies and protocols, the need for validated bioinformatics software that enables efficient large-scale data processing is growing. Here we present GenPipes, a flexible Python-based framework that facilitates the development and deployment of multi-step workflows optimized for High Performance Computing clusters and the cloud. GenPipes already implements 12 validated and scalable pipelines for various genomics applications, including RNA-Seq, ChIP-Seq, DNA-Seq, Methyl-Seq, Hi-C, capture Hi-C, metagenomics and PacBio long read assembly. The software is available under a GPLv3 open source license and is continuously updated to follow recent advances in genomics and bioinformatics. The framework has been already configured on several servers and a docker image is also available to facilitate additional installations. In summary, GenPipes offers genomic researchers a simple method to analyze different types of data, customizable to their needs and resources, as well as the flexibility to create their own workflows.

bioinformatics↗

bioRxiv2018

Scoring of pathogenic non-coding variants in Mendelian diseases through supervised learning on ancient, recent and ongoing purifying selection signals in human

The study of rare Mendelian diseases through exome sequencing typically yields incomplete diagnostic rates, ~8-70% depending on the disease type. Whole genome sequencing of the unresolved cases allows addressing the hypothesis that causal variants could lay in non-coding regions with damaging regulatory consequences. The large amount of rare and singleton variants found in each individual genome requires computational filtering and scoring strategies to gain power in downstream statistical genetics tests. However, state-of-the-art methods estimating the functional relevance of non-coding genomic regions have been mostly characterized on sets of variants largely composed of trait-associated polymorphisms and associated to common diseases, yet with modest accuracy and strong positional biases. In this work we first curated a collection of n=737 high-confidence pathogenic non-coding single-nucleotide variants in proximal cis-regulatory genomic regions associated to monogenic Mendelian diseases. We then systematically evaluated the ability to predict causal variants of a comprehensive set of natural selection features extracted at three genomic levels: the affected position, the flanking region and the associated gene. In addition to inter-species conservation, a comprehensive set of recent and ongoing purifying selection signals in human was explored, allowing to capture potential constraints associated to recently acquired regulatory elements in the human lineage. A supervised learning approach using gradient tree boosting on such features reached a high predictive performance characterized by an area under the ROC curve = 0.84 and an area under the Precision-Recall curve = 0.47. The figures represent a relative improvement of >10% and >34% respectively upon the performance of current state-of-the-art methods for prioritizing non-coding variants. Performance was consistent under multiple configurations of the sets of variants used for learning and for independent testing. The supervised learning design allowed the assessment of newly seen non-coding variants overcoming gene and positional bias. The scores produced by the approach allow a more consistent weighting and aggregation of candidate pathogenic variants from diverse non-coding regions within and across genes in the context of statistical tests for rare variant association analysis.

bioinformatics↗