Biology subjects

Kherif, F.

Publications and source records attributed to Kherif, F..

5 recordsLinked to original sources

bioRxiv2025

White Matter Microstructure Fingerprint of Cerebral Small Vessel Disease

BackgroundWhite matter hyperintensities (WMHs) are the imaging hallmark of cerebral small vessel disease (SVD), yet their microstructural composition, spatial heterogeneity, and relationship to diffuse normal-appearing white matter (NAWM) damage and cardiovascular risk remain incompletely defined. MethodsIn 363 community-dwelling adults from the BrainLaus cohort (mean age 55.5 years; range 19.8-89.4; 48.8% male), we combined quantitative relaxometry (MTsat, R1, R2*) and diffusion-derived metrics (FA, MD, NODDI, g-ratio). WMHs were automatically segmented on FLAIR and microstructure was quantified across lobes, white matter compartments, and geodesic layers extending from the WMH core into surrounding NAWM. Multivariate organisation was assessed using principal component analysis, and associations with cardiovascular risk factors were tested using partial least squares. ResultsOur analysis revealed demyelination, axonal loss, and extracellular fluid accumulation, particularly in periventricular regions. Layer-specific profiles showed a centrifugal gradient, with myelin loss and oedema at the core and axonal alteration in surrounding tissue. These signatures were associated with age-related cardiovascular risk factors, including higher blood pressure, bioimpedance, and lower hemoglobin levels. ConclusionWMHs index the endpoint of a broader, spatially structured white matter injury process that extends into NAWM, is regionally concentrated in periventricular tissue, and covaries with systemic vascular/metabolic factors. These findings support sustained vascular risk management to mitigate CSVD-related white matter degeneration.

neuroscience↗

bioRxiv2023

Statistical analyses of motion-corrupted MRI relaxometry data

Consistent noise variance across data points (i.e. homoscedasticity) is required to ensure the validity of statistical analyses of MRI data conducted using linear regression methods. However, head motion leads to degradation of image quality, introducing noise heteroscedasticity into ordinary-least square analyses. The recently introduced QUIQI method restores noise homoscedasticity by means of weighted least square analyses in which the weights, specific for each dataset of an analysis, are computed from an index of motion-induced image quality degradation. QUIQI was first demonstrated in the context of brain maps of the MRI parameter R2*, which were computed from a single set of images with variable echo time. Here, we extend this framework to quantitative maps of the MRI parameters R1, R2*, and MTsat, which are computed from multiple sets of images. QUIQI allows for optimization of the noise model by using metrics quantifying heteroscedasticity and free energy. QUIQI restores homoscedasticity more effectively than insertion of an image quality index in the analysis design and yields higher sensitivity than simply removing the datasets most corrupted by head motion from the analysis. In sum, QUIQI provides an optimal approach to group-wise analyses of a range of quantitative MRI parameter maps that is robust to inherent homoscedasticity.

neuroscience↗

bioRxiv2021

Restoring statistical validity in group analyses of motion-corrupted MRI data

Motion during the acquisition of magnetic resonance imaging (MRI) data degrades image quality, hindering our capacity to characterize disease in patient populations. Quality control procedures allow the exclusion of the most affected images from analysis. However, the criterion for exclusion is difficult to determine objectively and exclusion can lead to a suboptimal compromise between image quality and sample size. We provide an alternative, data-driven solution that assigns weights to each image, computed from an index of image quality using restricted maximum likelihood. We illustrate this method through the analysis of brain MRI data. The proposed method restores the validity of statistical tests, and performs near optimally in all brain regions, despite local effects of head motion. This method is amenable to the analysis of a broad type of MRI data and can accommodate any measure of image quality.

neuroscience↗

bioRxiv2018

Multiple Linear Regression: Bayesian Inference for Distributed and Big Data in the Medical Informatics Platform of the Human Brain Project

We propose a Multiple Linear Regression (MLR) methodology for the analysis of distributed and Big Data in the framework of the Medical Informatics Platform (MIP) of the Human Brain Project (HBP). MLR is a very versatile model, and is considered one of the workhorses for estimating dependences between clinical, neuropsychological and neurophysiological variables in the field of neuroimaging. One of the main concepts behind MIP is to federate data, which is stored locally in geographically distributed sites (hospitals, customized databases, etc.) around the world. We restrain from using a unique federation node for two main reasons: first the maintenance of data privacy, and second the efficiency in management of big volumes of data in terms of latency and storage resources needed in the federation node. Considering these conditions and the distributed nature of data, MLR cannot be estimated in the classical way, which raises the necessity of modifications of the standard algorithms. We use the Bayesian formalism that provides the armamentarium necessary to implement the MLR methodology for distributed Big Data. It allows us to account for the heterogeneity of the possible mechanisms that explain data sets across sites expressed through different models of explanatory variables. This approach enables the integration of highly heterogeneous data coming from different subjects and hospitals across the globe. Additionally, it offers general and sophisticated ways, which are extendable to other statistical models, to suit high-dimensional and distributed multimodal data. This work forms part of a series of papers related to the methodological developments embedded in the MIP.

neuroscience↗

bioRxiv2017

Convex Optimized Population Receptive Field (CO-pRF) Mapping

Population receptive field (pRF) mapping represents an invaluable non-invasive tool for the study of sensory organization and plasticity within the human brain. Despite the very appealing result that fMRI derived pRF measures agree well with measurements made from other fields of neuroscience, current techniques often require very computationally expensive non-linear optimization procedures to fit the models to the data which are also vulnerable to bias due local minima issues. In this work we present a general framework for pRF model estimation termed Convex Optimized Population Receptive Field (CO-pRF) mapping and show how the pRF fitting problem can be linearized in order to be solved by extremely fast and efficient algorithms. The framework is general and can be readily applied to a variety of pRF models and measurement schemes. We provide an example of the CO-pRF methodology as applied to a computational neuroimaging approach used to map sensory processes in human visual cortex - the CSS-pRF model. Via simulation and in-vivo fMRI results we demonstrate that the CO-pRF approach achieves robust model fitting even in the presence of noise or reduced data, providing parameter estimates closer to the global optimum across 93% of in-vivo responses as compared to a typical nonlinear optimization procedure. Furthermore the example CO-pRF application substantially reduced model fitting times by a factor of 50. We hope that the availability of such highly accelerated and reliable pRF estimation algorithms will facilitate the spread of pRF techniques to larger imaging cohorts and the future study of neurological disorders and plasticity within the human brain.

neuroscience↗