140 research outputs found
FeLebrities: a user-centric assessment of Federated Learning frameworks
No full textFederated Learning (FL) is a new paradigm aimed at solving data access problems. It provides a solution by moving the focus from sharing data to sharing models. The FL paradigm involves different entities (institutions) holding proprietary datasets that, contributing with each other to train a global Artificial Intelligence (AI) model using their own locally available data. Although several studies have proposed methods to distribute the computation or aggregate results, few efforts have been made to cover on how to implement FL pipelines. With the aim of accelerating the exploitation of FL frameworks, this paper proposes a survey of public tools that are currently available for building FL pipelines, an objective ranking based on the current state of user preferences, and an assessment of the growing trend of the tool’s popularity over a one year time window, with measurements performed every six months. These measurements include objective metrics, like the number of “Watch,” “Star” and “Follow” available from software repositories as well as thirteen custom metrics grouped into three main categories: Usability, Portability, and Flexibility. Finally, a ranking of the maturity of the tools is derived based on the key aspects to consider when building a FL pipeline
FeLebrities: a user-centric assessment of Federated Learning frameworks
No full textFederated Learning (FL) is a new paradigm aiming to solve the data access problem. It is gaining an increasing interest in a variety of research fields, including the Biomedical and Financial environments, where lots of valuable data sources are available but not often directly accessible due to the regulations that protect sensitive information. FL provides a solution by moving the focus from sharing data to sharing models. The FL paradigm involves different entities (institutions) holding proprietary datasets, contributing with each other to train a global Artificial Intelligence (AI) model using their own locally available data. Although several studies propose ways to distribute the computation or aggregate results, fewer efforts have been made on how to implement it. With the ambition of helping accelerate the exploitation of FL frameworks, this paper proposes a survey of public tools that are currently available for building FL pipelines, an objective ranking based on the current state of user preferences, and the assessment of the growing trend of the tool’s popularity over a six months time window. Finally, a ranking of the maturity of the tools is derived based on keyaspects to consider when building an FL pipeline. </p
Functional Sensitivity of Dual-Echo ASL in Localizing Active and Imagery Hand Movements
No full textDual-echo arterial spin labeling (DE-ASL) techniques have been recently proposed for thesimultaneous acquisition of ASL and blood-oxygenation-level-dependent (BOLD) functionalmagnetic resonance imaging (fMRI) data (Woolrich et al., 2006). The images acquired at the firstecho time are perfusion weighted (ASL), while the images from the second echo are primarilyT2* weighted, thus sensitive to the BOLD signal (Leontiev and Buxton, 2007). The sequence isuseful when the simultaneous estimation of blood flow and BOLD signal are targeted. Thepurpose of this study was to assess the sensitivity of the DE-ASL sequence in comparison to theconventional one (BOLD-fMRI) in detecting brain activations elicited by active and motor imageryhand movements
Leukocyte Telomere Length and Cardiac Structure and Function: A Mendelian Randomization Study
No full textBackground Existing research demonstrates the association of shorter leukocyte telomere length with increased risk of age‐related health outcomes including cardiovascular diseases. However, the direct causality of these relationships has not been definitively established. Cardiovascular aging at an organ level may be captured using image‐derived phenotypes of cardiac anatomy and function. Methods and Results In the current study, we use 2‐sample Mendelian randomization to assess the causal link between leukocyte telomere length and 54 cardiac magnetic resonance imaging measures representing structure and function across the 4 cardiac chambers. Genetically predicted shorter leukocyte telomere length was causally linked to smaller ventricular cavity sizes including left ventricular end‐systolic volume, left ventricular end‐diastolic volume, lower left ventricular mass, and pulmonary artery. The association with left ventricular mass (β =0.217, Pfalse discovery rate=0.016) remained significant after multiple testing adjustment, whereas other associations were attenuated. Conclusions Our findings support a causal role for shorter leukocyte telomere length and faster cardiac aging, with the most prominent relationship with left ventricular mass
Dual-echo ASL based assessment of motor networks: a feasibility study
No full textOBJECTIVE: Dual-echo arterial spin labeling (DE-ASL) techniques have been recently proposed for the simultaneous acquisition of ASL and blood-oxygenation-level-dependent (BOLD)-functional magnetic resonance imaging (fMRI) data. The assessment of this technique in detecting functional connectivity at rest or during motor and motor imagery tasks is still unexplored both per-se and in comparison with conventional methods. The purpose is to quantify the sensitivity of the DE-ASL sequence with respect to the conventional fMRI sequence (cvBOLD) in detecting brain activations, and to assess and compare the relevance of node features in decoding the network structure. APPROACH: Thirteen volunteers were scanned acquiring a pseudo-continuous DE-ASL sequence from which the concomitant BOLD (ccBOLD) simultaneously to the ASL can be extracted. The approach consists of two steps: (i) model-based analyses for assessing brain activations at individual and group levels, followed by statistical analysis for comparing the activation elicited by the three sequences under two conditions (motor and motor imagery), respectively; (ii) brain connectivity graph-theoretical analysis for assessing and comparing the networks model properties. MAIN RESULTS: Our results suggest that cvBOLD and ccBOLD have comparable sensitivity in detecting the regions involved in the active task, whereas ASL offers a higher degree of co-localization with smaller activation volumes. The connectivity results and the comparative analysis of node features across sequences revealed that there are no strong changes between rest and tasks and that the differences between the sequences are limited to few connections. SIGNIFICANCE: Considering the comparable sensitivity of the ccBOLD and cvBOLD sequences in detecting activated brain regions, the results demonstrate that DE-ASL can be successfully applied in functional studies allowing to obtain both ASL and BOLD information within a single sequence. Further, DE-ASL is a powerful technique for research and clinical applications allowing to perform quantitative comparisons as well as to characterize functional connectivity.ventional fMRI sequence (cvBOLD) in detecting brain activations, and to assess and compare the relevance of node features in decoding the network structure
Imaging Genetics through Brain Age Estimation and Image Derived Phenotypes
No full textIn this thesis, we investigated brain aging using different simple and complex models through brain age estimation using IDPs extracted from brain MRI.We have also applied simple methods and machine learning explainability models to identify the most informative features to model brain age. We further estimated brain age for fiber groups within brain white matter tracts. In addition, we revealed the effects of daily life style, cardiac risk factors and morbidity in brain aging. Finally, we used causal models to explore the role of TL in healthy aging and Alzheimer’s disease in unhealthy aging to cause alterations within brain structures and functions
The Brain is a Social Network
No full textSocial Network Analysis is employed widely as a means to compute the probability that a given message flows through a social net- work. This approach is mainly grounded upon the correct usage of three basic graph-theoretic measures: degree centrality, closeness centrality and betweeness centrality. We developed a model, using Semantic Social Net- work Analysis, that overcomes the drawbacks of general indices and we found that this model can be applied, after appropriate adaptations, to a very different domain such as brain connectivity
Test-Retest Reliability of Graph Metrics in Functional Brain Network
No full textThe statistical link between spontaneous fluctuations occurring in different parts of the brain can provide insights into its functional organization. Here, we used high-quality resting-state fMRI (rs-fMRI) data acquired with a test-retest (TRT) paradigm to assess the reliability of graph metrics. After applying time/frequency methods to generate FC matrices, we restricted our focus on global, local and central graph measure through different statistical measures, including but not limited to the intraclass correlation coefficient (ICC). We found that full correlation and magnitude square coherence yielded more reproducible measurements than the other metrics, as revealed by ICC values. These results have important implications when choosing metrics for quantifying FC in rs-fMRI studies, adding novel information to the current panorama of information on TRT reliability topic
Dual-Echo ASL contributes to decrypting the link between functional connectivity and cerebral blow flow
No full textArterial spin labeling (ASL) MRI with a dual-echo readout module (DE-ASL) enables noninvasive simultaneous acquisition of cerebral blood flow (CBF)-weighted images and blood oxygenation level dependent (BOLD) contrast. Up to date, resting-state functional connectivity (FC) studies based on CBF fluctuations have been very limited, while the BOLD is still the method most frequently used. The purposes of this technical report were: i) to assess the potentiality of the DE-ASL sequence for the quantification of resting-state FC and brain organization, with respect to the conventional BOLD (cvBOLD); ii) to investigate the relationship between a series of complex network measures and the CBF information. Thirteen volunteers were scanned on a 3T scanner acquiring a pseudo-continuous multi-slice DE-ASL sequence, from which the concomitant BOLD (ccBOLD) simultaneously to the ASL can be extracted. In the proposed comparison, the brain FC and graph-theoretical analysis were used for quantifying the connectivity strength between pairs of regions and for assessing the network model properties in all the sequences. The main finding was that the ccBOLD part of the DE-ASL sequence provided highly comparable connectivity results compared to cvBOLD. As expected, because of its different nature, ASL sequence showed different patterns of brain connectivity and graph indices compared to BOLD sequences. To conclude, the resting-state FC can be reliably detected using DE-ASL, simultaneously to CBF quantifications, whereas a single fMRI experiment precludes the quantitative measurement of BOLD signal changes
On the Viability of Diffusion MRI-Based Microstructural Biomarkers in Ischemic Stroke
No full textRecent tract-based analyses provided evidence for the exploitability of 3D-SHORE microstructural descriptors derived from diffusion MRI (dMRI) in revealing white matter (WM) plasticity. In this work, we focused on the main open issues left: (1) the comparative analysis with respect to classical tensor-derived indices, i.e., Fractional Anisotropy (FA) and Mean Diffusivity (MD); and (2) the ability to detect plasticity processes in gray matter (GM). Although signal modeling in GM is still largely unexplored, we investigated their sensibility to stroke-induced microstructural modifications occurring in the contralateral hemisphere. A more complete picture could provide hints for investigating the interplay of GM and WM modulations. Ten stroke patients and ten age/gender-matched healthy controls were enrolled in the study and underwent diffusion spectrum imaging (DSI). Acquisitions at three and two time points (tp) were performed on patients and controls, respectively. For all subjects and acquisitions, FA and MD were computed along with 3D-SHORE-based indices [Generalized Fractional Anisotropy (GFA), Propagator Anisotropy (PA), Return To the Axis Probability (RTAP), Return To the Plane Probability (RTPP), and Mean Square Displacement (MSD)]. Tract-based analysis involving the cortical, subcortical and transcallosal motor networks and region-based analysis in GM were successively performed, focusing on the contralateral hemisphere to the stroke. Reproducibility of all the indices on both WM and GM was quantitatively proved on controls. For tract-based, longitudinal group analyses revealed the highest significant differences across the subcortical and transcallosal networks for all the indices. The optimal regression model for predicting the clinical motor outcome at tp3 included GFA, PA, RTPP, and MSD in the subcortical network in combination with the main clinical information at baseline. Region-based analysis in the contralateral GM highlighted the ability of anisotropy indices in discriminating between groups mainly at tp1, while diffusivity indices appeared to be altered at tp2. 3D-SHORE indices proved to be suitable in probing plasticity in both WM and GM, further confirming their viability as a novel family of biomarkers in ischemic stroke in WM and revealing their potential exploitability in GM. Their combination with tensor-derived indices can provide more detailed insights of the different tissue modulations related to stroke pathology
- …
