Silvia Obertino

Infinite Feature Selection on SHORE based biomarkers reveals connectivity modulation after stroke

Connectomics is gaining increasing interest in the scientific and clinical communities. It consists in deriving models of structural or functional brain connections based on some local measures. Here we focus on structural connectivity as detected by diffusion MRI. Connectivity matrices are derived from microstructural indices obtained by the 3D-SHORE. Typically, graphs are derived from connectivity matrices and used for inferring node properties that allow identifying those nodes that play a prominent role in the network. This information can then be used to detect network modulations induced by diseases. In this paper we take a complementary approach and focus on link as opposed to node properties. We hypothesize that network modulation can be better described by measuring the connectivity alteration directly in the form of modulation of the properties of white matter fiber bundles constituting the network communication backbone. The goal of this paper is to detect the paths that are most altered by the pathology by exploiting a feature selection paradigm. Temporal changes on connection weights are treated as features and those playing a leading role in a patient versus healthy controls classification task are detected by the Infinite Feature Selection (Inf-FS) method. Results show that connection paths with high discriminative power can be identified that are shared by the considered microstructural descriptors allowing a classification accuracy ranging between 83% and 89%.

On the Viability of Diffusion MRI-Based Microstructural Biomarkers in Ischemic Stroke

Recent 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.

Shore-based microstructural indices: do they tell us more?

Recent methods for diffusion weighted magnetic resonance convey information about tissue microstructure. In the last years, many models have been proposed for recovering the diffusion signal and extracting information to constitute new families of microstructural indices. Here we focus on three leading diffusion MRI models: NODDI (Neurite Orientation Dispersion and Density Imaging), 3D-SHORE (3D Simple Harmonic Oscillator-based Reconstruction and Estimation) and its formulation in the Cartesian space, the MAPMRI (Mean Apparent Propagator MRI) and analyze the information conveyed by the respective set of indices based on information-theoretic measures. This will allow to objectively assess the ability of each index of capturing microstructural features and thus to shed light on their exploitability in discriminative tasks. To this end, the microstructural descriptors are treated as machine learning features and analyzed via information-theoretic methods. First results on in-vivo data suggest that 3D-SHORE and MAPMARI could be more eloquent in describing microstructure and that a combination of descriptors obtained from all models may provide the best subset of features for a classification task.

Shore-based biomarkers allow patient versus control classification in stroke

In diffusion MRI, numerical biomarkers are usually calculated for research and clinical purposes as Generalized Fractional Anisotropy (GFA). Recently, more eloquent indices allowing a more accurate description of tissue microstructure were derived from the SHORE model. Under certain experimental conditions, such indices express the morphological properties of the compartments where spins diffuse. Evidence of the suitability of such indices as biomarkers for stroke was provided in a previous study based on diffusion spectrum imaging (DSI) and focusing on the cortical motor loop. The goal of this work was to investigate the suitability of such indices for stratification, namely for distinguishing pathological from healthy subjects. To this end, two different paths were followed. First, the same approach used in the previous work for longitudinal analysis (statistics-based) was applied to detect inter-group variations. Then, a new approach based on the LASSO regressor was proposed. Results provided evidence of the suitability of the proposed indices for stratification purposes.

Exploiting Machine Learning Principles for Assessing the Fingerprinting Potential of Connectivity Features

To which extent connectivity measures are able to characterize subjective features? The pipeline leading from the signal acquisition to the connectivity matrix allows numerous degrees of freedom each having an impact on the final result. In this paper, we investigated the sensitivity and specificity of the connectivity models within a machine learning framework through the assessment of the detectability of repeated measures of the same subject versus other subjects. Two fiber Orientation Distribution Function (fODF) reconstruction methods, one of which firstly proposed in this paper, three tractography algorithms and four connectivity features were considered and performance was expressed in terms of Area Under the Curve of the test-retest recognition task. Results suggest that there is a trade-off between the selectivity of the fODF reconstruction methods and the conservativeness of the fiber tracking algorithms across all microstructural indices . The best solution was provided by using an high angular resolution fODF estimation method and the most restrictive deterministic tractography algorithm.