Areas of interest: Neuroimaging, Anatomical Brain Connectivity, Diffusion-Weighted MRI (DTI, GDTI, HARDI, DSI, Q-space), Fiber Tracking, Functional MRI, Bayesian Probability Theory, Markov Chain Monte Carlo Methods, Methods of Mathematical Physics.
Softwares of interest: Matlab, Mathematica, Maple, FSL, SPM, FreeSurfer, Diffusion Toolkit, TrackVis, MRIcro/MRIcroN, Photoshop, etc.
Professional Studies: 1996-1997 - Computer Science at Mathematic & Cybernetics Faculty, Havana University, Cuba. 1997-2003 - Physics at Physics Faculty, Havana University, Cuba.
Abstract: Cognitive impairment is an established feature of schizophrenia. However, little is known about its relationship to the structural and functional brain abnormalities that characterise the disorder. Aims To identify structural and/or functional brain abnormalities associated with schizophrenic cognitive impairment.
Abstract: Neuroimaging classification procedures between normal and pathological subjects are sparse and highly dependent of an expert's clinical criterion. Here, we aimed to investigate whether possible brain structural network differences in the shiverer mouse mutant, a relevant animal model of myelin related diseases, can reflect intrinsic individual brain properties that allow the automatic discrimination between the shiverer and normal subjects. Common structural networks properties between shiverer (C3Fe.SWV Mbp(shi)/Mbp(shi), nâ=â6) and background control (C3HeB.FeJ, nâ=â6) mice are estimated and compared by means of three diffusion weighted MRI (DW-MRI) fiber tractography algorithms and a graph framework. Firstly, we found that brain networks of control group are significantly more clustered, modularized, efficient and optimized than those of the shiverer group, which presented significantly increased characteristic path length. These results are in line with previous structural/functional complex brain networks analysis that have revealed topologic differences and brain network randomization associated to specific states of human brain pathology. In addition, by means of network measures spatial representations and discrimination analysis, we show that it is possible to classify with high accuracy to which group each subject belongs, providing also a probability value of being a normal or shiverer subject as an individual anatomical classifier. The obtained correct predictions (e.g., around 91.6-100%) and clear spatial subdivisions between control and shiverer mice, suggest that there might exist specific network subspaces corresponding to specific brain disorders, supporting also the point of view that complex brain network analyses constitutes promising tools in the future creation of interpretable imaging biomarkers.
Abstract: Evidence for interregional structural asymmetries has been previously reported for brain anatomic regions supporting well-described functional lateralization. Here, we aimed to investigate whether the two brain hemispheres demonstrate dissimilar general structural attributes implying different principles on information flow management. Common left hemisphere/right hemisphere structural network properties are estimated and compared for right-handed healthy human subjects and a nonhuman primate, by means of 3 different diffusion-weighted magnetic resonance imaging fiber tractography algorithms and a graph theory framework. In both the human and the nonhuman primate, the data support the conclusion that, in terms of the graph framework, the right hemisphere is significantly more efficient and interconnected than the left hemisphere, whereas the left hemisphere presents more central or indispensable regions for the whole-brain structural network than the right hemisphere. From our point of view, in terms of functional principles, this pattern could be related with the fact that the left hemisphere has a leading role for highly demanding specific process, such as language and motor actions, which may require dedicated specialized networks, whereas the right hemisphere has a leading role for more general process, such as integration tasks, which may require a more general level of interconnection.
Abstract: Diffusion orientation transform (DOT) is a powerful imaging technique that allows the reconstruction of the microgeometry of fibrous tissues based on diffusion MRI data. The three main error sources involving this methodology are the finite sampling of the q-space, the practical truncation of the series of spherical harmonics and the use of a mono-exponential model for the attenuation of the measured signal. In this work, a detailed mathematical description that provides an extension to the DOT methodology is presented. In particular, the limitations implied by the use of measurements with a finite support in q-space are investigated and clarified as well as the impact of the harmonic series truncation. Near- and far-field analytical patterns for the diffusion propagator are examined. The near-field pattern makes available the direct computation of the probability of return to the origin. The far-field pattern allows probing the limitations of the mono-exponential model, which suggests the existence of a limit of validity for DOT. In the regimen from moderate to large displacement lengths the isosurfaces of the diffusion propagator reveal aberrations in form of artifactual peaks. Finally, the major contribution of this work is the derivation of analytical equations that facilitate the accurate reconstruction of some orientational distribution functions (ODFs) and skewness ODFs that are relatively immune to these artifacts. The new formalism was tested using synthetic and real data from a phantom of intersecting capillaries. The results support the hypothesis that the revisited DOT methodology could enhance the estimation of the microgeometry of fiber tissues.
Abstract: Neuroimaging studies have found evidence of altered brain structure and function in schizophrenia, but have had complex findings regarding the localization of abnormality. We applied multimodal imaging (voxel-based morphometry (VBM), functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI) combined with tractography) to 32 chronic schizophrenic patients and matched healthy controls. At a conservative threshold of P=0.01 corrected, structural and functional imaging revealed overlapping regions of abnormality in the medial frontal cortex. DTI found that white matter abnormality predominated in the anterior corpus callosum, and analysis of the anatomical connectivity of representative seed regions again implicated fibres projecting to the medial frontal cortex. There was also evidence of convergent abnormality in the dorsolateral prefrontal cortex, although here the laterality was less consistent across techniques. The medial frontal region identified by these three imaging techniques corresponds to the anterior midline node of the default mode network, a brain system which is believed to support internally directed thought, a state of watchfulness, and/or the maintenance of one's sense of self, and which is of considerable current interest in neuropsychiatric disorders.
Abstract: Diffusion spectrum magnetic resonance imaging (DSI) allows the estimation of the displacement probability density function (pdf) of water molecules, which contain valuable information about the microgeometry of the medium where the diffusion process occurs. It provides a more general approach to disentangle complex fiber structures in biological tissues because it does not assume any particular model of diffusion; even so, it has a number of limitations that remain unstudied. For instance, the theoretical model used to compute the displacement pdf is based on a Fourier transformation defined in the whole measurement space; however, in practice, it is computed using discrete signals with a finite support. As a consequence, the displacement pdf obtained from the experiments is the convolution between the true pdf and a point spread function (PSF) that completely depends on the experimental sampling scheme. In this work, a general framework to rectify and decontaminate the displacement pdf reconstructed from DSI is introduced. This framework is based on model-free deconvolution techniques that allow obtaining clearer and sharper DSI estimates. The method was tested in synthetic data as well as in real data measured from a healthy human volunteer. The results demonstrated that the angular resolution of DSI can be increased, potentially revealing new real fiber components and reducing both the artefactual peaks and the uncertainty of the local diffusion orientational distribution. Furthermore, the deconvolution process provides scalar maps of quantities derived from the propagator, such as the zero displacement probability, with higher tissue contrast.
Abstract: Novel methodologies have been recently developed to characterize the microgeometry of neural tissues and porous structures via diffusion MRI data. In line with these previous works, this article provides a detailed mathematical description of q-space in spherical coordinates that helps to highlight the differences and similarities between various related q-space methodologies proposed to date such as q-ball imaging (QBI), diffusion spectrum imaging (DSI), and diffusion orientation transform imaging (DOT). This formulation provides a direct relationship between the orientation distribution function (ODF) and the diffusion data without using any approximation. Under this relationship, the exact ODF can be computed by means of the Radon transform of the radial projection (in q-space) of the diffusion MRI signal. This new methodology, termed exact q-ball imaging (EQBI), was put into practice using an analytical ODF estimation in terms of spherical harmonics that allows obtaining model-free and model-based reconstructions. This work provides a new framework for combining information coming from diffusion data recorded on multiple spherical shells in q-space (hybrid diffusion imaging encoding scheme), which is capable of mapping ODF to a high accuracy. This represents a step toward a more efficient development of diffusion MRI experiments for obtaining better ODF estimates.
Abstract: Our goal is to study the human brain anatomical network. For this, the anatomical connection probabilities (ACP) between 90 cortical and subcortical brain gray matter areas are estimated from diffusion-weighted Magnetic Resonance Imaging (DW-MRI) techniques. The ACP between any two areas gives the probability that those areas are connected at least by a single nervous fiber. Then, the brain is modeled as a non-directed weighted graph with continuous arc weights given by the ACP matrix. Based on this approach, complex networks properties such as small-world attributes, efficiency, degree distribution, vulnerability, betweenness centrality and motifs composition are studied. The analysis was carried out for 20 right-handed healthy subjects (mean age: 31.10, S.D.: 7.43). According to the results, all networks have small-world and broad-scale characteristics. Additionally, human brain anatomical networks present bigger local efficiency and smaller global efficiency than the corresponding random networks. In a vulnerability and betweenness centrality analysis, the most indispensable and critical anatomical areas were identified: putamens, precuneus, insulas, superior parietals and superior frontals. Interestingly, some areas have a negative vulnerability (e.g. superior temporal poles, pallidums, supramarginals and hechls), which suggest that even at the cost of losing in global anatomical efficiency, these structures were maintained through the evolutionary processes due to their important functions. Finally, symmetrical characteristic building blocks (motifs) of size 3 and 4 were calculated, obtaining that motifs of size 4 are the expanded version of motif of size 3. These results are in agreement with previous anatomical studies in the cat and macaque cerebral cortex.
Abstract: In this paper we introduce a new method to characterize the intravoxel anisotropy based on diffusion-weighted imaging (DWI). The proposed solution, under a fully Bayesian formalism, deals with the problem of joint Bayesian Model selection and parameter estimation to reconstruct the principal diffusion profiles or primary fiber orientations in a voxel. We develop an efficient stochastic algorithm based on the reversible jump Markov chain Monte Carlo (RJMCMC) method in order to perform the Bayesian computation. RJMCMC is a good choice for this problem because of its ability to jump between models of different dimensionality. This methodology provides posterior estimates of the parameters of interest (fiber orientation, diffusivities etc) unconditional of the model assumed. It also gives an empirical posterior distribution of the number of primary nerve fiber orientations given the DWI data. Different probability maps can be assessed using this methodology: 1) the intravoxel fiber orientation map (or orientational distribution function) that gives the probability of finding a fiber in a particular spatial orientation; 2) a three-dimensional map of the probability of finding a particular number of fibers in each voxel; 3) a three-dimensional MaxPro (maximum probability) map that provides the most probable number of fibers for each voxel. In order to study the performance and reliability of the presented approach, we tested it on synthetic data; an ex-vivo phantom of intersecting capillaries; and DWI data from a human subject.
Abstract: A new methodology is introduced that characterizes the intravoxel orientation distribution function (ODF) based on a single-fiber model of the diffusion MRI signal. Using a Bayesian framework the probability of finding a fiber in a specific orientation is obtained. The proposed ODF estimation relies on a cigar-like diffusion tensor model, the methodology is thus denominated Bayesian cigar-like diffusion tensor (BCDT). This work makes two major contributions: 1) the study of single-fiber models in detecting fibers with different volume fractions in a voxel, and 2) the introduction of the Nth-root correction to improve the detection of fibers with smaller volume fractions, where N is the number of diffusion MRI measurements. It is demonstrated that the incomplete signal modeling fails to reconstruct the relative fiber volume fractions, especially when the intravoxel diffusion profiles have dissimilar contributions to the diffusion MRI signal. In this situation the fibers with smaller contributions are hardly detectable. The BCDT method proposed here reduces this effect by introducing the Nth-root correction, making multiple fibers estimable. The performance of the new methodology is illustrated using synthetic and real data, as well as the data from a phantom of intersecting capillaries.
Abstract: A new methodology based on Diffusion Weighted Magnetic Resonance Imaging (DW-MRI) and Graph Theory is presented for characterizing the anatomical connections between brain gray matter areas. In a first step, brain voxels are modeled as nodes of a non-directed graph in which the weight of an arc linking two neighbor nodes is assumed to be proportional to the probability of being connected by nervous fibers. This probability is estimated by means of probabilistic tissue segmentation and intravoxel white matter orientational distribution function, obtained from anatomical MRI and DW-MRI, respectively. A new tractography algorithm for finding white matter routes is also introduced. This algorithm solves the most probable path problem between any two nodes, leading to the assessment of probabilistic brain anatomical connection maps. In a second step, for assessing anatomical connectivity between K gray matter structures, the previous graph is redefined as a K+1 partite graph by partitioning the initial nodes set in K non-overlapped gray matter subsets and one subset clustering the remaining nodes. Three different measures are proposed for quantifying anatomical connections between any pair of gray matter subsets: Anatomical Connection Strength (ACS), Anatomical Connection Density (ACD) and Anatomical Connection Probability (ACP). This methodology was applied to both artificial and actual human data. Results show that nervous fiber pathways between some regions of interest were reconstructed correctly. Additionally, mean connectivity maps of ACS, ACD and ACP between 71 gray matter structures for five healthy subjects are presented.
Abstract: The diffusion of water in white matter is anisotropic because of high orientational complexity of the
nervous fibers distribution. This characteristic is useful to infer the directions of these fibers through
Diffusion Weighted Magnetic Resonance Imaging. In this work a new method to characterize the
orientational anisotropy of the nervous fibers distribution in white matter based on the Bayesian
Information Criterion model selection and Markov Chains Monte Carlo sampling techniques is presented.
This method detects the number of fibers ( less than four as a priori information ) and itâs spatial
orientations at each voxel.
Abstract: In the so-called âCentury of the brainâ, physics can play an important role in the development of neurosciences. Current research in the field of Complexity could gather a relevant meaning when applied to the study of one of the most attractive natural complex systems: the brain. Realizing that the brain can be understood as a network of many neuronal populations functioning in a critical state, will allow the use of techniques for studying complex systems in the analysis of the brain activity. In this work, we briefly describe novel methods based on the use of physical modeling and modern mathematical algorithms, for the better understanding of cerebral functioning from a âcomplexityâ point of view. These methods apply in three different spatiotemporal scales: anatomical, functional and mental scales. A recent method for characterizing anatomical connections from magnetic resonance imaging of the diffusion of water inside the brain is based in a probabilistic approach on the spatial distribution of the nervous fibers. A fiber-tracking algorithm allows the estimation of trajectories of fiber bunches connecting different anatomical zones. On the other scale, space-time-frequency methods and multivariate autoregressive models can be used for characterizing the functional networks in the brain. Application to experimental data corresponding to a superior highly organized mental activity can shed light on the understanding of the mechanisms which underlie the integration of information between different cerebral areas.
Abstract: There is much current interest in identifying the anatomical and functional circuits that are the basis of the brain's computations, with hope that functional neuroimaging techniques will allow the in vivo study of these neural processes through the statistical analysis of the time-series they produce. Ideally, the use of techniques such as multivariate autoregressive (MAR) modelling should allow the identification of effective connectivity by combining graphical modelling methods with the concept of Granger causality. Unfortunately, current time-series methods perform well only for the case that the length of the time-series Nt is much larger than p, the number of brain sites studied, which is exactly the reverse of the situation in neuroimaging for which relatively short time-series are measured over thousands of voxels. Methods are introduced for dealing with this situation by using sparse MAR models. These can be estimated in a two-stage process involving (i) penalized regression and (ii) pruning of unlikely connections by means of the local false discovery rate developed by Efron. Extensive simulations were performed with idealized cortical networks having small world topologies and stable dynamics. These show that the detection efficiency of connections of the proposed procedure is quite high. Application of the method to real data was illustrated by the identification of neural circuitry related to emotional processing as measured by BOLD.
