Neuroimaging Processing and Analysis
In this project we focus on the processing and analysis of neuroimaging. Neuroimaging is the set of techniques to image the structure and function of the brain. We collaborate with psychologists and clinical doctors working in several clinical and technical problems.
Our first works deal with Attention-Deficit/Hyperactivity Disorder (ADHD) diagnosis using anatomical/structural and functional Magnetic Resonance Images (MRI).
Structural MRI is a medical imaging technique used in neuroradiology to visualize the brain internally in detail.
ADHD is a developmental disorder characterized by: Inattentiveness, Motor hyperactivity, and Impulsiveness. It is the most prevalent psychiatric disorder in childhood and the diagnosis is difficult. There are contradictory studies pointing to overdiagnosis, underdiagnosis, and undertreatment of ADHD.
Neuroanatomical Abnormalities in ADHD
Studies of volumetric brain MRI showed neuroanatomical abnormalities in pediatric ADHD. One of the most replicated findings is the diminished volume of the right caudate nucleus. New diagnostic test based on a ratio between head and body of the caudate have been presented showing statistically different in ADHD and control children.
In general, previous studies on ADHD are based on manual procedure, which is extremely time consuming, and prone to inter-rater discrepancies, limiting the power of the analysis.
Automatic ADHD Diagnostic Test
Thus, there is interest in: 1) Developing automatic methods for segmentation and diagnostic test to accelerate the analysis and make the procedure feasible for large amounts of data. 2) Finding new biomarkers to ADHD diagnosis using MRI which could be useful for clinical practice.
In [Igual2011, Igual2012a, Igual2012b] we introduce a fully automatic diagnostic imaging test for ADHD diagnosis assistance based on previously found evidences of caudate nucleus volumetric abnormalities. The proposed method consists of different steps: a new automatic method for external and internal segmentation of caudate based on machine learning methodologies; the definition of a set of new volume relation features used for caudate representation and classification. We separately validate the contributions using real data from a pediatric population. We present precise internal caudate segmentation results and good discrimination power of the diagnostic test, showing significant performance improvements in comparison to other state-of-the-art methods.
This work was developed in collaboration with the research group in cognitive neurosciences from Universitat Autònoma de Barcelona.
Functional MRI (fMRI) measures brain activity by detecting changes in blood flow. It is used to analyse neural activity during different time-points in a task or resting.
Resting-State fMRI (RS-fMRI) is a brain imaging technique useful for exploring functional connectivity. A major point of interest in RS-fMRI analysis is to isolate connectivity patterns characterising disorders such as for instance ADHD. Such characterisation is usually performed in two steps: first, all connectivity patterns in the data are extracted by means of Independent Component Analysis (ICA); second, standard statistical tests are performed over the extracted patterns to find differences between control and clinical groups. In [Tabas2014], we introduce a novel, single-step, approach for this problem termed Spatial Discriminant ICA. The algorithm can efficiently isolate networks of functional connectivity characterising a clinical group by combining ICA and a new variant of the Fisher’s Linear Discriminant also introduced in this work. As the characterization is carried out in a single step, it potentially provides for a richer characterisation of inter-class differences. The algorithm is tested using synthetic and real fMRI data, showing promising results in both experiments.
RS-fMRI can reveal information of functional connectivity between distant regions during rest. In [Nuñez2015], we present a novel method, called Functional-Anatomical Discriminative Regions (FADR), for selecting a discriminative subset of functional-anatomical regions of the brain in order to characterize functional connectivity abnormalities in mental disorders. FADR integrates Independent Component Analysis with a sparse feature selection strategy, namely Elastic Net, in a supervised framework to extract a new sparse representation. In particular, ICA is used for obtaining group Resting State Networks and functional information is extracted from the subject-specific spatial maps. Anatomical information is incorporated to localize the discriminative regions. Thus, functional-anatomical information is combined in the new descriptor, which characterizes areas of different networks and carries discriminative power. Experimental results on the public database ADHD-200 validate the method being able to automatically extract discriminative areas and extending results from previous studies. The classification ability is evaluated showing that our method performs better than the average of the teams in the ADHD-200 Global Competition while giving relevant information about the disease by selecting the most discriminative regions at the same time.
References:
[Igual2011] L. Igual, J. C. Soliva, A. Hernández-Vela, S. Escalera, X. Jiménez, O. Vilarroyaand P. Radeva. A Fully-Automatic Caudate Nucleus Segmentation of Brain MRI: Application in Volumetric Analysis of Pediatric Attention-Deficit Hyperactivity Disorder, BioMedical EngineeringOnLine 2011, 10:105.
[Igual2012a] L Igual, J. C. Soliva, R. Gimeno, S. Escalera, O. Vilarroya, and P. Radeva. Automatic Internal Segmentation of Caudate Nucleus for Diagnosis of Attention-Deffcit/Hyperactivity Disorder. International Conference on Image Analysis and Recognition (ICIAR) 2012. 25-27|06|2012, Aveiro (Portugal)
[Igual2012b] L. Igual, Joan C. Soliva, S. Escalera, R. Gimeno, O. Vilarroya, P. Radeva. Automatic Brain Caudate Nuclei Segmentation and Classification in Diagnostic of Attention-Deficit/Hyperactivity Disorder. Computerized Medical Imaging and Graphics, 36(8), pp.591-600, 2012. doi:10.1016/j.compmedimag.2012.08.002.
[Nuñez2015] M. Nuñez-Garcia, S. Simpraga, M. A. Jurado, M. Garolera, R. Pueyo, and L. Igual. FADR: Functional-Anatomical Discriminative Regions for rest fMRI Characterization. 6th International Workshop on Machine Learning in Medical Imaging, MLMI in conjunction with MICCAI, 2015.
[Tabas2014] A. Tabas, E. Balaguer-Ballester, L. Igual. Spatial Discriminant ICA for RS-fMRI Characterisation. 4th International Workshop on Pattern Recognition in Neuroimaging, PRNI 2014.
Automatic Detection and Characterization of Atherosclerotic Plaque in Carotid Ultrasound
Cardiovascular diseases (CVD) are the leading cause of death in Western countries. The common basis of this group of diseases is atherosclerosis, a chronic degenerative process characterized morphologically by an asymmetric focal thickening of the innermost layer of the artery. The carotid arteries may develop atherosclerosis so that blood flow may be partially or totally blocked by the atherosclerotic plaque. The atherosclerotic plaque can be indirectly detected using carotid artery images. In this project, we have proposed a fully-automatic method to detect the presence of plaque in the far wall common carotid artery image [Zhang2015] and we are working in the extension of this method to other segments of the carotid.
Characterization of carotid plaque composition, more specifically the amount of lipid core, fibrous tissue, and calcified tissue, is an important task for the identification of plaques that are prone to rupture, and thus for early risk estimation of cardiovascular and cerebrovascular events. Due to its low costs and wide availability, carotid ultrasound has the potential to become the modality of choice for plaque characterization in clinical practice. However, its significant image noise, coupled with the small size of the plaques and their complex appearance, makes it difficult for automated techniques to discriminate between the different plaque constituents.
Figure: Four examples of carotid ultrasound images, incorporating noise, artifacts, shadowing, and reverberation. All of these examples contain plaques, but their detection and characterization is challenging even for an expert clinician, resulting in tedious and inconsistent assessments.
In this project, we have addressed this challenging problem by exploiting the unique capabilities of the emerging deep learning framework. More specifically, and unlike existing works, which require a priori definition of specific imaging features or thresholding values, we propose to build a convolutional neural network (CNN) that will automatically extract from the images the information that is optimal for the identification of the different plaque constituents [Lekadir2016]. Based on our previous results, our goal is to progressively build a larger database of training cases together with our clinical partner to further enrich the plaque CNN classification as new datasets become available and to make it more robust to larger variability.
In this project, we will also explore the relation of the plaque composition with the early prediction of cardiovascular and cerebrovascular events using other clinical datasets.
References:
[Zhang2015] Chen Zhang, Maria Del Mar Vila Muñoz, Petia Radeva, Roberto Elosua, María Grau, Angels Betriu, Elvira Fernandez-Giraldez and Laura Igual. Carotid Artery Segmentation in Ultrasound Images. CVII-STENT: Computing and Visualization for Intravascular Imaging and Computer Assisted Stenting in conjunction with MICCAI, 2015.
[Lekadir2016] Karim Lekadir, Alfiia Galimzianova, Àngels Betriu, Maria del Mar Vila, Laura Igual, Daniel L. Rubin, Elvira Fernández, Petia Radeva, and Sandy Napel. A Convolutional Neural Network for Automatic Characterization of Plaque Composition in Carotid Ultrasound. IEEE Journal of Biomedical and Health Informatics (J-BHI), 2016.