Brain-Computer Interfaces (BCI) that could decode thoughts into commands would improve the quality of life of patients who have lost control over voluntary muscles. Imagined speech consists in imagining the pronunciation of words, without moving or emitting sounds. In this study, we introduce a new open access database of electroencephalogram (EEG) signals recorded while 15 subjects imagined the pronunciation of two groups of Spanish words. The first one contained the vowels /a/, /e /, /i/, /o/, /u/; and the second one corresponds to the commands up, down, left, right, backward and forward. Each subject repeated each word 50 times in a random order, meanwhile EEG signals were recorded using a six channel acquisition system and sampled at 1024 Hz. For comparison, some blocks were recorded using the pronounced speech condition, in which audio and EEG signals were acquired simultaneously. The EEG signals were filtered for artifact’s removal between 2 Hz and 40 Hz using a finite impulse response (FIR) pass-band filter. As a preliminary analysis of the EEG data, an offline classification method is presented. Accuracy rate is above chance level for almost all subjects, suggesting that EEG signals possess discriminative information about the imagined word.

Transfer entropy (TE) is a nonlinear metric employed recently in polysomnography (PSG) recordings to quantify the topological characteristics of the brain-heart physiological network. The present study applies the TE to evaluate its usefulness to identify quantitative differences in PSG registers of patients diagnosed with oclusive sleep apnea (OSA), before and after a continuous positive air pressure (CPAP) therapy. PSG recordings corresponding to 19 OSA patients were analysed under the rationale that the set of EEG subbands represents the sympathetic activity of the autonomic nervous system (ANS), and the high frequency component of the heart rate variability (HRV) represents the parasympathetic activity. The TE was computed based on a binning estimation and the results were analyzed via effect size calculation. The results showed that the sympathetic activity is increased in the presence of OSA, which is represented by the increased flow of information among brain subsystems and dropping to values close to zero during CPAP therapy. In contrast, the parasympathetic activity showed to be reduced in the presence of OSA and augmented during the CPAP therapy.

Quasi-periodic blood flow signals, obtained by means of Doppler ultrasound techniques during graft verification, may be contaminated with random distortions (artifacts not periodical not homogeneous and affecting only some cycles in the signal). These distortions are not distributed regularly and cannot be characterized statistically or modelled with a known probability distribution function. Also, it is not possible to estimate when or where they will be presented and can cause the major deformations in the cycles where they occur, including the total loss of cycle morphology. In this paper, an improved analysis method to detect the presence of such distortions is proposed. It is based on a modified mean square error method to identify the affected cycles and then, it uses the well-known and widely used method Dynamic Time Warping for reducing the false positives detection. The identification, in time domain, of the affected cycles, and their exclusion of the signal analysis, allows to estimate parameters and extract the clinically useful information needed for a correct characterization of the blood vessel and to improve the results of coronary revascularization procedures. We tested the proposed algorithm in real signals and to evaluate the results we compute the pulsatility index. The results improve the false positive reduction comparing with the method based only in the modified mean square error reported previously.

To search for possible interactions between the autonomic heart control and the brain activity we have evaluated the relation between heart rate variability (HRV) and the alpha and beta EEG band power dynamics. The experiments consisted in the alteration of HRV induced by respiratory rhythm changes. Modifications in the time series of the spectral density in each band {causally related to the HRV{ were recognized in the central and occipital regions using a Granger causality approach. The number of subjects and distribution of channels in which there was a causal relation changed with the different tasks assigned to the subjects and the analyzed bands. The causal relation between HRV and beta power series was observed in at least one channel, in eight out of nine subjects for a resting condition. On the other hand, in the HRV and alpha power series analysis the change in distribution was less pronounced, and detected in a smaller number of subjects. The results suggest that the HRV may be associated with the spectral band power time series dynamics, and this relationship is altered by the respiration.

During the first trimester of pregnancy fetal health assessment is especially important. In the clinical practice, the gestational sac (GS) volume estimation is manually done using a tedious procedure which is prone to physicians' subjectivity. The method proposed in this paper consists on a semiautomatic delimitation of the GS and a segmentation of its content with minimal expert intervention. It is based on spreading active contours (SAC), following a planimetric strategy to define the GS' edges. Additionally, an optimal thresholding method was used to separate solid matter and amniotic fluid. The comparison between manual GS segmentations and those obtained with the proposed SAC method, shows Dice similarities of 90% and a mean Hausdor distance of 5.63 ± 1.94 mm, while the correlation index between SAC and the clinical reference (VOCAL) is 0.997. However, with statistical tests (t-paired) a value of p < 0.05 was obtained, which suggests a difference in the measured volume by the compared methods. The proposed method (SAC) has shown to be reliable, besides of being easy to implement.

The thickness of the nuchal fold is one of the main markers for the detection of Down syndrome during the second trimester of pregnancy. In this paper are reported our preliminary results of the automatic segmentation and measurement of the nuchal fold thickness in ultrasound images of the fetal brain. The method is based on a 2D active shape model used to segment the brain structures involved in the measurement of the nuchal fold: cerebellum; brain midline; the outer edge of the occipital plate; and the outer skin edge. The algorithm was trained and tested in 10 different ultrasound images, using leave one out cross validation. We have obtained an average difference of 0.23 mm from the expert measurement of the nuchal fold, with a standard deviation of 0.1 mm.

Three dimensional ultrasound imaging has become the main modality for fetal health diagnostics, with extensive use in fetal brain imaging. According to the fetal position and the stage of development of the fetal skull, a specific plane of image acquisition is required. In most cases for a single plane of acquisition, the image quality is limited by the shadows produced by the skull. In this work a new method for registration of multiple views of 3D ultrasound of the fetal brain is reported, which results in improved imaging of the internal brain structures. In the initial stage, texture, intensity and edge features are used, with a support vector machine (SVM) for the segmentation of the skull in each of the 3D ultrasound views to be registered. The segmentation of each skull is modelled as a set of points with the centre determined with a Gaussian mixture model, where each point is assigned a probability of membership to a Gaussian determined by the posterior probability assigned by the SVM. Our method has shown improved results compared to intensity based registration, with a 52% reduction in the target registration error (TRE), and a 39% reduction in the TRE compared to feature based registration. These are encouraging results for the future development of an automatic method for registration and fusion of multiple views of 3D fetal ultrasound.

Medical ultrasound image acquisition is a process that depends heavily on the skills of the operator. Simulation- based medical education has proven effective as a means to improve professional's performance and reduce patient's stress. In this paper, we implement and compare two modern ultrasound simulation models aimed for training. Both methods are able to produce images at a speed suitable for real-time application, and are based in CT images as a source for the simulation.

Sickle cell disease (SCD) is a genetic hematological disease in which the hemoglobin molecule in red blood cells is abnormal. It is closely associated with many symptoms, including pain, anemia, chest syndrome and neurocognitive impairment. One of the most debilitating symptoms is elevated risk for cerebro-vascular accidents. The corpus callosum (CC), as the largest and most prominent white matter (WM) structure in the brain, can reflect the chronic cerebrovascular damage resulting from silent strokes or infarctions in asymptomatic SCD patients. While a lot of studies have reported WM alterations in this cohort, little is known about the shape deformation of the CC. Here we perform the first surface morphometry analysis of the CC in SCD patients using four different shape metrics on T1-weighted magnetic resonance images. We detect regional surface morphological differences in the CC between 11 patients and 10 healthy control subjects. Differences are located in the genu, posterior midbody and splenium, potentially casting light on the anatomical substrates underlying neuropsychological test differences between the SCD and control groups.

Clinical practice on magnetic resonance imaging of the brain has been historically based on a comparative analysis using a well-trained eye to see whether different features corresponded to a healthy pattern or not. Several studies have described that healthy aging is associated with loss in tissue volume and expansion of cerebrospinal fluid cavities, making this healthy pattern a dynamical and complex model. For these reasons we propose that structural neurorradiology should be assisted by a quantitative and statistical model that can give meaning to a patient's brain morphometric measurements, giving additional information to the clinician about possible deviations from health. With this aim we obtained normative brain morphometric values by applying an automated voxel and surface-based processing pipeline using the well-known software package FreeSurfer. Employing the publicly available IXI Dataset created by Imperial College London we obtained 135 metrics of the aging process from 538 participant between 20 and 86 years old. In concordance with previous studies we found evidence of change in almost all analyzed features, for both brain's volumes and thicknesses, reproducing findings from several previous brain's morphometric studies. Finally, we explored how different stratified percentiles evolve with age, finding that aging is not a process that can be described by a mean descriptor but on the contrary should be analyzed by considering different percentile layers with its own specific aging dynamic.

The quest to identify genetic factors that shape the human brain has been greatly accelerated by imaging. The human brain functions as a complex network of integrated systems and connected processes, and a vast number of features can be observed and extracted from structural brain images -- including regional volume, shape, and other morphological features of given brain structures. This feature set can be considered as part of the structural network of the brain, which is under strong genetic control. However, it is unclear which of the imaging derived features serve as the most promising traits for discovering specific genes that affect brain structure. Here, we aim to create the first ever network of genetically correlated cortical sulcal features, and through a twin model, determine the degree of genetic correlation across the entire network. Building on functional brain network analysis, we consider the high-dimensional genetic correlation structure as a undirected graph with a complex network of multi-weighted hubs to uncover the underlying genetic core of sulcal morphometry.

The population is aging as the years pass. There is an increase in life expectancy, but also a decrease in the quality of life for the presence of chronic degenerative diseases. Processing medical images can identify brain changes typical of Alzheimer's disease (AD) and Mild Cognitive Impairment (MCI) at an early stage. We propose a new method of segmentation technique using Mean Shift algorithm applying probabilistic maps and Support Vector Machine with Linear and Radial Basis Kernel for segmentation of the hippocampus on Magnetic Resonance Images (MRI). The similarity index of DICE for a 8 control subject was calculated obtaining a mean value of 0:7053±0:0996 using Linear kernel and 0:7275±0:1335 using RBF kernel compared with the manual segmentation made by radiologist.

We studied the developmental trajectory of the putamen in 13-21 months old children using multivariate surface tensor-based morphometry. Our results indicate surface changes between 12 and 15 months’ age groups in the middle superior part the left putamen. The growth of the left putamen at earlier ages slows down after 15 months. The most important surface changes were detected in the right putamen between 18 and 21 months and were located in the anterior part of the structure. Our results demonstrate the heterochronic growth of the right and left putamen related to different functional subregions within putamen. Our results are compatible with previous studies devoted to total putamen volume changes during normal development.

A previously introduced variation of a conventional P300 speller, consisting on a modifiable image background and asymmetrically arranged stimulation markers for controlling wheelchair navigation, was used in this study. Five commonly used classifiers for solving P300 speller-like tasks, namely, Linear-SVM, RBF-SVM, LASSO-LDA, Shrinkage-LDA and SWLDA, were designed and trained and their performances contrasted, seeking the classifier with highest performance on our proposed screen. 19 able-bodied subjects participated in this study. The highest median sensitivity and specificity were respectively 1.00 (IQR = 0.61-1.00) and 1.00 (IQR = 0.96-1.00), which were obtained with the LASSO approach. These performances are suitable for the planned application and they are comparable with the conventional P300 speller performances reported, despite of our speller variation. Friedman tests showed that there are no statistical differences on the sensitivity and specificity performances among the five classifiers evaluated. However, the customized selection of the classifier approach improves the sensitivity by 66.7% in some cases.

The purpose of this study was to introduce a balance evaluation system using multiple motion gaming devices and verify its reliability and feasibility in the clinical setting. This research collected data from 12 (4 males, 8 females) healthy participants without neurological or musculoskeletal impairment (21.58 ± 1.44 years, 168.25 ± 8.66 cm, 61.67 ± 16.01 kg) and the participants performed balance tests on the Wii Balance Board, in front of the Kinect. The present study analyzed the total path distance of the center of pressure (COP), the area of COP displacement, and the number of postural error while maintaining static position for 20 seconds. Six different conditions consisted of double leg stance (DLS), single-dominant leg stance (SDLS), and single-non-dominant leg stance (SNLS) with both eyes open (EO) or closed (EC). Correlation analyses and repeated-measures ANOVAs with post-hoc tests were implemented by using SPSS 23 and the statistical significance level was set with α=0.05. The result of analysis related to the reliability of the system is from r=.173 to r=.917. COP (path distance, area) was significantly higher for each stance with EC than with EO, except the COP area between DLS conditions. There were no significant difference between SDLS_EO and SNLS_EO and between SDLS_EC and SNLS_EC. In the future, this system could be a new balance assessment tool in the clinical setting.

Several approaches using auditory feedback have been proposed to improve gait rehabilitation in Parkinson Disease. Despite auditory cues have shown to be useful, there are still unanswered questions about their optimal usage regarding parameters like frequency, number of beats and their integration with rehabilitation protocols, among others. Most approaches have attempted to resolve these questions by measuring their direct effect on spatiotemporal gait variables. However, few studies have assessed how synchronized the auditory feedback and the gait pattern are. The main goal was to quantify synchronization between the gait temporal patterns and the auditory stimuli. The group of participants consisted of seven (7) healthy subjects, aged between 50-70 years (average 57.28, ± 5.87 years), with average height of 1.64±0.09m and independent community ambulation. Each candidate was asked to sign an informed consent, given their good cognitive conditions for understanding the nature and purpose of the study. Participants were instructed to follow the sounds provided by a metronome. Feet tracking yielded the temporal gait pattern. The temporal coherence metric was developed to evaluate synchronization between audio signal and subject motion, in terms of phase shift (π radian). Results show a good fit to auditory stimulus in metronome rates between 140-150 and 60-80 beats/min (bpm) for the selected participants. A lower temporal coherence was observed at the beginning and the end of the test. The proposed metric allows quantification of the temporal coherence between gait and auditory cues in healthy elder subjects. Other exploratory trials should be directed to evaluate the temporal coherence between auditory stimuli and generated movements in population with Parkinson Disease.

The analysis of deconvolution algorithms performance is crucial when is wanted to use deconvolution as an effective image restoration approach. The performance of deconvolution algorithms available in Open Source software was compared by using three-dimensional (3D) microscopy images. DeconvolutionLab and MitivBlindDeconvolution were the plugins chosen from FIJI and Icy respectively. In the first place, analyses included 3D bead stack measurements both pre- and post-deconvolution by using theoretical and empirical Point Spread Functions (PSFs) as well as parameter variation. The set of parameters that resulted in the improvement of both the 3D morphology and intensity of the beads was applied to 3D autofluorescence colon tissue images from BALB/c mice to evaluate if original morphology and intensity features were restored.

Purpose: Evaluate the feasibility of spectral analysis, particularly fast fourier transform (FFT), to help clinicians differentiate clear cell renal cell carcinoma (ccRCC) tumor grades using contrast-enhanced computed tomography (CECT) images of renal masses, quantitatively, and compare their performance to the Fuhrman grading system. Materials and Methods: Regions of interest of the whole lesion were manually segmented and co-registered from multiphase CT acquisitions of 95 patients with ccRCC. Here, FFT is employed to objectively quantify the texture of a tumor surface by evaluating tissue gray-level patterns and automatically measure frequency-based texture metrics. An independent t-test or a Wilcoxon rank sum test (depending on the data distribution) was used to determine if the spectral analysis metrics would produce statistically significant differences between the tumor grades. Receiver operating characteristic (ROC) curve analysis was used to evaluate the usefulness of spectral metrics in predicting the ccRCC grade. Results: The Wilcoxon test showed that there was a significant difference in complexity index between the different tumor grades, p < 0.01 at all the four phases of CECT acquisition. In all cases a positive correlation was observed between tumor grade and complexity index. ROC analysis revealed the importance of the entropy of FFT amplitude, FFT phase and complexity index and its ability to identify grade 1 and grade 4 tumors from the rest of the population. Conclusion: Our study suggests that FFT-based spectral metrics can differentiate between ccRCC grades, and in combination with other metrics improve patient management and prognosis of renal masses.

A novel algorithm to evaluate the quality of surface segmentations extracted from 3D images is presented. The procedure calculates the volume enclosed between the segmented object represented by a triangular mesh and a reference one. The indicator is computed by means of a robust and efficient ray-tracing algorithm. This algorithm is fully parallelizable, and it can run even on GPUs architectures. The method is validated against synthetic cases and segmentations of real medical images.

In this paper we present an Eulerian motion magnification technique using a spatial decomposition based on the Steered Hermite Transform (SHT) which is inspired in the Human Vision System (HVS). We test our method in one sequence of the breathing of a newborn baby and on a video sequence that shows the heartbeat on the wrist. We estimate the heart pulse applying the Fourier transform on the magnified sequences. Our motion magnification approach is compared to the Laplacian and the Cartesian Hermite decomposition strategies by means of quantitative metrics.

We present an algorithm to align cortical surface models based on structural connectivity. We follow the continuous connectivity approach,^{1, 2} assigning a dense connectivity to every surface point-pair. We adapt and modify an approach for aligning low-rank functional networks based on eigenvalue decomposition of individual connectomes.³ The spherical demons framework then provides a natural setting for inter-subject connectivity alignment, enforcing a smooth, anatomically plausible correspondence, and allowing us to incorporate anatomical as well as connectivity information. We apply our algorithm to 98 diffusion MRI images in an Alzheimer's Disease study, and 731 healthy subjects from the Human Connectome Project. Our method consistently reduces connectome variability due to misalignment. Further, the approach reveals subtle disease effects on structural connectivity which are not seen when registering only cortical anatomy.

Among the most dangerous cancers, there is the Melanoma that affects millions of people. As this is a type of malignant pigmented skin lesion and it can be recognized by medical experts, computer-aided diagnostic systems are developed in order to assist dermatologists in clinical routine. One of the more difficult tasks is to find the right segmentation of lesions whose precision is very important to distinguish benign from malignant cases. In this work, we propose a new method based on sparse representation. First, an alternative representation of the image is obtained from the texture information. A sparse non-negative dictionary is computed and every image is projected onto this space. The reconstruction is calculated using only the most active atoms, which allows to obtaining an enhanced version of the texture where the morphological post-processing can effectively extract the lesion area. The experiments were carried out on a publicly available database and performance was evaluated in terms of segmentation error, accuracy, and specificity. Results showed that this first approach performs better than methods reported in the literature on this same data.

There are many different methods to perform gray level segmentation in microscope cell images, however in some circumstances texture features and roughness are not as relevant as the color for the segmentation task. In many biomedical applications, where these types of images are analyzed, the aim is to segment nuclei for their clinical analysis. In this work, a fuzzy color mathematical morphology reconstruction technique was developed, based on a new locally defined ordering, to achieve microscope cell images segmentation. We show experimental results for this proposed fuzzy color morphological reconstruction, which show that this tool can be efficiently applied in cells segmentation without generating false colors.

This paper presents a novel color separation method for Hematoxylin-Eosin (H&E) stained Histopathology Images. The whole (R;G;B) space of the input image is projected to a family of surfaces connecting the distributions of a series of [(R + B)²=B;G] planes that divide the cloud of H&E tones. Such projection is then used to cluster both the Hematoxylin and Eosin samples, from where the color basis are then derived. The projection presented herein is more resilient to noise and therefore the separation process is less hampered when the staining properties largely vary. Unlike other normalization methods evaluated by comparing with gold standard images, the power of the method was demonstrated by thresholding the resulting Hematoxylin image with the well-known Otsu's method and the number of detected nuclei was compared with two manually annotated datasets. Despite the simplicity of this approach, the detection sensitivity was 71 % and 67 %, respectively

Evidence based medicine aims to provide a quantifiable framework to support cancer optimal treatment selection. Pathological examination is the main evidence used in medical management, yet the level of quantification is low and highly dependent on the examiner expertise. This paper presents and evaluates a method to extract graph based topological features from skin tissue images to identify cancerous regions associated to basal cell carcinoma. The graph features constitute a quantitative measure of the architectural tissue organization. Results show that graph topological features extracted from a nuclei based distance graph, particularly those related to local density, have a high predictive value in the automated detection of basal cell carcinoma. The method was evaluated using a leave-one-out validation scheme in a set of 9 skin Whole Slide Images obtaining a 0.76 F-score in distinguishing basal cell carcinoma regions in skin tissue whole slide images.

In geometric control model based methods, it is well known that if a nonlinear system has its relative degree equal to the number of states in the neighborhood of a point of equilibrium, thus it is possible to perform a coordinate transformation and a state feedback that transforms the nonlinear system into a linear and controllable one. Stabilization of this system is possible by a simple linear control technique. This method is based on the exact cancellation of nonlinear terms. If the parameters are either time-varying or there is uncertainty in the nonlinear terms of the model, then cancellation would not be longer accurate. The modification of the control method with adaptive parameters, makes asymptotically exact the cancellation of the nonlinear terms and maintains the efficiency of the transformation. This paper presents a nonlinear adaptive control method based on exact linearization techniques applied to the automation of blood glucose regulation in Type-1 diabetes. Using continuous blood glucose monitoring as the input, the method provides the insulin infusion function as the output. Since the insulin infusion calculated drives blood glucose to normal levels, the method mimics the healthy pancreas function and it could be applied in artificial pancreas to control blood glucose in type 1 diabetic patients.

A marked increase in the population aged 60 years and over is evident; the proportion of the older adult population will rise 18.6% in 2025. On the other hand, obesity, metabolic syndrome (MS), diabetes and insulin resistance (or low insulin sensitivity-IS) are diseases related to lifestyle, they have become a social and public health problem. IS is the ability of cells to react due to insulin´s presence; when this ability is diminished, low insulin sensitivity or insulin resistance (IR) is considered. Studies show that IS decreases with age, though no one knows exactly if it is directly due to aging or changes in muscle mass. IS can be determined using direct or indirect methods. This paper aims to propose an insulin sensitivity method design from anthropometries and lipid measures. The methodology consist in a simple correspondence analysis for determine the variables, and a parametrical optimization using Avignon method as optimal function. The database used is composed by 120 Ecuadorian older adults with and without MS. The results show that the proposed optimized method got a better correlation with Avignon compared to non-optimized method. The proposed method could discriminate between subjects with and without IR and with and without MS. This is an important contribution since other methods like HOMA-IR, which is the most used in clinical practice, cannot find these differences, this means that HOMA-IR is not sensitive for IS estimation in elderly people. Future works will focus in the determination of cutoffs for insulin resistance diagnosis in the proposed method.

Diabetic retinopathy has several clinical data sources for medical diagnosis, but the lack of tools to process the data generates a subjective and unclear diagnosis. The use of convolutional networks to analyze and extract features in eye fundus images may help with an automatic detection to support medical personnel in the grading of diabetic retinopathy. This paper presents a description of convolutional neural networks as a good methodology to detect and discriminate between exudate and healthy regions in eye fundus images.

Most current systems for automated glaucoma detection in fundus images rely on segmentation-based features, which are known to be influenced by the underlying segmentation methods. Convolutional Neural Networks (CNNs) are powerful tools for solving image classification tasks as they are able to learn highly discriminative features from raw pixel intensities. However, their applicability to medical image analysis is limited by the non-availability of large sets of annotated data required for training. In this article we present results of analysis of the viability of using CNNs that are pre-trained from non-medical data for automated glaucoma detection. Two different CNNs, namely OverFeat and VGG-S, were applied to fundus images to generate feature vectors. Preprocessing techniques such as vessel inpainting, contrast-limited adaptive histogram equalization (CLAHE) or cropping around the optic nerve head (ONH) area were explored within this framework to evaluate the improvement in feature discrimination, combined with both ℓ₁ and ℓ₂ regularized logistic regression models. Results on the Drishti-GS1 dataset, evaluated in terms of area under the average ROC curve, suggests the viability of this approach and offer significant evidence of the importance of well-chosen image pre-processing for transfer learning when the amount of data is not sufficient for fine-tuning the network.

Outcomes after traumatic brain injury (TBI) are variable and only partially predicted by acute injury factors. With rich datasets, we can examine how numerous factors – cognitive scores, acute injury variables, demographic variables, and brain imaging variables – are interrelated and aid in outcome prediction. To help study this rich data, we applied CorEx, a novel method for unsupervised machine learning. CorEx decodes the hierarchical structure, identifying latent causes of dependence in the data. It groups predictor variables based on their joint information and inter-dependence. We examined 21 TBI patients 2-5 months post-injury along with healthy controls; both groups were assessed again 12 months later. Although we were limited in the number of participants, this tool for exploratory analysis found potential relationships between change in cognitive scores over the 12-month period and baseline brain volumes. Certain regional brain volumes measured post-injury could serve as predictors of patient recovery. As future planned analyses will examine greater sample sizes, we hope to perform follow-up statistical analysis of variables identified by CorEx in independent data.

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder with atypical cortical maturation leading to a deficiency in social cognition and language. Numerous studies have attempted to classify ASD using brain measurements such as cortical thickness, surface area, or volume with promising results. However, the underpowered sample sizes of these studies limit external validity and generalizability at the population level. Large scale collaborations such as Enhancing NeuroImaging Genetics through Meta Analysis (ENIGMA) or the Autism Brain Imaging Data Exchange (ABIDE) aim to bring together like-minded scientists to further improve investigations into brain disorders. To the best of our knowledge, this study represents the largest classification analysis for detection of ASD vs. healthy age and sex matched controls using cortical thickness brain parcellations and intracranial volume normalized surface area and subcortical volumes. We were able to increase classification accuracy overall from 56% to 60% and for females only by 6%. These novel findings using Evolving Partitions to Improve Connectomics (EPIC) underscore the importance of large-scale data-driven approaches and collaborations in the discovery of brain disorders.

Autism Spectrum Disorder (ASD) is a very complex neuro-developmental entity characterized by a wide range of signs. The high variability of reported anatomical changes has arisen the interest of the community to characterize the different patterns of the disorder. Studies so far have focused on measuring the volume of the cerebral cortex as well as the inner brain regions of the brain, and some studies have described consistent changes. This paper presents an automatic method that separates cases with autism from controls in a population between 18 to 35 years extracted from the open database Autism Brain Imaging Data Exchange (ABIDE). The method starts by segmenting a new case, using the delineations associated to the template MNI152. For doing so, the template is non rigidly registered to the input brain. Once these cortical and sub-cortical regions are available, each region is characterized by the histogram of intensities which is normalized. The Kullback-Leibler distance is used as a metric for training a binary SVM classifier, region per region. The highest discrimination values were found for the Right Superior Temporal Gyrus, region which the Area is Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) curve was 0.67.

Compared to many neurological disorders, for which imaging biomarkers are often available, there are no accepted imaging biomarkers to assist in the diagnosis of major depressive disorder (MDD). One major barrier to understanding MDD has been the lack of a practical and efficient platform for collaborative efforts across multiple data centers; integrating the knowledge from different centers should make it easier to identify characteristic measures that are consistently associated with the illness. Here we applied our newly developed “distributed Lasso” method to brain MRI data from multiple centers to perform feature selection and classification. Over 1,000 participants were involved in the study; our results indicate the potential of the proposed framework to enable large-scale collaborative data analysis in the future.

This work demonstrates the usage of Convolutional Neural Networks (CNNs) to explore and identify the brain regions most contributing to Alzheimer’s disease in two-dimensional images extracted from structural magnetic resonance (MRI) images. In a first stage, we set up different CNN configurations which are trained in a supervised mode reaching classification accuracy similar to that in other works. Then, the best performing CNN is chosen and we create brain models for each filter at the CNN first layer as they convolve throughout MRI images of patient cases. The brain models are further explored as their corresponding filter activations throughout brain regions reveals different anatomical patterns for different patient class, and thus, allowing us to identify the CNN filters with greatest discriminating power and which brain regions contribute most. Specifically, the CNN shows the largest differentiation between patients in the frontal pole region, which is known to host intellectual deficits related to the disease. This shows how CNNs could be used to provide interpretability on Alzheimer’s and constitute an additional tool to support decision making in clinical practice.

Heart diseases are one of the most important causes of death in the Western world. It is, then, important to implement algorithms to aid the specialist in analyzing the heart motion. We propose a new strategy to estimate the cardiac motion through a 3D optical flow differential technique that uses the Steered Hermite transform (SHT). SHT is a tool that performs a decomposition of the images in a base that model the visual patterns used by the human vision system (HSV) for processing the information. The 3D + t analysis allows to describe most of motions of the heart, for example, the twisting motion that takes place on every beat cycle and to identify abnormalities of the heart walls. Our proposal was tested on two phantoms and on two sequences of cardiac CT images corresponding to two different patients. We evaluate our method using a reconstruction schema, for this, the resulting 3D optical flow was applied over the volume at time t to obtain a estimated volume at time t + 1. We compared our 3D optical flow approach to the classical Horn and Shunk's 3D algorithm for different levels of noise.

This article presents the Core Expansion method to automatically detect border-zone corridors in MRI images of the left ventricle, to serve as guidance to Ventricular Tachycardia (VT) ablation. The method relies on the fact that the different gray level intensities of Delayed Contrast-Enhanced Magnetic Resonance Images (DE-MRI) encode information about the fibrotic tissue. These differences in intensities among tissue types allow separating dense scar from healthy areas of the myocardium, and identify the border-zone region. After generating an onion-like layer-based 3D model of the left ventricle, the method detects potential corridors in the border-zone that can become electrical circuits of low conductivity. These circuits can be responsible for arrhythmic events. The method has been tested both in phantoms and patients. In patients there was a high degree of correlation between the channels detected and those visually identified by an expert on the MRI. Whenever electroanatomical maps were available post-intervention, the MRI detected channels were found to have a high degree of correlation with them.

Background: Intravascular ultrasound (IVUS) provides axial gray-scale images, allowing the assessment of vessel morphology and tissues. Automated segmentation of lumen-intima and media-adventitia interfaces is valuable to identify artery occlusion.

Purpose: Bifurcations, shadows and echogenic plaques usually affect proper segmentation of the vessel wall. Thus, identification of these morphological structures is an advisable step when developing segmentation techniques, which have been dealing with this issue by using different features and methods in the past. The aim of this work is to develop a simultaneous classification method for IVUS image sectors into bifurcations, shadows, echogenic plaques and normal, as an intermediate step for the arterial wall segmentation.

Methods: A 22-dimensional feature vector, mainly composed by current existing methods, is computed for each column in the polar image. To deal with this multiclass classification problem, Random Forest (RF) is used as classifier. Due to the high skewness of the problem, RFs are successively trained by resampling the training data, specifically the majority class.

Results: Fscore reaches 0.62, when the RF is trained with 15% of the normal samples of the training set. Thresholds found in the RF are close to the previously reported values for the features in the literature.

Conclusion: Random Forest demonstrates good performance to classify morphological structures in IVUS. Random undersampling for training was useful to deal with the imbalanced data, and to manage the trade-off between precision and recall of minority classes. However, better features must be developed to improve the classification of the structures, specially in the case of the echogenic plaque.

Ultrasound is widely used as an inexpensive, real-time method for imaging vascular tissue. However, sonographs often lack automatic or semi-automatic software for measuring vascular diameter precisely, especially in low- and mid-income countries or institutions. Tools can be developed to perform this task, but they must be validated before being accepted for clinic use. For that purpose, in this work we present low-cost phantoms that resemble vascular tissue when subjected to ultrasound. Several materials are analysed and a step-by-step recipe for building a simple phantom is presented. Qualitatively, models were imaged by an ultrasound expert physician, and several characteristic are assessed. Quantitatively, a comparison between ultrasound and caliper measurements of the phantoms is presented. Finally, a discussion about the results and the recommended materials for low-cost vascular phantoms is carried out.

Introduction: Vascular reactivity (VR) can be defined as the arterial ability to react to vasoactive stimulus in terms of vasodilator or vasoconstrictor responses, where blood vessel tone is determined both by smooth muscle and endothelial functions (EF). Flow mediated dilation (FMD) and Peripheral Arterial Tonometry (PAT) measurements are recognized as the most relevant techniques for VR and EF evaluation. Objective: To design an integrated system for vascular analysis based on the limitations reported in FMD technique and incorporating PAT as a simultaneous measure. Materials and Methods: A complete software platform and a specific fastening system were proposed, in order to obtain simultaneous measurements of continuous arterial pressure (by applying the applanation tonometry technique) and continuous vessel diameter variations (by processing B-mode ultrasound images) at the same vascular site. Results: Beat to beat pressure-diameter loops were obtained, along with the assessment of stiffness indices and pulse transit time values. The volunteer who used the system felt comfortable and relaxed, unworried about making about involuntary movements Conclusion: A preliminary study was carried out using this new module designed for the acquisition, realtime processing, visualization and storage of vascular behavior related data.

The Picture Administration and Communications System (PACS) was designed to replace the old film archiving system in hospitals in order to store and move varying medical image modalities. Using the standard Internet transport protocol, PACS creators designed a robust digital signaling platform to optimize media use, availability, and confidentiality. Nowadays PACS has become ubiquitous in medical facilities but lacks imaging analytical capabilities. A myriad of initiatives have been launched in the hope of achieving this goal, but current solutions face issues with security and ease-of-use that have precluded their widespread adoption.

Here, we present a PACS-based image processing tool that safeguards patient confidentiality, is user-friendly and is easy to implement. The final product is platform-independent, has a small degree of intrusiveness and is well suited to clinical and research work flows.

State-of-the-art data analysis methods in genetics and related fields have advanced beyond massively univariate analyses. However, these methods suffer from the limited amount of data available at a single research site. Recent large-scale multi-centric imaging-genetic studies, such as ENIGMA, have to rely on meta-analysis of mass univariate models to achieve critical sample sizes for uncovering statistically significant associations. Indeed, model parameters, but not data, can be securely and anonymously shared between partners. We propose here partial least squares (PLS) as a multivariate imaging-genetics model in meta-studies. In particular, we propose an online estimation approach to partial least squares for the sequential estimation of the model parameters in data batches, based on an approximation of the singular value decomposition (SVD) of partitioned covariance matrices. We applied the proposed approach to the challenging problem of modeling the association between 1,167,117 genetic markers (SNPs, single nucleotide polymorphisms) and the brain cortical and sub-cortical atrophy (354,804 anatomical surface features) in a cohort of 639 individuals from the Alzheimer's Disease Neuroimaging Initiative. We compared two different modeling strategies (sequential- and meta-PLS) to the classic non-distributed PLS. Both strategies exhibited only minimal approximation errors of model parameters. The proposed approaches pave the way to the application of multivariate models in large scale imaging-genetics meta-studies, and may lead to novel understandings of the complex brain phenotype-genotype interactions.

Radiomics workflows are high-throughput disease descriptive or predictive tools that extract mineable quantitative data of pathological phenotypes from standard-of-care grayscale images using advanced image processing algorithms. The success of these workflows rely on establishing large image datasets from which diverse disease descriptors can be extracted, with the expectation that large numbers may be able to overcome some of the inherent heterogeneities inherent in standard-of-care medical imaging workflows. Here, we present such a radiomics platform which relies on a combination of existing standard-of-care imaging clinical and research software as well as custom written code. The key components of the workflow include a file organization schema for centralized data storage, deployment of image registration strategies, and frontend GUI design for ease of use by the clinical researcher, all of which increase the transparency, flexibility, and portability of our radiomics platform. Widespread establishment of such radiomics platform can greatly revolutionize radiomics research and aid in successful translation into clinical decision support systems.

Presented are three preliminary studies completed using our proposed radiomics research workflow to investigate various diseases. The radiomics research workflow is modality and disease independent which allow it to serve as a general platform for medical image post-processing experimentation.

This article introduces a computer-aided solution for radiology education integrated with the clinic practice, inherited from modern technologies that facilitate the processing of large amounts of stored information. Radiology training may have several challenges such as image retrieval, extraction of knowledge, education towards solving problems and the interaction with huge repositories known as Picture Archiving and Communication Systems (PACS). This project proposes a user-based system that learns from user interaction, retrieving not just the requested information but recommending related cases and interesting images. The recommended images are retrieved using a Click-through rate (CTR) strategy for defining the most similar cases in the database. This is a fully web-based proposal, potentially useful at classroom or home, that allows students to develop the clinical skills needed in a more realistic scenario.

Traumatic brain injury (TBI) causes extensive damage to the white matter (WM) of the brain, which can be evaluated with diffusion-weighted magnetic resonance imaging (dMRI). Diffusion MRI can be used to map the WM tracts and their integrity, but offers limited understanding of the biochemical basis of any differences. Magnetic resonance spectroscopy (MRS) measures neural metabolites that reflect neuronal health, inflammation, demyelination, and other consequences of TBI. We combined whole-brain MRS with dMRI to investigate WM dysfunction following pediatric TBI, using “tract-based spectroscopy”. Deficits in N-acetylaspartate (NAA) correspond to regions of deficits in WM integrity, but choline showed minimal overlap with WM deficits. NAA is a marker of neuronal health, while choline is an inflammatory marker. A partial F-test showed that MRS measures improved our ability to predict long-term cognitive function. This is the first paper to combine MRS with dMRI-derived tracts on a whole-brain scale, offering insights into the biochemical correlates of WM tract dysfunction, following injury and potentially in other WM disorders.

White matter tractography is non-invasive method to study white matter microstructure within the brain and its connectivity across the different regions. Various neuro-degenerative diseases affect the white matter connectivity in the brain. In order to study the neurodegeneration and localize the affected fiber bundles, it is important to cluster the white matter fibers in an anatomically consistent manner. Clustering white matter fiber bundles in the brain is a challenging problem. The present approaches include region of interest (ROI) based clustering as well as template based clustering. A novel clustering technique using support vector machine framework is introduced. In this method, a conformal volumetric bijective mapping between the brain and the topologically equivalent sphere is established. The white matter fibers are then parameterized in this domain. Such a parameterization also introduces a spatial normalization without requiring any prior registration. We show that such a mapping is useful to learn statistical models of white matter fiber bundles and use it for clustering in a new subject.

Gray-level co-occurrence matrix (GLCM) is a statistical method widely used to characterize images and specifically, for Second Harmonic Generation (SHG) collagen images characterization. This method takes into account the spatial relationship between the image pixels, at specific angle. It is usually calculated for four orientations, at specific distances. Over these matrix, a textural feature function is calculated. Often, results of different orientations are compared or averaged to get a unique statistic parameter. In the present report, we will demonstrate the error that bring with this methodology, and following, we offer the correction formula. Preferred orientation of SHG images is proposed as structural property to characterize biological samples. For example, for determining the parallelism grade of collagen fibers regarding the ovarian epithelium. Here, we present a robust method to calculate this parameter, based on the two-dimensional Fourier transform. Finally, we show how these two elements help improve the discrimination between normal and pathological ovarian tissues.

Flow diverters are widely extended in clinical practice for intracraneal aneurysms treatment. They are formed by a dense mesh of braided wires that partially occludes the aneurysm neck and restores the blood flow into the parent vessel. The occlusion degree is highly dependant on the distribution of the wires under the aneurysm, which is affected by the vessel geometry. Nowadays, there are no clinical indicators of the covering ratio once the flow diverter is deployed. We propose a novel method for the simulation of the flow diverter local porosity before its deployment into the parent vessel. We validate the method on curved silicon models, obtaining a correlation of 0.9 between the simulated values and the measured porosity on the deployed flow diverter.

An innovative approach to treat abdominal aortic aneurysms, based on an endovascular aneurysm sealing system, claims to reduce both endoleak and graft migration with respect to conventional devices with proximal fixation technologies. However, the aortic bifurcation anatomy is significantly modified with this novel proposal and the hemodynamic influences on blood flow have not been addressed until now. In this work we evaluated the aortic fluid dynamics changes introduced after the implantation of a sealing device with respect to a conventional endograft on four adults with abdominal aorta aneurysms. An adaptive Geometrical Deformable Model was used for aortic segmentation and Finite Volume mesh generation. Inlet boundary conditions were set to reproduce normal physiological conditions at the abdominal aorta, and maximum pressure drop and maximum peak velocity for the models were estimated at 3 sections (proximal, mid and distal) using Computational Fluid Dynamics simulations. We found a systematic pressure increase in the proximal abdominal aorta segment for patients treated with the sealing device with respect to the more conventional endograft. Pressure values at the level of the renal arteries averaged a ≈3 mmHg pressure increase for the sealing device, compared to the ≈1 mmHg for the conventional device. Velocities inside the endograft were 4-fold higher for the sealing device with respect to the conventional device, reaching 0.41 m/s vs 0.13 m/s, respectively. Distal velocity also remained higher: 0.45 m/s vs 0.24 m/s, respectively. Although these results should be analyzed carefully due to the small number of participants, the orders of magnitude and tendencies evidence the influence that the novel sealing device has on aortic blood flow.

In radiotherapy for prostate cancer the bladder presents the largest inter-fraction shape variations during treatment resulting in random geometric uncertainties that may increase the risk of developing side-effects. In this setting, our interest is thus to propose a hierarchical population model, based on longitudinal data, to characterize bladder motion and deformation between fractions. This method is based on a principal component analysis (PCA) of bladder shapes to obtain the dominant eigenmodes that describe bladder geometric variations between fractions. However, PCA may not properly capture the latent structure of complex data like longitudinal data of organs with large inter and intra-patient shape variations. With this, we propose hierarchical modes to separate intra- and inter-patient bladder variability of the longitudinal data following a dimensionality reduction by means of spherical harmonics (SPHARM). The training data base was used to derive a top-level PCA model that describes the entire structure of the bladder surface space. This space was subsequently divided into subspaces by lower-level PCA models that describe their internal structures. The model was evaluated using a reconstruction error and compared with a conventional PCA model following leave-one-out cross validation.

Intracranial aneurysm treatment with flow diverters stent (FDs) is a minimally invasive approach for use in human patients. Because this treatment is strongly related to blood flow, flow simulation by CFD is an attractive method to study FDs. Such flow simulations generally define geometries of aneurysms and stents in the computation by creating calculation meshes in the fluid space. For the other hand, generating a mesh in porous media (PM) sometimes represents a smaller computational load than generating realistic stent geometries with CFD, particularly for the small gaps between stent struts. For this reason, PMs become attractive to simulate FDs. To find the proper parameters, we investigated Darcy-Forchheimer model for porous media. The model describes the relation between the pressure drop and flow velocity considering a viscous permeability (linear model's term), and an inertial permeability (quadratic model's term). Finally, two stage studies were performed. First, we verified flow model validity at different angles in known flow conditions. Second, model validation was checked for a channel with no-slip boundary conditions. Results indicate that resistance calculated according to model has a difference of less than 3.5 % which is appropriate to characterize the FDs.

Crawling wave sonoelastography (CWS) provides quantitative stiffness information from an examined tissue by the application of two mechanical vibrations that generates an interference pattern which is analyzed to reconstruct a shear wave speed (SWS) image. While CWS performance using parallel excitation setup (PES) has been widely studied in the literature, its implementation is complicated to apply in tissue (i.e. breast, liver) of the technique in human tissue often requires a normal excitation setup. The aim of this study is to validate the SWS estimation using a normal excitation setup with two different types of coupling attachments for the surface -i.e. surface plate (SP), rounded head (RH)- and perform a quantitative comparison among the three techniques (PES,SP,RH) in homogeneous and heterogeneous gelatin phantoms. The comparison of the three excitation methods was performed using signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), coefficient of variation(CV) and bias over a range of excitation frequencies. (PES:3.74 ±0.14 m/s, SP: 3.43 ±0.14 m/s and RH: 3.32 ± 0.07 m/s) and in the inclusion (PES:4.96 ± 0.18 m/s, SP: 4.88 ± 0.32 m/s and RH: 4.64 ± 0.29 m/s) show that both normal excitation setups are comparable to the PES to estimate the SWS. This suggests that this type of setup could be used for ex vivo and in vivo analysis.

Magnetic resonance imaging (MRI) is widely used in medicine nowadays, yet a significant disadvantage is the amount of artifacts that affect the image during the acquisition process. This paper presents a strategy for automatic damage detection when the image is altered by movement or there is a loss of information due to magnetic susceptibility. This approach uses a conventional SV D to detect the variability between slices of the image and a region of damaged voxels within the volume. Using a simple derivative algorithm, the method was tested in several cases automatically revealing the distortion's location with a performance of 74% for slice damage and 55% for the volume's damaged region.

Studies have shown that the brain is constituted by anatomically segregated and functionally specific regions working in synergy as a complex network. In this context, the brain at rest does not passively retrieve environmental information and respond but instead it maintains an active representation modulated by sensory information. Using independent component analysis (ICA) over resting state recordings a discrete set of resting state networks (RSNs) has been found, which proven to be systematically present across individuals and to be modified by the state of consciousness and also in disease. ICA's main drawback is that its output consists of a series of 3D z-score maps where noise and physiological components are randomly mixed. In this work we present a computational method composed by an ICA-based noise filtering preprocessing pipeline and a template-based identification algorithm that combines spatial comparison metrics through a voting system developed to find RSNs in a subject-by-subject basis. To validate it, we use a publicly available dataset consisting of 75 resting state fMRI sessions from 25 participants scanned three different times each one. For most common RSNs the correct candidate won the voting 93% of the times and it was voted at least once in 99%. Then we probe within-subject consistency in detected RSNs by showing augmented correlation in networks from the same subject. Finally, by comparing obtained mean RSNs with the ones from nearly 30,000 participants we show that our method constitutes a personalized-medicine oriented approach to shorten the gap between RSN research and clinical applications.

In this paper, a Bayesian super resolution (SR) method obtains high resolution (HR) brain Diffusion-Weighted Magnetic Resonance Imaging (DMRI) images from degraded low resolution (LR) images. Under a Bayesian formulation, the unknown HR image, the acquisition process and the unknown parameters are modeled as stochastic processes. The likelihood model is modeled using a Gaussian distribution to estimate the error between the a linear representation and the observations. The prior is introduced as a Multivariate Gaussian Distribution, for which the inverse of the covariance matrix is approximated by Laplacian-like functions that model the local relationships, capturing thereby non-homogeneous relationships between neighbor intensities. Experimental results show the method outperforms the base line by 2.56 dB when using PSNR as a metric of quality in a set of 35 cases.

Fractional anisotropy derived from the single-tensor model (FA^DTI) in diffusion MRI (dMRI) is the most widely used metric to characterize white matter (WM) micro-architecture in disease, despite known limitations in regions with extensive fiber crossing. Models such as the tensor distribution function (TDF), which represents the diffusion profile as a probabilistic mixture of tensors, have been proposed to reconstruct multiple underlying fibers. Although complex HARDI acquisition protocols are rare in clinical studies, the TDF and TDF-derived scalar FA metric (FA^TDF) have been shown to be advantageous even for data with modest angular resolution. However, further evaluation and validation of the metric are necessary. Here we compared the test-retest reliability of FA^TDF and FA^DTI in clinical quality data by computing the intra-class correlation (ICC) between dMRI scans collected 3 months apart. When FA^TDF and FA^DTI were calculated at various angular resolutions, FA^TDF ICC in both the corpus callosum and in a full axial slice were consistently more stable across scans, as compared to FA^DTI.

Functional MRI (fMRI) studies typically analyze data by applying a single function – across the entire brain – to relate what is measured (blood oxygenation fluctuations) to the underlying neural activity. However, this hemodynamic response function (HRF), is known to vary considerably across brain regions in healthy individuals, and even more prominently in clinical populations (e.g., AIDS, Alzheimer’s). An improved characterization of HRF variability would improve cognitive science experimentation, effective connectivity analysis, and may be crucial for early detection of certain diseases. Here, a method is suggested for altering stimulus presentation timing during task related fMRI experiments that aims to maximize characterization of HRF variability while minimizing the number of trials required to accomplish this. To do so, d-optimality constraints are applied for sparse sampling of the HRF in the temporal domain. We first demonstrate this approach using simulated data over a range of background noise fluctuations. Using simulated data, we were able to recover HRF signal estimates with <10% sum of squared error (SSE) using 73% and 47% less stimulus events using D-optimal sampling compared to fixed or random designs respectively. We then utilized this method for designing the stimulus timing in an event-related fMRI experiment. Empirically, we were able to detect the initial dip in 53% of subjects, a part of the HRF signal that is thought to reflect oxygen usage and often obscured when using conventional experimental design paradigms.

Attentional processing in the absence of conscious vision has yet to be understood in terms of neurophysiological mechanisms. Therefore, we used the visual mismatch negativity (vMMN) to determine if automatic detection of changes can be followed by an attentional switch without visual awareness. Random moving dots changing in direction were presented in the periphery, while participants carried out an effortful Stroop test in the central visual field to fully engage their attention on this primary task. The results revealed a posterior vMMN at 200 ms that was maximal in the parietal regions, revealing an automatic detection of change in the absence of visual awareness related to a dorsal/magnocellular pathway. Moreover, a frontal and central positivity, with a more pronounced activity in the left frontal areas was found at 300 ms possibly reflecting (1) unconscious attentional switch, (2) inhibition of explicit attentional switch by the left frontal areas acting on the right frontal areas via interhemispheric connections (3) inhibition of explicit attentional switch by the frontal areas acting on the central area via top-down connections. In conclusion, our results showed that vMMN could be a useful tool to study detection of changes and attentional mechanisms in the absence of visual consciousness.