Development and characterization of deep learning techniques for neuroimaging data

Deep learning methods are extremely promising machine learning tools to analyze neuroimaging data. However, their potential use in clinical settings is limited because of the existing challenges of applying these methods to neuroimaging data. In this study, first a data leakage type caused by slice-level data split that is introduced during training and validation of a 2D CNN is surveyed and a quantitative assessment of the model’s performance overestimation is presented. Second, an interpretable, leakage-fee deep learning software written in a python language with a wide range of options has been developed to conduct both classification and regression analysis. The software was applied to the study of mild cognitive impairment (MCI) in patients with small vessel disease (SVD) using multi-parametric MRI data where the cognitive performance of 58 patients measured by five neuropsychological tests is predicted using a multi-input CNN model taking brain image and demographic data. Each of the cognitive test scores was predicted using different MRI-derived features. As MCI due to SVD has been hypothesized to be the effect of white matter damage, DTI-derived features MD and FA produced the best prediction outcome of the TMT-A score which is consistent with the existing literature. In a second study, an interpretable deep learning system aimed at 1) classifying Alzheimer disease and healthy subjects 2) examining the neural correlates of the disease that causes a cognitive decline in AD patients using CNN visualization tools and 3) highlighting the potential of interpretability techniques to capture a biased deep learning model is developed. Structural magnetic resonance imaging (MRI) data of 200 subjects was used by the proposed CNN model which was trained using a transfer learning-based approach producing a balanced accuracy of 71.6%. Brain regions in the frontal and parietal lobe showing the cerebral cortex atrophy were highlighted by the visualization tools.

Evaluation of plaque composition by intravascular ultrasound ""virtual histology"": the impact of dense calcium on the measurement of necrotic tissue

Aims: We aimed to evaluate if the co-localisation of calcium and necrosis in intravascular ultrasound virtual histology (IVUS-VH) is due to artefact, and whether this effect can be mathematically estimated. Methods and results: We hypothesised that, in case calcium induces an artefactual coding of necrosis, any addition in calcium content would generate an artificial increment in the necrotic tissue. Stent struts were used to simulate the ""added calcium"". The change in the amount and in the spatial localisation of necrotic tissue was evaluated before and after stenting (n=17 coronary lesions) by means of a especially developed imaging software. The area of ""calcium"" increased from a median of 0.04 mm(2) at baseline to 0.76 mm(2) after stenting (p<0.01). In parallel the median necrotic content increased from 0.19 mm(2) to 0.59 mm(2) (p<0.01). The ""added"" calcium strongly predicted a proportional increase in necrosis-coded tissue in the areas surrounding the calcium-like spots (model R(2)=0.70; p<0.001). Conclusions: Artificial addition of calcium-like elements to the atherosclerotic plaque led to an increase in necrotic tissue in virtual histology that is probably artefactual. The overestimation of necrotic tissue by calcium strictly followed a linear pattern, indicating that it may be amenable to mathematical correction.

Gamma ventral capsulotomy for treatment of resistant obsessive-compulsive disorder: A structural MRI pilot prospective study

Objective: The purpose of this study was to investigate regional structural abnormalities in the brains of five patients with refractory obsessive-compulsive disorder (OCD) submitted to gamma ventral capsulotomy. Methods: We acquired morphometric magnetic resonance imaging (MRI) data before and after 1 year of radiosurgery using a 1.5-T MRI scanner. Images were spatially normalized and segmented using optimized voxel-based morphometry (VBM) methods. Voxelwise statistical comparisons between pre- and post-surgery MRI scans were performed using a general linear model. Findings in regions predicted a priori to show volumetric changes (orbitofrontal cortex, anterior cingulate gyrus, basal ganglia and thalamus) were reported as significant if surpassing a statistical threshold of p<0.001 (uncorrected for multiple comparisons). Results: We detected a significant regional postoperative increase in gray matter volume in the right inferior frontal gyri (Brodmann area 47, BA47) when comparing all patients pre and postoperatively. Conclusions: Our results support the current theory of frontal-striatal-thalamic-cortical (FSTC) circuitry involvement in OCD pathogenesis. Gamma ventral capsulotomy is associated with neurobiological changes in the inferior orbitofrontal cortex in refractory OCD patients. (C) 2008 Elsevier Ireland Ltd. All rights reserved.

Detection of Auditory Cortex Activity by fMRI Using a Dependent Component Analysis

Functional MRI (fMRI) data often have low signal-to-noise-ratio (SNR) and are contaminated by strong interference from other physiological sources. A promising tool for extracting signals, even under low SNR conditions, is blind source separation (BSS), or independent component analysis (ICA). BSS is based on the assumption that the detected signals are a mixture of a number of independent source signals that are linearly combined via an unknown mixing matrix. BSS seeks to determine the mixing matrix to recover the source signals based on principles of statistical independence. In most cases, extraction of all sources is unnecessary; instead, a priori information can be applied to extract only the signal of interest. Herein we propose an algorithm based on a variation of ICA, called Dependent Component Analysis (DCA), where the signal of interest is extracted using a time delay obtained from an autocorrelation analysis. We applied such method to inspect functional Magnetic Resonance Imaging (fMRI) data, aiming to find the hemodynamic response that follows neuronal activation from an auditory stimulation, in human subjects. The method localized a significant signal modulation in cortical regions corresponding to the primary auditory cortex. The results obtained by DCA were also compared to those of the General Linear Model (GLM), which is the most widely used method to analyze fMRI datasets.

Quantificação por imagem ponderada em difusão (DWI) das lesões tumorais da mama

Introdução – A técnica de Difusão por Ressonância Magnética (RM), ao avaliar o movimento das moléculas de água nos tecidos, permite inferir sobre a arquitetura dos mesmos, em particular relativamente à celularidade, volume celular e permeabilidade das membranas. O Coeficiente de Difusão Aparente (ADC) é um parâmetro quantificável da imagem ponderada em difusão (DWI). A sua análise poderá fornecer informação clínica adicional sobre estas lesões, sobretudo em relação à sua caracterização histológica. Objetivos – Caracterizar e diferenciar tipos e alguns subtipos de lesões mamárias através da análise do ADC. Metodologia – 20 Mulheres com 23 lesões mamárias foram submetidas a RM mamária: 3 lesões benignas (3 Fibroadenomas-FA) e 20 malignas (16 Carcinomas Ductais Invasivos-CDI, 2 Carcinomas Ductais In Situ-CDIS e 2 Carcinomas Invasivos Lobulares-CLI). Num equipamento 1.5T aplicou-se uma sequência de Difusão (b=0,50,250,500,750,1000 s/mm2). Obteve-se o ADC através do ajuste exponencial da intensidade de sinal das lesões em função do valor de b, fazendo-se corresponder os valores de ADC à respetiva caracterização histológica e posterior comparação com a literatura. Resultados e Discussão – As lesões malignas apresentaram ADCs significativamente (p=0,014) inferiores [(0,94±0,22)x10-3 mm2/s] aos das benignas [(1,43±0,25)x10-3 mm2/s]. A justificação pode residir no aumento da celularidade e consequente restrição da Difusão que se observa nas lesões malignas. Os CDI apresentaram ADCs baixos [(0,88±0,17)x10-3 mm2/s], enquanto que os CDIS apresentaram ADCs mais elevados [(1,33±0,29)x10-3 mm2/s]. Estes resultados estão de acordo com o facto dos CDIS estarem limitados aos ductos mamários, mantendo-se menos alterada a estrutura do tecido adjacente e resultando numa menor restrição à difusão que nos CDI. Verificaram-se diferenças significativas entre FA e CDI (p=0,010) e entre CDI e CDIS (p=0,049). Conclusões – O ADC possibilita a diferenciação entre lesões mamárias benignas e malignas, bem como entre alguns tipos histológicos. O desenvolvimento deste conceito pode representar um avanço no papel da RM na avaliação destas neoplasias. ABSTRACT - Introduction – The Magnetic Resonance (MR) diffusion technique measures the movement of water molecules in tissues. Therefore, it provides useful information about tissue architecture, specially regarding tissue cellularity, cell volume and membrane permeability. The quantification of diffusion weighted imaging (DWI) data is done by measuring the so-called. Apparent Diffusion Coefficient (ADC). This parameter provides additional clinical information about breast lesions, and potentially allows for in-vivo histological characterization. Purpose – To characterize and differentiate breast lesions through ADC analysis. Methodology – The study comprised 20 women, with 23 breast lesions: 3 benign lesions - 3 Fibroadenomas (FA); and 20 malignant - 16 Invasive Ductal Carcinomas (CDI), 2 Ductal Carcinomas In Situ (CDIS), 2 Invasive Lobular Carcinoma (CLI). On a 1.5T equipment a diffusion-weighted sequence with 6 b-values (b=0,50,250,500,750,1000 s/mm2) was used to examine the patients. ADC was obtained by fitting an exponential to data of lesion signal intensity vs. b values. A correspondence of ADC values to histological lesion characterization was done and finally, the results were comparison with the literature. Results and Discussion – Malignant lesions showed inferior ADCs significantly (p=0.014) lower ((0.94±0.22)x10-3 mm2/s) than the benign lesions ((1.43±0.25)x10-3 mm2/s). This may be associated to increasead cellularity in malignant lesions, which result in higher tissue restriction to diffusion. CDI showed low ADC values ((0.88±0.17)x10-3 mm2/s), while the CDIS showed higher ADCs ((1.33±0.29)x10-3 mm2/s). These results agree with the fact that CDIS are limited to mammary ducts, maintaining a less altered neighboring tissue structure, which results in a lower restriction to diffusion than observed in CDI. Significant differences between FA and CDI (p=0.010) and between CDI and CDIS (p=0.049) were observed. Conclusion – The ADC parameter is able to differentiate between malignant and benign breast lesions, as well as between some histological types.

COMMIT: Convex Optimization Modeling for Microstructure Informed Tractography.

Tractography is a class of algorithms aiming at in vivo mapping the major neuronal pathways in the white matter from diffusion magnetic resonance imaging (MRI) data. These techniques offer a powerful tool to noninvasively investigate at the macroscopic scale the architecture of the neuronal connections of the brain. However, unfortunately, the reconstructions recovered with existing tractography algorithms are not really quantitative even though diffusion MRI is a quantitative modality by nature. As a matter of fact, several techniques have been proposed in recent years to estimate, at the voxel level, intrinsic microstructural features of the tissue, such as axonal density and diameter, by using multicompartment models. In this paper, we present a novel framework to reestablish the link between tractography and tissue microstructure. Starting from an input set of candidate fiber-tracts, which are estimated from the data using standard fiber-tracking techniques, we model the diffusion MRI signal in each voxel of the image as a linear combination of the restricted and hindered contributions generated in every location of the brain by these candidate tracts. Then, we seek for the global weight of each of them, i.e., the effective contribution or volume, such that they globally fit the measured signal at best. We demonstrate that these weights can be easily recovered by solving a global convex optimization problem and using efficient algorithms. The effectiveness of our approach has been evaluated both on a realistic phantom with known ground-truth and in vivo brain data. Results clearly demonstrate the benefits of the proposed formulation, opening new perspectives for a more quantitative and biologically plausible assessment of the structural connectivity of the brain.

Resting-state temporal synchronization networks emerge from connectivity topology and heterogeneity.

Spatial patterns of coherent activity across different brain areas have been identified during the resting-state fluctuations of the brain. However, recent studies indicate that resting-state activity is not stationary, but shows complex temporal dynamics. We were interested in the spatiotemporal dynamics of the phase interactions among resting-state fMRI BOLD signals from human subjects. We found that the global phase synchrony of the BOLD signals evolves on a characteristic ultra-slow (<0.01Hz) time scale, and that its temporal variations reflect the transient formation and dissolution of multiple communities of synchronized brain regions. Synchronized communities reoccurred intermittently in time and across scanning sessions. We found that the synchronization communities relate to previously defined functional networks known to be engaged in sensory-motor or cognitive function, called resting-state networks (RSNs), including the default mode network, the somato-motor network, the visual network, the auditory network, the cognitive control networks, the self-referential network, and combinations of these and other RSNs. We studied the mechanism originating the observed spatiotemporal synchronization dynamics by using a network model of phase oscillators connected through the brain's anatomical connectivity estimated using diffusion imaging human data. The model consistently approximates the temporal and spatial synchronization patterns of the empirical data, and reveals that multiple clusters that transiently synchronize and desynchronize emerge from the complex topology of anatomical connections, provided that oscillators are heterogeneous.

Validation of network communicability metrics for the analysis of brain structural networks.

Computational network analysis provides new methods to analyze the brain's structural organization based on diffusion imaging tractography data. Networks are characterized by global and local metrics that have recently given promising insights into diagnosis and the further understanding of psychiatric and neurologic disorders. Most of these metrics are based on the idea that information in a network flows along the shortest paths. In contrast to this notion, communicability is a broader measure of connectivity which assumes that information could flow along all possible paths between two nodes. In our work, the features of network metrics related to communicability were explored for the first time in the healthy structural brain network. In addition, the sensitivity of such metrics was analysed using simulated lesions to specific nodes and network connections. Results showed advantages of communicability over conventional metrics in detecting densely connected nodes as well as subsets of nodes vulnerable to lesions. In addition, communicability centrality was shown to be widely affected by the lesions and the changes were negatively correlated with the distance from lesion site. In summary, our analysis suggests that communicability metrics that may provide an insight into the integrative properties of the structural brain network and that these metrics may be useful for the analysis of brain networks in the presence of lesions. Nevertheless, the interpretation of communicability is not straightforward; hence these metrics should be used as a supplement to the more standard connectivity network metrics.

Prediction of Recanalization Trumps Prediction of Tissue Fate: The Penumbra: A Dual-edged Sword.

BACKGROUND AND PURPOSE: To determine whether infarct core or penumbra is the more significant predictor of outcome in acute ischemic stroke, and whether the results are affected by the statistical method used. METHODS: Clinical and imaging data were collected in 165 patients with acute ischemic stroke. We reviewed the noncontrast head computed tomography (CT) to determine the Alberta Score Program Early CT score and assess for hyperdense middle cerebral artery. We reviewed CT-angiogram for site of occlusion and collateral flow score. From perfusion-CT, we calculated the volumes of infarct core and ischemic penumbra. Recanalization status was assessed on early follow-up imaging. Clinical data included age, several time points, National Institutes of Health Stroke Scale at admission, treatment type, and modified Rankin score at 90 days. Two multivariate regression analyses were conducted to determine which variables predicted outcome best. In the first analysis, we did not include recanalization status among the potential predicting variables. In the second, we included recanalization status and its interaction between perfusion-CT variables. RESULTS: Among the 165 study patients, 76 had a good outcome (modified Rankin score ≤2) and 89 had a poor outcome (modified Rankin score >2). In our first analysis, the most important predictors were age (P<0.001) and National Institutes of Health Stroke Scale at admission (P=0.001). The imaging variables were not important predictors of outcome (P>0.05). In the second analysis, when the recanalization status and its interaction with perfusion-CT variables were included, recanalization status and perfusion-CT penumbra volume became the significant predictors (P<0.001). CONCLUSIONS: Imaging prediction of tissue fate, more specifically imaging of the ischemic penumbra, matters only if recanalization can also be predicted.

Caracterização mineralógica de solos tropicais por sensoriamento remoto hiperespectral

Dados hiperespectrais coletados no Brasil pelo sensor AVIRIS (Airborne Visible/Infrared Imaging Spectrometer) foram utilizados para a caracterização espectral de uma típica cena agropastoril e para testar o uso da técnica Spectral Feature Fitting (SFF) na identificação de minerais argilosos na imagem. Utilizou-se um modelo linear de mistura espectral, usando como membros de referência a vegetação verde e seca, a água, e os solos Nitossolo Vermelho, Latossolo Vermelho e Neossolo Quartzarênico órtico. Na identificação dos minerais, foram selecionados espectros de referência da biblioteca espectral do JPL/NASA. Os espectros dos pixels e das referências foram normalizados pelo método do contínuo espectral, entre 2.100 e 2.330 nm, e depois comparados quanto à similaridade com o uso da técnica SFF. A caulinita predomina na cena, cuja identificação remota é dependente do tipo de solo e das proporções dos componentes da cena no interior do pixels. Os melhores resultados foram obtidos em solos de reflectância intermediária a alta e em pixels com valor de abundância da fração solo superior a 70%. Isto ocorreu devido, respectivamente, à menor quantidade de substâncias opacas nestes solos e à redução nos pixels dos efeitos espectrais da lignina-celulose. Estes fatores tendem a mascarar as bandas de absorção das argilas.

Altered processing of contextual information during fear extinction in PTSD: an fMRI study.

Medial prefrontal cortical areas have been hypothesized to underlie altered contextual processing in posttraumatic stress disorder (PTSD). We investigated brain signaling of contextual information in this disorder. Eighteen PTSD subjects and 16 healthy trauma-exposed subjects underwent a two-day fear conditioning and extinction paradigm. On day 1, within visual context A, a conditioned stimulus (CS) was followed 60% of the time by an electric shock (conditioning). The conditioned response was then extinguished (extinction learning) in context B. On day 2, recall of the extinction memory was tested in context B. Skin conductance response (SCR) and functional magnetic resonance imaging (fMRI) data were collected during context presentations. There were no SCR group differences in any context presentation. Concerning fMRI data, during late conditioning, when context A signaled danger, PTSD subjects showed dorsal anterior cingulate cortical (dACC) hyperactivation. During early extinction, when context B had not yet fully acquired signal value for safety, PTSD subjects still showed dACC hyperactivation. During late extinction, when context B had come to signal safety, they showed ventromedial prefrontal cortex (vmPFC) hypoactivation. During early extinction recall, when context B signaled safety, they showed both vmPFC hypoactivation and dACC hyperactivation. These findings suggest that PTSD subjects show alterations in the processing of contextual information related to danger and safety. This impairment is manifest even prior to a physiologically-measured, cue-elicited fear response, and characterized by hypoactivation in vmPFC and hyperactivation in dACC.

Physiological noise in human cerebellar fMRI.

OBJECTIVES: To compare physiological noise contributions in cerebellar and cerebral regions of interest in high-resolution functional magnetic resonance imaging (fMRI) data acquired at 7T, to estimate the need for physiological noise removal in cerebellar fMRI. MATERIALS AND METHODS: Signal fluctuations in high resolution (1 mm isotropic) 7T fMRI data were attributed to one of the following categories: task-induced BOLD changes, slow drift, signal changes correlated with the cardiac and respiratory cycles, signal changes related to the cardiac rate and respiratory volume per unit of time or other. [Formula: see text] values for all categories were compared across regions of interest. RESULTS: In this high-resolution data, signal fluctuations related to the phase of the cardiac cycle and cardiac rate were shown to be significant, but comparable between cerebellar and cerebral regions of interest. However, respiratory related signal fluctuations were increased in the cerebellar regions, with explained variances that were up to 80 % higher than for the primary motor cortex region. CONCLUSION: Even at a millimetre spatial resolution, significant correlations with both cardiac and respiratory RETROICOR components were found in all healthy volunteer data. Therefore, physiological noise correction is highly likely to improve the temporal signal-to-noise ratio (SNR) for cerebellar fMRI at 7T, even at high spatial resolution.

Capçal extrusor per a un sistema de tecnologia additiva

El departament de GREP de la UdG disposa d’un sistema de tecnologia additiva, laFab@home model 1, una impressora 3D o també anomenada de prototipat ràpid. Ésuna màquina de dimensions reduïdes, versàtil i de cost reduït, que a partir d’un modelgenerat a CAD, es capaç de fabricar objectes sòlids tridimensionals mitjançant lasuperposició de capes horitzontals.Les impressores 3D neixen de la necessitat de realitzar, en un temps relativament curt,proves de geometries complexes, que executades en un procés de manofacturaciónormal pot ser llarg i d’elevat cost.Gran part de les màquines de característiques similars a la Fab@home, tenen modelspreparats fabricar / extrudir peces a partir de silicones, aliments o polímers. Ara bé, laconformació per extrusió de plàstic és a partir de polímer verge en forma de fil.El present projecte té l’objectiu de dissenyar un sistema de conformació que permetimanufacturar polímer en forma de gra i s’adapti al funcionament de la Fab@home, amb l’argot corresponent, es tracta d’adaptar la Fab@home al fused depositionmodeling (FDM)

Electro-Optical Properties of Metal Phthalocyanines and Naphthalocyanines

The thesis deals with the preparation of chemical, optical, thermal and electrical characterization of five compounds, namely metal free naphthalocyanine, vanadyl napthalocyanine, zinc naphlocyanine, europium dinaphthalocyanine, and europium diphthalocyanine in the pristine and iodine-doped forms. Two important technological properties of these compounds have been investigated. The electrical properties are important in applications sensors and semiconductor lasers. Opto-thermal properties assume significance for optical imaging and data recording. The electrical properties were investigated by dc and ac techniques. This work has revealed some novel information on the conduction mechanism in five macrocyclic compounds and their iodine-doped forms. Also useful data on the thermal diffusivity of the target compounds have been obtained by optical techniques.

Comparison of bottom-up protein identification methods

With the rapid development of proteomics, a number of different methods appeared for the basic task of protein identification. We made a simple comparison between a common liquid chromatography-tandem mass spectrometry (LC-MS/MS) workflow using an ion trap mass spectrometer and a combined LC-MS and LC-MS/MS method using Fourier transform ion cyclotron resonance (FTICR) mass spectrometry and accurate peptide masses. To compare the two methods for protein identification, we grew and extracted proteins from E. coli using established protocols. Cystines were reduced and alkylated, and proteins digested by trypsin. The resulting peptide mixtures were separated by reversed-phase liquid chromatography using a 4 h gradient from 0 to 50% acetonitrile over a C18 reversed-phase column. The LC separation was coupled on-line to either a Bruker Esquire HCT ion trap or a Bruker 7 tesla APEX-Qe Qh-FTICR hybrid mass spectrometer. Data-dependent Qh-FTICR-MS/MS spectra were acquired using the quadrupole mass filter and collisionally induced dissociation into the external hexapole trap. Proteins were in both schemes identified by Mascot MS/MS ion searches and the peptides identified from these proteins in the FTICR MS/MS data were used for automatic internal calibration of the FTICR-MS data, together with ambient polydimethylcyclosiloxane ions.