Image segmentation and image registration are two fundamental problems in computer vision and medical image processing. In image segmentation, one seeks to partition an image into meaningful segments by assigning a label to each pixel indicating which segment it belongs to. In image registration, one seeks to recover a spatial transformation that geometrically aligns two or more images, which allows downstream image analyses in which the registered images share a coordinate system. Image processing pipelines typically apply these procedures sequentially even though the segmentation of an image could improve its registration and registration of an image could improve its segmentation. With an appropriate parametrization, one can view these two tasks as an inference problem in which the spatial transformation and segmentation are latent variables. In this work, registration and segmentation are integrated through a hierarchical Bayesian generative framework. The framework models the data generating process of a set of magnetic resonance (MR) images of ischemic stroke lesioned brains. Under this framework, we simultaneously estimate a lesion tissue segmentation along with a spatial diffeomorphic transformation that maps a subject image into spatial correspondence with a healthy template image. The framework is evaluated on two-dimensional images both real and synthetic. Experimental results on real MR images show that simultaneous segmentation and registration can significantly improve the accuracy of lesion segmentation as well as the accuracy of registration near the lesion.

L’observation du système nerveux, de son métabolisme et de certaines de ses structures est possible grâce à la neuroimagerie. Une littérature importante issue du « neurodroit » véhicule des imaginaires et des fantasmes relatifs aux possibilités judiciaires qu’offriraient ces technologies. Qu’il s’agisse de détection du mensonge, d’identification cérébrale des individus dangereux ou encore de prédiction de comportements déviants, la neuroimagerie, en l’état actuel des technologies, ne peut pourtant être sérieusement conçue comme pouvant faire l’objet de telles applications. L’utilisation de la neuroimagerie dans le cadre d’expertises est néanmoins une réalité, dans les tribunaux canadiens comme dans la loi française. Cette thèse souligne que les conceptions des technologies dont témoignent les deux systèmes juridiques étudiés s’avèrent lacunaires, ce qui engendre des risques. Elle évoque les conditions du recours à une normativité extra-juridique, la normalisation technique, qui pourrait s’élaborer dans ce contexte controversé, et esquisse les traits d’un dialogue amélioré entre les normativités juridique et technologique. / Neuroimaging allows the observation of the nervous system, of both its metabolism and some of its structures. An important literature in “neurolaw” conveys illusions and fantaisies about the judicial possibilities that imaging technologies would contain. Whether it is about lies detection, cerebral identifications of dangerous individuals through their neurobiology or predictions of criminal behaviors, neuroimaging, in the current state of technologies, can not be seriously conceived as being able to offer such applications. Judicial uses of neuroimaging through expertise are a reality nonetheless, in Canadian courts as in French law. This thesis emphasizes that the conceptions of imaging technologies integrated in the two legal systems studied are incomplete, which creates an important amount of risks. It discusses the conditions for the use of an extra-legal normativity, the international technical standardization, which could be elaborated in this particular and controversial context, and outlines several features of an increased dialogue between legal and technological norms

The traditional diagnosis of patients with disorders of consciousness relies solely on behavioral responses. In 1996 it was estimated that 43% of patients diagnosed with unresponsive wakefulness syndrome (vegetative state) receive the wrong diagnosis. Assessing consciousness is perhaps the most crucial part of the diagnostic process. The challenging task of identifying covert consciousness in this patient group seems to be the biggest issue. In 2006 willful modulation of brain activity in response to a mental imagery task was discovered in a patient with unresponsive wakefulness syndrome. The brain activity was measured with functional magnetic resonance imaging. It was concluded that consciousness was preserved in this patient and new research investigating this novel method began to take place. The aim of this thesis was to conduct a systematic review of the literature and thereby arrive at the best current estimate of the proportion of patients who receive a diagnosis that wrongfully defines them as “unconscious” although they in fact are “covertly conscious”. In this review, 11 studies were examined. The results showed that patients with unresponsive wakefulness syndrome, who still receive the wrong diagnosis, decreased to 22-28% by the use of neuroimaging. This improvement points to the possible use of neuroimaging methods in the diagnosis of disorders of consciousness. However, this result cannot be taken without reservations. The limitations of the studies have to be taken into consideration. For example, most studies included a limited sample size and healthy controls did not always give the expected response to mental imagery tasks.

Adolescence is an exceptional period of life, not least in terms of social and brain development. Friends become increasingly important, susceptibility to peer rejection increases, and brain regions involved in social cognition are predicted to go through major changes. Adolescents’ social lives today may, to different extents, take place on social-media platforms online. This systematic review investigates how social-media use (SMU) affects adolescents’ brains. Out of 626 studies from the initial search, seven met the inclusion criteria. Out of these, five studies used functional magnetic resonance imaging, one study used diffusion tensor imaging, and one study used diffusion-weighted imaging. Functional findings suggest the reward circuit of the brain, as well as brain regions implicated in social cognition, to be involved in SMU. Activity in the nucleus accumbens was elicited by both giving and getting likes on posted pictures, whereas more SMU related to increased activity in the medial prefrontal cortex during physical self-judgement. Structural findings indicate frequent SMU to be associated with more reward sensitivity in terms of increased white matter in reward-processing pathways. These studies provide an initial understanding of the neural mechanisms of adolescents’ SMU. Future research is needed to draw inferences about how SMU affects the brains of adolescents.

Indiana University-Purdue University Indianapolis (IUPUI) / Over 5 million older adults have Alzheimer's disease (AD) in the US, and this number is projected to double by 2050. Clinical trials of potential pharmacological treatments for AD have largely shown that once cognitive decline has occurred, targeting AD pathology in the brain does not improve cognition. Therefore, it is likely that the most effective treatments for AD will need to be administered before cognitive symptoms occur, necessitating a biomarker for the early, preclinical stages of AD. Cerebral blood flow (CBF) is a promising early biomarker for AD. CBF is decreased in individuals with AD compared to their normally aging counterparts, and it has been shown that CBF is altered in mild cognitive impairment (MCI) and earlier stages and may occur prior to amyloid or tau aggregation. In addition, CBF can be measured using arterial spin labeled (ASL) MRI, a noninvasive imaging technique that can be safely repeated over time to track prognosis or treatment efficacy. The complex temporal and spatial patterns of altered CBF over the course of AD, as well as the relationships between CBF and AD-specific and -nonspecific factors, will be critical to elucidate in order for CBF to be an effective early biomarker of AD. Here, we begin to characterize the relationships between CBF and risk factors, pathologies, and symptoms of AD. Chapter 1 is a systematic review of published literature that compares CBF in individuals with AD and MCI to CBF in cognitively normal (CN) controls and assesses the relationship between CBF and cognitive function. Chapter 2 reports our original research assessing the relationships between CBF, hypertension, and race/ethnicity in older adults without dementia from the the Indiana Alzheimer’s Disease Research Center (IADRC) and Alzheimer’s Disease Neuroimaging Initiative (ADNI). Chapter 3 reports our original research assessing the relationships between CBF and amyloid beta and tau aggregation measured with PET, as well as whether hypertension or APOEε4 positivity affects these relationships, in older adults without dementia from the IADRC. Chapter 4 reports our original research assessing the relationship between the spatial distribution of tau and subjective memory concerns. / 2023-05-24

The neuronal substrates underlying aspects of comorbidity in brain disease states may be described over psychiatric and neurologic conditions that include affective disorders, cognitive disorders, schizophrenia, obsessive-compulsive disorder, substance abuse disorders as well as the neurodegenerative disorders. Regional and circuitry analyses of biogenic amine systems that are implicated in neural and behavioural pathologies are elucidated using neuroimaging, electrophysiological, neurochemical, neuropharmacological and neurobehavioural methods that present demonstrations of the neuropathological phenomena, such as behavioural sensitisation, cognitive impairments, maladaptive reactions to environmental stress and serious motor deficits. Considerations of neuronal alterations that may or may not be associated with behavioural abnormalities examine differentially the implications of discrete areas within brains that have been assigned functional significance; in the case of the frontal lobes, differential deficits of ventromedial and dorsolateral prefrontal cortex may be associated with different aspects of cognition, affect, remission or response to medication thereby imparting a varying aspect to any investigation of comorbidity.

Resting-state functional magnetic resonance imaging (rs-fMRI) is a non-invasive neuroimaging technique that utilizes spontaneous low-frequency fluctuations of blood-oxygenation-level dependent (BOLD) signals to examine resting-state functional connectivity in the brain. In the past two decades, this technique has been increasingly utilized to investigate properties of large-scale functional neural networks as well as their alterations in various cognitive and disease states. However, much less is known about large-scale functional neural networks of the rodent brain, particularly in the awake state. Therefore, we attempted to unveil local and global functional connectivity in awake rat through a combination of seed-based analysis, independent component analysis and graph-theory analysis. In the current studies, we revealed elementary local networks and their global organization in the awake rat brain. We further systematically compared the functional neural networks in awake and anesthetized states, revealing that the rat brain was locally reorganized while maintaining global topological properties from awake to anesthetized states. Furthermore, specific neural circuitries of the rat brain were examined using resting-state fMRI. First anticorrelated functional connectivity between infralimbic cortex and amygdala were found to be evident with different preprocessing methods (global signal regression, regression of ventricular and white matter signal and no signal regression). Secondly the thalamocortical connectivity was mapped for individual thalamic groups, revealing group-specific functional cortical connections that were generally consistent with known anatomical connections in rat. In conclusion, large-scale neural networks can be robustly and reliably studied using rs-fMRI in awake rat, and with this technique we established a baseline of local and global neural networks in the awake rat brain as well as their alterations in the anesthetized condition.

The dissertation discusses various node identification techniques as well as their downstream effects on network characteristics using task-activated fMRI data from two working memory paradigms: a verbal n-back task and a visual n-back task. The three node identification techniques examined within this work include: a group-aggregated approach, a subject-specific approach, and a voxel wise approach. The first chapters highlight crucial differences between group-aggregated and subject-specific methods of isolating nodes prior to undirected functional connectivity analysis. Results show that the two techniques yield significantly different network interactions and local network characteristics, despite having their network nodes restricted to the same anatomical regions. Prior to the introduction of the third technique, a chapter is dedicated to explaining the differences between a priori approaches (like the previously introduced group-aggregated and subject-specific techniques) and no a priori approaches (like the voxel wise approach). The chapter also discusses two ways to aggregate signal for node representation within a network: using the signal from a single voxel or aggregating signal across a group of neighboring voxels. Subsequently, a chapter is dedicated to introducing a novel processing pipeline which uses a data driven voxel wise approach to identify network nodes. The novel pipeline defines nodes using spatial temporal features generated by a deep learning algorithm and is validated by an analysis showing that the isolated nodes are condition and subject specific. The dissertation concludes by summarizing the main takeaways from each of the three analyses as well as highlighting the advantages and disadvantages of each of the three node identification techniques. / Includes bibliography. / Dissertation (Ph.D.)--Florida Atlantic University, 2020. / FAU Electronic Theses and Dissertations Collection

In this research two functional magnetic resonance tomography experiments were conducted to identify core regions of language comprehension processes. The focus of the studies was on inferencing, i.e. the activation of information which has not been explicitly mentioned in a given utterance but which is somehow implied because of general world knowledge. The research strategy was two-fold. First, text materials were used which allowed to isolate inference processes from more basic language processes. Second, two tasks verification in Experiment 1 and recognition in Experiment 2 were assigned to the participants to selectively enhance or attenuate processing at different levels of representation. In both experiments a network of brain areas was found to be active during language comprehension including areas all along the left superior temporal sulcus, the left lateral and medial prefrontal areas, as well as the right anterior temporal lobe and the posterior cingulate cortex. The results of Experiment 1 indicated that the dorsomedial prefrontal cortex was most prominently associated with inferencing in the context of the verification task. As expected, activity in this region was attenuated in Experiment 2 during recognition. No indications were found that the right hemisphere plays a particular role for inferencing as has been suggested by some authors. The results of both experiments are discussed with respect to the neuroimaging literature on language comprehension and with respect to recent approaches to memory systems in the brain particularly the episodic memory system. Finally, a functional neuroanatomical model of inferencing is sketched.

Cette thèse présente de nouvelles méthodes d’apprentissage structuré et parcimonieux sur les graphes, ce qui permet de résoudre une large variété de problèmes d’imagerie cérébrale, ainsi que d’autres problèmes en haute dimension avec peu d’échantillon. La première partie de cette thèse propose des relaxation convexe de pénalité discrète et combinatoriale impliquant de la parcimonie et bounded total variation d’un graphe, ainsi que la bounded `2. Ceux-ci sont dévelopé dansle but d’apprendre un modèle linéaire interprétable et on démontre son efficacacité sur des données d’imageries cérébrales ainsi que sur les problèmes de reconstructions parcimonieux.Les sections successives de cette thèse traite de la découverte de structure sur des modèles graphiques “undirected” construit à partir de peu de données. En particulier, on se concentre sur des hypothèses de parcimonie et autres hypothèses de structures dans les modèles graphiques gaussiens. Deux contributions s’en dégagent. On construit une approche pour identifier les différentes entre des modèles graphiques gaussiens (GGMs) qui partagent la même structure sous-jacente. On dérive la distribution de différences de paramètres sous une pénalité jointe quand la différence des paramètres est parcimonieuse. On montre ensuite comment cette approche peut être utilisée pour obtenir des intervalles de confiances sur les différences prises par le GGM sur les arêtes. De là, on introduit un nouvel algorithme d’apprentissage lié au problème de découverte de structure sur les modèles graphiques non dirigées des échantillons observés. On démontre que les réseaux de neurones peuvent être utilisés pour apprendre des estimateurs efficacaces de ce problèmes. On montre empiriquement que ces méthodes sont une alternatives flexible et performantes par rapport aux techniques existantes. / This dissertation presents novel structured sparse learning methods on graphs that address commonly found problems in the analysis of neuroimaging data as well as other high dimensional data with few samples. The first part of the thesis proposes convex relaxations of discrete and combinatorial penalties involving sparsity and bounded total variation on a graph as well as bounded `2 norm. These are developed with the aim of learning an interpretable predictive linear model and we demonstrate their effectiveness on neuroimaging data as well as a sparse image recovery problem.The subsequent parts of the thesis considers structure discovery of undirected graphical models from few observational data. In particular we focus on invoking sparsity and other structured assumptions in Gaussian Graphical Models (GGMs). To this end we make two contributions. We show an approach to identify differences in Gaussian Graphical Models (GGMs) known to have similar structure. We derive the distribution of parameter differences under a joint penalty when parameters are known to be sparse in the difference. We then show how this approach can be used to obtain confidence intervals on edge differences in GGMs. We then introduce a novel learning based approach to the problem structure discovery of undirected graphical models from observational data. We demonstrate how neural networks can be used to learn effective estimators for this problem. This is empirically shown to be flexible and efficient alternatives to existing techniques.