Blood oxygenation level dependent (BOLD) functional magnetic resonance imaging (fMRI) is the dominant tool used for mapping human brain function because it is non-invasive, does not use ionizing radiation, and offers relatively high spatial and temporal resolution compared to other neuroimaging techniques. Unfortunately, conventional fMRI techniques cannot map brain function in the inferior temporal cortex (ITC) and orbitofrontal cortex (OFC). These brain regions experience severe magnetic field distortions due to magnetic susceptibility mismatch with the neighboring air-filled ear-canals (ITC) or sinus cavities (OFC), causing loss of the fMRI signal. Functional imaging capability is important for gaining a better understanding of these brain regions and the diseases that commonly affect them (Alzheimer’s disease and epilepsy (ITC), Parkinson’s disease and schizophrenia (OFC)). Balanced steady state free precession (balanced SSFP) is a relatively new fMRI technique that can measure function in all brain regions. Rather than diffuse signal loss, balanced SSFP images exhibit signal loss in spatially periodic, narrow bands. Banding artifacts cannot be eliminated in a single scan, but the phase of the banding artifacts can be controlled by the experimenter, permitting the combination of two antiphase balanced SSFP images to produce a single image free of banding artifacts. Unfortunately, image-corrupting transient signal oscillations limit the rate at which the banding artifact phase can be modified, such that the banding-artifact-free image acquisition rate is prohibitively slow for most clinical and neuroscience applications. This work describes the development of a modified balanced SSFP fMRI technique, alternating SSFP, which permits rapid, banding-artifact-free balanced SSFP fMRI. Theoretical modeling was used to find a rapid transition between antiphase balanced SSFP images with minimal transient signal oscillations. Monte Carlo simulations were used to optimize alternating SSFP acquisition parameters for BOLD sensitivity, with comparison to established balanced SSFP acquisitions. Rat fMRI was used to confirm these predictions. Finally, the ability of alternating SSFP to provide rapid, banding-artifact-free balanced SSFP fMRI in humans at 4 T was demonstrated.

Les ganglions de la base (GB) sont des structures cérébrales profondes participant à la sélection de comportements adaptés, avec ses composantes motrices, cognitives et émotionnelles. L’étude par IRMf BOLD de ces structures présente un grand intérêt pour explorer leur rôle et leur dysfonctionnement dans certaines pathologies, comme la maladie de Parkinson (MP). Cette technique permet, par l’étude du signal BOLD, de mettre en évidence des activations cérébrales suite à une activation neuronale. Or l’IRMf BOLD a été optimisée pour l’étude des activations corticales, et la mise en évidence d’activations dans les GB est difficile, surtout au niveau individuel. Ceci est en parti lié au fait que le signal BOLD est plus faible dans ces structures par rapport au cortex. Plusieurs raisons peuvent expliquer ce faible signal BOLD. Ainsi la charge en fer de ses structures, modifiant le paramètre de relaxation T 2 *, peut en être une des causes. En effet, la sensibilité de mesure du signal BOLD est optimale lorsque le temps d’écho (TE) de la séquence d’acquisition égale le T 2 * de la structure cérébrale d’intérêt. Notre premier travail a consisté à étudier l’hétérogénéité du T 2 * dans différentes structures cérébrales en tenant compte des effets de la MP, pathologie connue pour entrainer des accumulations de fer dans certaines régions. Nous avons par ailleurs étudié l’évolution du T 2 * de manière longitudinale, et ce paramètre est apparu comme un biomarqueur intéressant de l’évolutivité de la MP. Le deuxième travail a été consacré à étudier les activations des GB en tenant compte de l’hétérogénéité du T 2 *. Nous avons étudié les activations cérébrales suite à la réalisation d’une tâche motrice, en explorant entre autres l’effet TE. Nous avons montré que le choix du TE a finalement peu d’impact sur la capacité de détection des activations au niveau des GB. Nous proposons une stratégie pour l’étude individuelle de l’activité cérébrale au niveau des GB en utilisant le pourcentage de changement du signal BOLD dans des régions cérébrales d’intérêt préalablement définies sur l’analyse de groupe. / The basal ganglia (BG) are deep brain structures involved in the selection of appropriate behavior, with motor, cognitive and emotional components. The BOLD fMRI study of these structures is of great interest to explore their role and dysfunction in certain diseases, such as Parkinson's disease (PD). By studying the BOLD signal, this technique allows to identify brain activation following neuronal activation. However BOLD fMRI has been optimized for the study of cortical activations and detection of activations in the BG is difficult, mainly at the individual level. This is partly due to the fact that the BOLD signal is lower in these structures in relation to the cortex. Several reasons may explain the BOLD signal attenuation. Thereby, iron load in its structures, which changes the relaxation parameter T 2 *, may be a cause. Indeed, the BOLD signal is optimal when the echo time (TE) of the MRI acquisition sequence equal T 2 * of the considered brain structure. Our first work was to study the heterogeneity of T 2 * in different brain structures, taking into account the effects of PD. Indeed, PD is known to induce iron accumulation some regions. We also studied the evolution of T 2 * longitudinally, and this parameter has emerged as an interesting biomarker to track PD progression. The second work was to study the activation of BG taking into account T 2 * heterogeneity. We studied brain activation during a motor task, exploring in particular the effect of TE. We showed that the choice of TE has a low impact on BG activation detection sensitivity. We propose a strategy for individual quantification of neuronal activity in the BG, using the BOLD percentage signal change in pre-defined regions of interest, obtained from the group analysis.

This master’s thesis deals with problematic of correlates finding of heart rate changes in fMRI data. The first part describes principle of fMRI, creation of BOLD signal, data acquisition, their pre-processing and analysis. The next part describes heart rate variability and its impact on fMRI data. The following section is dedicated to pre-processing of heart rate time series to the form, which can be used in correlates finding of heart rate variability and fMRI data with generalized linear model. The process of statistical testing and its result with discussion can be found in the last part of this thesis.

Nos organes sensoriels sont sous un bombardement permanent d'informations provenant du monde extérieur dont seule une infime fraction est intéressante ou pertinente pour l'individu à chaque instant, comme par exemple le visage d'une personne recherchée au milieu d'une foule. Filtrer et extraire cette information sont des fonctions vitales du cerveau, mais les mécanismes neuronaux qui décident de la réussite ou l'échec de cette épreuve restent partiellement compris. La visée de ce projet est d'étudier la relation entre activité neurale de fond et performance comportementale dans des paradigmes nécessitant différents types d'attention, au moyen de techniques comportementales et de la neuroimagerie fonctionnelle. Le modèle théorique que j'évalue dans cette thèse postule une organisation hiérarchique de ces fonctions avec l'éveil comme fonction basique d'activation physiologique, la vigilance comme fonction intermédiaire et l'attention sélective comme niveau le plus élevé et spécifique. Nous proposons en outre que l'attention sélective et la vigilance emploient deux mécanismes orthogonaux ; ces deux versants du spectre attentionnel seront le focus des deux études expérimentales présentées dans ce manuscrit. Dans les deux cas, il s'agit de chercher un lien entre l'activité neuronale spontanée et la variabilité comportementale à travers des essais répétés.J'ai utilisé le paradigme de Stroop pour opérationnaliser l'attention sélective soumise à la distraction ; la couleur du mot écrit étant la cible et le mot lui-même un distracteur en conflit direct avec cette cible. Les résultats indiquent qu'une activité pré-stimulus plus élevée dans les aires frontales et ventrales pertinentes pertinentes pour la tâche corrèle essai par essai avec des temps de réponse plus rapides ; l’effet inverse est observé dans l’aire liée au traitement de la distraction (i.e. sensible à la forme des mots). De plus, la variabilité comportementale observée à travers les sujets corrèle avec la variabilité de l’effet préstimulus dans ces régions. Pour étudier la vigilance j’ai conçu un paradigme expérimental au cours duquel les sujets ont pour instruction de rapporter tout événement saillant dans la modalité auditive ou visuelle, sans aucun signal préparatoire permettant de prédire l’apparition du prochain stimulus ou sa nature. Confirmant nos résultats précédents, des temps de réactions plus rapides sont associés à une activité pré-stimulus plus élevée dans le réseau cingulo-thalamo-insulaire. De surcroît, une activité plus élevée dans le réseau du mode par défaut (DMN) est également associée à des réponses plus rapides alors que l’activité spontanée du réseau attentionnel dorsal (DAN) n’a aucun effet sur les temps de réaction aux essais visuels et un effet négatif sur ceux des essais auditifs. Des temps de réaction plus rapide sont également associés à une activité pré-stimulus plus élevée dans les aires sensorielles primaires, mais uniquement lorsque la modalité du stimulus subséquent est congruente avec l’aire de traitement sensoriel. Les conclusions générales qui découlent de cette thèse sont donc doubles : d’une part elle confirme la pertinence fonctionnelle des fluctuations neuronales spontanées pour le comportement, d’autre part elle apporte une première identification des structures neuronales impliquées dans la vigilance sans la confondre avec l’attention sélective. / How does the brain manage to efficiently select from the abundance of sensory input that information which is currently relevant, as, for example, a known face in a crowd? Filtering and extracting this information are essential functions of the brain, but the neural mechanisms underlying the success or failure of this cognitive operation are only partially understood. Together, the aim of this project is to study, using behavioral and imaging techniques, the relationship between spontaneous neural activity and behavioral performance in task settings involving several types of attention. The related working model I wish to assess in my thesis is that attentional functions are organized in a hierarchical manner, with arousal as a basic function, alertness as an intermediate function and selective attention as the highest and most specific level. I further propose that selective attention and tonic alertness employ two orthogonal mechanisms, those two sides of the attentional spectrum will be the focus of the two experiments presented in this manuscript. In both studies, our aim was to establish a link between spontaneous fluctuations of brain activity and behavioral variability across repeated trials I used the Stroop paradigm to operationalize selective attention under distraction, the target being the color of the ink in which the word is written, the word itself is then a distractor conflicting with the target. Results indicate that higher prestimulus activity in frontal and ventral regions relevant for the task correlates trial-by-trial with faster reaction times, the opposite effect is found in the area involved in the processing of distracting features (i.e. sensitive to the word form). Moreover, inter-subjects behavioral variability also correlated with the prestimulus effect variability in those regions. To study alertness, I designed an experimental paradigm where subjects are instructed to report whatever salient event occurs in the auditory or visual modality during the recording, without any preparatory cue allowing them to predict the timing or type of an upcoming stimulus. Confirming our previous results, faster reaction times were associated with a higher prestimulus activity in the cingulo-thalamo-insular network. Furthermore, higher prestimulus activity in the default mode network (DMN) was also tied to faster responses whereas dorsal attention network (DAN) activity was overall irrelevant and, on auditory trials, even detrimental to performance. Similarly, higher prestimulus activity in the primary sensory cortices was associated with faster responses, but those effects were confined to the respective modality or, for visual trials, most pronounced in the relevant retinotopic representation. The general conclusions resulting from this thesis are two-fold: first, it confirms the functional relevance of spontaneous neuronal fluctuations for behavior; on the other hand, it brings a first identification of the brain structures involved in alertness, without confusing it with selective attention.

Circadian Rhythms (CR) are driven by a biological clock called as suprachiasmaticnucleus (SCN), located in a brain region called the hypothalamus. These rhythms are very much necessary in maintaining the sleep and wake cycle at appropriate times in a day. As a starting step towards non-invasive investigation of CR, aim is to study changes in the physiological processes of two Regions of Interest (ROI), the hypothalamus and the visual cortex. This was studied using a functional Magnetic Resonance Imaging (fMRI) technique to investigate for any changes or differences in the Blood Oxygen Level Dependent (BOLD)signals extracted from the ROI during a visual stimulation. We acquired and processed fMRI data to extract BOLD signals from ROI and the extracted signals are again further used to study the correlation with the experimental ON-OFF design paradigm. The extracted BOLD signals varied a lot between the two specified brain regions within the same subject and between three types of fMRI data. These variations were found in terms of number of activated voxels and also Signal to Noise ratio(SNR) level present in the signals. The number of activated voxels and SNR werehigh in visual cortex whereas low number of activated voxels and low SNR were found in hypothalamus. The correlation between BOLD responses from primaryvisual cortex were shown as positive with the experimental stimulation whereas BOLD responses extracted from hypothalamus have shown a negative correlation in time with the experimental stimulation. As a start up of the project, these BOLD responses can provide references for a future use in research studies, especially to further study about change in phase of the BOLD signal extracted exactly from the SCN. These phase responses can then be used to study physiological processing in subjects affected by sleep disorders.

<p dir="ltr">The brain is a dynamic tissue that is passively driven by a combination of the cardiac cycle, respiration, and slow wave oscillations. The function of the brain relies on its ability to maintain a normal homeostatic balance between its mechanical environment and metabolic demands, which can be greatly altered in the cases of neurodegeneration or traumatic brain injury. It has been a challenge in the field to quantify the dynamics of the tissue and cerebrospinal fluid flow in human subjects on a patient-specific basis over the many spatial and temporal scales that it relies upon. Non-invasive imaging tools like structural, functional, and dynamic MRI sequences provide modern researchers with an unprecedented view into the human brain. Our work leverages these sequences by developing novel, open-source pipelines to 1) quantify the biomechanical environment of the brain tissue over 133 functional brain regions, and 2) estimate real-time cerebrospinal fluid velocity from flow artifacts on functional MRI by employing breathing regimens to enhance fluid motion. These pipelines provide a comprehensive view of the macroscale tissue and fluid motion in a given patient. Additionally, we sought to understand how the transmission of macroscale forces, in the context of traumatic brain injury, contribute to neuronal damage by 3) developing a digital twin to simulate 30-200 g-force loading of 2D neuronal cultures and observing the morphological and electrophysiological consequences of these impacts in vitro by our collaborators. Taken together, we believe these works are a steppingstone that will enable future researchers to deeply understand the mechanical contributions that underly clinical neurological outcomes and perhaps lead to the development of earlier diagnostics, which is of dire need in the case of neurodegenerative diseases.</p>

The aim of the thesis is to define the dependency of the classification of patients affected by neurodegenerative diseases on the choice of the parcellation atlas. Part of this thesis is the application of the functional connectivity analysis and the calculation of graph metrics according to the method published by Olaf Sporns and Mikail Rubinov [1] on fMRI data measured at CEITEC MU. The application is preceded by the theoretical research of parcellation atlases for brain segmentation from fMRI frames and the research of mathematical methods for classification as well as classifiers of neurodegenerative diseases. The first chapters of the thesis brings a theoretical basis of knowledge from the field of magnetic and functional magnetic resonance imaging. The physical principles of the method, the conditions and the course of acquisition of image data are defined. The third chapter summarizes the graph metrics used in the diploma thesis for analyzing and classifying graphs. The paper presents a brief overview of the brain segmentation methods, with the focuse on the atlas-based segmentation. After a theoretical research of functional connectivity methods and mathematical classification methods, the findings were used for segmentation, calculation of graph metrics and for classification of fMRI images obtained from 96 subjects into the one of two classes using Binary classifications by support vector machines and linear discriminatory analysis. The data classified in this study was measured on patiens with Parkinson’s disease (PD), Alzheimer’s disease (AD), Mild cognitive impairment (MCI), a combination of PD and MCI and subjects belonging to the control group of healthy individuals. For pre-processing and analysis, the MATLAB environment, the SPM12 toolbox and The Brain Connectivity Toolbox were used.

<p>Alzheimer’s disease (AD) is a degenerative disease with progressive deterioration of neural networks in the brain. Fractal dimension analysis (FD) of resting state blood oxygen level dependent (BOLD) signals acquired using functional magnetic resonance imaging (fMRI) allows us to quantify complex signalling in the brain and may offer a window into the network erosion. This novel approach can provide a sensitive tool to examine early stages of AD. As AD progresses, we expect to see a reduction in brain connectivity and signal complexity concurrent with biochemical changes (e.g. altered levels of N-acetyl aspartate (NAA), myoinositol (mI) and glutamate as measured using magnetic resonance spectroscopy, MRS), volumetric changes and abnormally high levels of brain iron.</p> <p>Over a series of 4 studies we examined the relationship of BOLD signal complexity and functional connectivity with documented MRI markers of pathology in AD (n=38) as compared to normal controls (NC) (n=16). AD subjects were in early stage of illness (mild to moderate impairment on the mini mental state exam, MMSE). We validated the temporal (short term (within minutes) and longer term (over a number of months)) consistency of FD measurement and choice of BOLD acquisition method (spiral vs. EPI), provided MRI sequence repeat time (TR) was kept constant. FD reduction (decrease in signal complexity) correlated with worsening pathological values on MRS (­NAA decrease and mI increase) and with a decrease in functional connectivity. This demonstrates that FD (signal complexity) reduces in proportion to AD severity. FD reduction is connected to functional connectivity measured through resting state network (RSN) analysis suggesting the reduction in FD relates to neuronal loss rather than altered vascularity. The narrow range of cognitive impairment (such as scores on the MMSE or the clinical dementia rating scale, CDR) likely precluded correlation between these measures and FD or RSN. Functional connectivity (RSN) was also reduced when brain iron levels were increased within certain network nodes (posterior cingulate cortex and lateral parietal cortex). Therefore iron deposition may play a role in network disruption of AD brains.</p> <p>The overall conclusion of this thesis is that signal complexity of BOLD fMRI signals, as measured with FD, may detect early pathology in the progression of AD. FD can detect neuronal changes in deep brain structures before volume loss in these structures and before significant changes in MRS markers were detectable between the AD and NC groups. An FD change mirrors disruptions in functional connectivity but detection is not limited to RSN nodes in the brain. This novel approach could further our understanding of AD and may be applied to other pathologies of the brain.</p> / Doctor of Philosophy (PhD)