New optical technologies were developed to continuously measure cerebral blood flow (CBF) and oxygen metabolism (CMRO2) non-invasively through the skull. Methods and devices were created to improve the performance of near-infrared spectroscopy (NIRS) and diffuse correlation spectroscopy (DCS) for use in experimental animals and humans. These were employed to investigate cerebral metabolism and cerebrovascular reactivity under different states of anesthesia and during models of pathological states. Burst suppression is a brain state arising naturally in pathological conditions or under deep general anesthesia, but its mechanism and consequences are not well understood. Electroencephalography (EEG) and cortical hemodynamics were simultaneously measured in rats to evaluate the coupling between cerebral oxygen metabolism and neuronal activity in the burst suppressed state. EEG bursts were used to deconvolve NIRS and DCS signals into the hemodynamic and metabolic response function for an individual burst. This response was found to be similar to the stereotypical functional hyperemia evoked by normal brain activation. Thus, spontaneous burst activity does not cause metabolic or hemodynamic dysfunction in the cortex. Furthermore, cortical metabolic activity was not associated with the initiation or termination of a burst. A novel technique, time-domain DCS (TD-DCS), was introduced to significantly increase the sensitivity of transcranial CBF measurements to the brain. A new time-correlated single photon counting (TCSPC) instrument with a custom high coherence pulsed laser source was engineered for the first-ever simultaneous measurement of photon time of flight and DCS autocorrelation decays. In this new approach, photon time tags are exploited to determine path-length-dependent autocorrelation functions. By correlating photons according to time of flight, CBF is distinguished from superficial blood flow. Experiments in phantoms and animals demonstrate TD-DCS has significantly greater sensitivity to the brain than existing transcranial techniques. Intracranial pressure (ICP) modulates both steady-state and pulsatile CBF, making CBF a potential marker for ICP. In particular, the critical closing pressure (CrCP) has been proposed as a surrogate measure of ICP. A new DCS device was developed to measure pulsatile CBF non-invasively. A novel method for estimating CrCP and ICP from DCS measurement of pulsatile microvascular blood flow in the cerebral cortex was demonstrated in rats. / 2018-03-08T00:00:00Z

Functional magnetic resonance imaging (fMRI) based on blood oxygenation level dependent (BOLD) contrast has become a universal methodology in functional neuroimaging. However, the BOLD signal consists of a mix of physiological parameters and has relatively poor reproducibility. As fMRI becomes a prominent research tool for rehabilitation studies involving repeated measures of the human brain, more quantitative and stable fMRI contrasts are needed. This dissertation enhances quantitative measures to complement BOLD fMRI. These additional markers, cerebral blood flow (CBF) and cerebral blood volume (CBV) (and hence cerebral metabolic rate of oxygen (CMRO₂) modeling) are more specific imaging markers of neuronal activity than BOLD. The first aim of this dissertation assesses feasibility of complementing BOLD with quantitative fMRI measures in subjects with central visual impairment. Second, image acquisition and analysis are developed to enhance quantitative fMRI by quantifying CBV while simultaneously acquiring CBF and BOLD images. This aim seeks to relax assumptions related to existing methods that are not suitable for patient populations. Finally, CBF acquisition using a low-cost local labeling coil, which improves image quality, is combined with simultaneous acquisition of two types of traditional BOLD contrast. The demonstrated enhancement of CBF, CBV and CMRO₂measures can lead to better characterization of pathophysiology and treatment effects.

La spectroscopie par résonance magnétique nucléaire (RMN) est un outil puissant permettant d’acquérir des profils biochimiques du cerveau et de quantifier de nombreux paramètres cellulaires in vivo. Au cours de ce travail de thèse, nous nous sommes intéressés à trois techniques : (i) la spectroscopie RMN du 1H pondérée en diffusion, (ii) la spectroscopie RMN du carbone-13 (13C) et (iii) de l’oxygène-17 (17O) pour étudier la microstructure et la fonction cellulaire in vivo.Les métabolites cérébraux sont des traceurs endogènes spécifiques d’un type cellulaire (neurones et astrocytes) dont la diffusion dépend des nombreuses propriétés cellulaires (par exemple la viscosité du cytosol et la restriction intracellulaire). L’étude de la dépendance du coefficient de diffusion (ADC) aux temps de diffusion (td) permet de quantifier chacun de ces paramètres. En particulier, la mesure de l’ADC aux td longs permet d’évaluer la compartimentation des métabolites. Dans une première étude, nous avons mesuré l’ADC de plusieurs métabolites neuronaux et astrocytaires sur une large gamme de td (de ~80 ms à ~1 s) dans un large voxel dans le cerveau du macaque. Aucune dépendance de l’ADC de l’ensemble des métabolites au td n’a été observée suggérant que les métabolites diffusent majoritairement dans les prolongements neuronaux (axones, dendrites) et astrocytaires et ne sont pas confinés dans le corps cellulaire ou les organelles (mitochondries, noyau). La grande taille du voxel, liée à la sensibilité de détection limitée, ne nous a pas permis d’étudier la compartimentation des métabolites dans la substance blanche (SB) et la substance grise (SG). C’est pourquoi, une nouvelle étude a été réalisée dans le cerveau de l’Homme. Les résultats montrent que les métabolites diffusent dans des structures fibrillaires dans la SG et la SB. Enfin, une dernière étude, avec une gamme de td jusqu’à 2 s chez le macaque, nous a permis d’estimer, à l’aide de modèles analytiques simples mimant la structure cellulaire, la longueur des fibres neuronales (~110 μm) et astrocytaires (~70 μm). L’oxydation du glucose au sein des mitochondries permet de produire l’ATP (adénosine triphosphate), la principale source d’énergie de l’organisme. La spectroscopie du 13C permet de mesurer la vitesse de dégradation du glucose dans le cycle de Krebs (VTCA). Cette méthode est largement reconnue pour l’étude du métabolisme. Néanmoins, de nombreuses limitations, en termes de modélisation des données en détection indirecte ou de puissance émise dans le contexte du découplage hétéronucléaire en détection directe, ont été rencontrées sur notre scanner IRM. C’est pourquoi, la spectroscopie du 17O a ensuite été développée afin de quantifier la vitesse de consommation de l’oxygène pendant la phosphorylation oxydative (CMRO2). Des développements méthodologiques et technologiques ont été nécessaires et sont encore en cours pour mettre en place et valider cette technique qui n’a encore jamais été utilisée chez le macaque. / Magnetic Resonance Spectroscopy is a unique tool that allows acquiring brain biochemical profiles and quantifying many cellular parameters in vivo. During this thesis, three different techniques have been developed: (i) 1H diffusion-weighted, (ii) carbone-13 (13C) and (iii) oxygen-17 (17O) NMR spectroscopy to study brain structure and function in vivo. Brain metabolites are cell-specific endogeneous tracers of the intracellular space whose translational diffusion depends on many cellular properties (e.g.: cytosol vicosity and intracellular restriction). Studying the variation of the diffusion coefficient (ADC) as a function of diffusion time (td) allows untangling and quantifying those parameters. In particular, measuring metabolites ADC at long diffusion times gives information about the metabolites compartmentation in cells. In a first study, we measured neuronal and astrocytic metabolites ADC over a large time window (from ~80 ms to ~1 s) in a large voxel in the macaque brain. No dependence of all metabolites ADC on td was observed suggesting that metabolites primarily diffuse in neuronal (dendrites and axons) and astrocytic processes and are not confined inside the cell body and organelles (nucleus, mitochondria). The large size of the voxel, due to low detection sensitivity, did not allow us to study metabolites compartmentation in pure white (WM) and grey matters (GM). Therefore, we performed a new study in the human brain. Results showed that in both WM and GM metabolites diffuse in fiber-like cell structure. Finally, using an even larger time window (up to 2 s) in the macaque brain and analytical models mimicking the cell structure, we estimated the length of neuronal (~110 μm) and astrocytic (~70 μm) processes. ATP (adenosine triphosphate), the main source of energy in the organism, is produced thanks to glucose oxidation inside the mitochondria. 13C NMR spectroscopy is a well-known technique to study brain energy metabolism and can be used to estimate the rate of glucose degradation within the Krebs cycle (VTCA). However, many limitations, concerning data modeling when performing indirect detection or power deposition due to heteronuclear decoupling during direct detection, were encountered on our MRI scanner. Therefore, 17O NMR spectroscopy was developed to quantify the rate of oxygen consumption during oxidative phosphorylation (CMRO2). Methodological and technological developments were necessary and are still ongoing to validate this technique, which has never been used with macaque.