Dissertations / Theses on the topic 'Brain imaging'

Thesis (Ph.D.)--Harvard--Massachusetts Institute of Technology Division of Health Sciences and Technology, 2000.
Includes bibliographical references.
Understanding how the human brain works, in both health and disease, requires data with both high spatial and temporal resolution. This thesis develops and applies a spatiotemporal neuroimaging method. I describe a linear estimation inverse approach, which is a method for the combination of functional magnetic resonance imaging (fMRI) with electroencephalography (EEG) and magnetoencephalography (MEG). fMRI provides millimeter spatial resolution, while EEG and MEG provide millisecond temporal resolution. The thesis is divided into two broad sections: Monte Carlo modeling studies and experimental studies. Improvements to both the bioelectromagnetic forward and inverse solutions are demonstrated. Through modeling studies, I characterize the accuracy of the method with and without functional and anatomic constraints, the effects of various model mis-specifications, and as a function of EEG/MEG sensor configuration. I describe a noise sensitivity normalization to the traditional linear estimation operator that improves the point spread function (a measure of spatial resolution), increases the spatial homogeneity of the point spread, and allows interpretation of the localization in terms of a statistical measure (F-statistic). Using experimentally generated current dipoles implanted an epilepsy patient, I examine the accuracy of both a realistic and spherical EEG head model. This experimental data demonstrates the improved accuracy of the realistic head model, and gives us confidence in using this realistic head model for EEG source localization. The optimized and validated forward and inverse methods are then applied to a variety of empirical measurements. First, the combined multi modality imaging approach is used to simultaneous EEG/fMRI measurements of a visual stimulus, demonstrating the feasibility of measuring and localizing simultaneously acquired electric potential and hemodynamic measurements. Second, combined MEG/fMRI measurements are used to analyze the spatiotemporal characteristics of a cortical network that is responsive to visual motion coherency. Finally, in epilepsy patients, I compare the non-invasive MEG localization of interictal spikes with verification from invasive recordings and surgical results. These studies, in both normal volunteers and patients, clearly demonstrate the utility, accuracy, and power of the combined use of fMRI, EEG and MEG. The tools demonstrated here provide "real time movies" of the human brain at work during a given task or behavior. This information is required to develop computational models of how the human brain/mind works.
by ARthur K. Lui.
Ph.D.

In questo lavoro si sono analizzate due diverse tecniche di imaging di perfusione implementate in Risonanza Magnetica e Tomografia Assiale Computerizzata (TAC). La prima analisi proposta riguarda la tecnica di Arterial Spin Labeling che permette di ottenere informazioni di perfusione senza la somministrazione di un mezzo di contrasto. In questo lavoro si è sviluppata e testata una pipeline completa, attraverso lo sviluppo sia di un protocollo di acquisizione che di post-processing. In particolare, sono stati definiti parametri di acquisizione standard, che permettono di ottenere una buona qualità dei dati, successivamente elaborati attraverso un protocollo di post processing che, a partire dall'acquisizione di un esperimento di ASL, permette il calcolo di una mappa quantitativa di cerebral blood flow (CBF). Nel corso del lavoro, si è notata una asimmetria nella valutazione della perfusione, non giustificata dai dati e probabilmente dovuta ad una configurazione hardware non ottimale. Risolta questa difficoltà tecnica, la pipeline sviluppata sarà utilizzata come standard per l’acquisizione e il post-processing di dati ASL. La seconda analisi riguarda dati acquisiti attraverso esperimenti di perfusione TAC. Si è presa in considerazione la sua applicazione a casi di infarti cerebrali in cui le tecniche di trombectomia sono risultate inefficaci. L'obiettivo di questo lavoro è stata la definizione di una pipeline che permetta il calcolo autonomo delle mappe di perfusione e la standardizzazione della trattazione dei dati. In particolare, la pipeline permette l’analisi di dati di perfusione attraverso l’utilizzo di soli software open-source, contrapponendosi alla metodologia operativa comunemente utilizzata in clinica e rendendo le analisi riproducibili. Il lavoro proposto è inserito in un progetto più ampio, che include future analisi longitudinali con coorti di pazienti più ampie per definire e validare parametri predittivi degli outcome dei pazienti.

Chapter 1 reviews literature, including a systematic and quantitative review of the volumetric MRI studies to date, and examines the links between such biological findings and the clinical features of the illness. Chapter 2 describes an original piece of research designed to examine the biological associations of treatment response in schizophrenia with MRI, SPET and detailed neuropsychological testing. Forty patients were selected as treatment responsive or resistant using standardised criteria. A quantitative analysis of particular regional volumes on MRI revealed that the treatment resistant group had a consistent, but not statistically significant, tendency to smaller cerebral structures. Qualitative ratings showed a tendency to greater atrophy in the treatment resistant patients. The forty SPET scans were subsequently used in a comparison of rCBF in medicated and unmedicated schizophrenia, other psychotic patients and normal controls. The schizophrenic patients showed the predicted hypofrontality, but this was limited to the anterior cingulate and medial pre-frontal cortex. Given some evidence, from post-mortem studies, of a preferential loss of gaba-ergic neurones in the anterior cingulate in schizophrenia, with a compensatory upregulation of non-specific GABA-A receptor binding, a study of GABA receptor binding on SPET was conducted using the benzodiazepine ligand Iomazenil. As described in Chapter 3, the expectation was that Iomazenil binding would be increased in frontal regions, but this was not confirmed in a comparison of ten schizophrenics and ten normal controls; and an apparent reduction of subcortical receptor binding was attributed to methodological problems. Finally, Chapter 4 describes the likely technical and experimental developments in brain imaging studies of schizophrenia in the foreseeable future. Some recommendations are made, based on these advances and the studies described in the thesis, that would help to exploit the full potential of neuroimaging to improve understanding of the pathophysiology of schizophrenia.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2011.
Cataloged from PDF version of thesis.
Includes bibliographical references.
Magnetic resonance imaging (MRI) has demonstrated the potential for non-invasive mapping of structure and function (fMRI) in the human brain. In this thesis, we propose a series of methodological developments towards improved fMRI of auditory processes. First, the inefficiency of standard fMRI that acquires brain volumes one slice at a time is addressed. The proposed single-shot method is capable, for the first time, of imaging the entire brain in a single-acquisition while still maintaining adequate spatial resolution for fMRI. This method dramatically increases the temporal resolution of fMRI (20 fold) and improves sampling efficiency as well as the ability to discriminate against detrimental physiological noise. To accomplish this it exploits highly accelerated parallel imaging techniques and MRI signal detection with a large number of coil elements. We then address a major problem in the application of fMVIRI to auditory studies. In standard fMRI, loud acoustic noise is generated by the rapid switching of the gradient magnetic fields required for image encoding, which interferes with auditory stimuli and enforces inefficient and slow sampling strategies. We demonstrate a fMRI method that uses parallel imaging and redesigned gradient waveforms to both minimize and slow down the gradient switching to substantially reduce acoustic noise while still enabling rapid acquisitions for fMRI. Conventional fMRI is based on a hemodynamic response that is secondary to the underlying neuronal activation. In the final contribution of this thesis, a novel image contrast is introduced that is aimed at the direct observation of neuronal magnetic fields associated with functional activation. Early feasibility studies indicate that the imaging is sensitive to oscillating magnetic fields at amplitudes similar to those observed by magnetoencephalography.
by Thomas Witzel.
Ph.D.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, February 2004.
Includes bibliographical references.
This thesis integrates hardware development, data analysis, and mathematical modeling to facilitate our understanding of brain cognition. Exploration of these brain mechanisms requires both structural and functional knowledge to (i) reconstruct the spatial distribution of the activity, (ii) to estimate when these areas are activated and what is the temporal sequence of activations, and (iii)to determine how the information flows in the large-scale neural network during the execution of cognitive and/or behavioral tasks. Advanced noninvasive medical imaging modalities are able to locate brain activities at high spatial and temporal resolutions. Quantitative modeling of these data is needed to understand how large-scale distributed neuronal interactions underlying perceptual, cognitive, and behavioral functions emerge and change over time. This thesis explores hardware enhancement and novel analytical approaches to improve the spatiotemporal resolution of single (MRI) or combined (MRI/fMRI and MEG/EEG) imaging modalities. In addition, mathematical approaches for identifying large-scale neural networks and their correlation to behavioral measurements are investigated. Part I of the thesis investigates parallel MRI. New hardware and image reconstruction techniques are introduced to improve spatiotemporal resolution and to reduce image distortion in structural and functional MRI. Part II discusses the localization of MEG/EEG signals on the cortical surface using anatomical information from AMTRI, and takes advantage of the high temporal resolution of MEG/EEG measurements to study cortical oscillations in the human auditory system. Part III introduces a multivariate modeling technique to identify "nodes" and "connectivity" in a
(cont.) large-scale neural network and its correlation to behavior measurements in the human motor system.
by Fa-Hsuan Lin.
Ph.D.

This chapter presents a brief introduction to the application of diffusion-weighted magnetic resonance imaging (MRI) to in vivo studies. Diffusion-weighted MRI has found application both in the clinic, and in basic neuroscience. In the former situation it is primarily used for the detection of brain lesions, in particular infarcted regions. The ability to follow fibre tracts in white matter via diffusion tensor imaging has also made this methodology of interest to the neurosurgeon wishing to avoid severance of essential fibre tracts, but also of interest to the cognitive neuroscientist exploring anatomical connectivity in the brain. The chapter starts with a brief recap of the theory of diffusionweighted MRI and moves on to examine the two major experimental confounds, eddy currents and bulk motion. Current correction schemes for these problems are touched upon. Diffusion anisotropy is introduced as a potential source of artefacts for lesion detection in white matter, and the diffusion tensor model presented. The chapter concludes with a short introduction to fibre tracking.

Autism is a life-long neurodevelopmental condition. Autistic individuals have difficulties in communicative and social ability, and repetitive and stereotypic behavior. It has proposed that these symptoms are caused by underconnectivity in the autistic brain. Functional imaging studies have reported functional underconnectivity in autism. In this thesis, the structural connectivity of the autistic brain was studied. White matter contains axon fibers, which connect different cortical and subcortical brain regions. To measure the structural connectivity, Diffusion tensor imaging (DTI) was applied. Since water diffusion in axons inside the white matter is directional, by measuring the magnitude and direction of water diffusion in white matter, the structural integrity of white matter fibers could be estimated. In this thesis, the background of autism as a genetic, neurological and behavioral condition is outlined. The methods needed to acquire and analyze DTI data are illustrated. A meta-analysis on abnormalities found in autistic brain using DTI was conducted and the most consistently reported regions with DTI differences in autism compared to typically developing controls are described. The results of the metaanalysis were localized to white matter tracts likely to be involved, and the possible associations between anatomy and autistic behavioral features are discussed. Finally, a DTI tractography study was conducted in a sample but clinically representative sample of patients with ASD and eighteen major white matter tracts were explored. Underconnectivity in several tracts was identified. It is hoped that the findings reported here will enhance our understanding of widespread underconnectivity in autism.
published_or_final_version
Psychiatry
Master
Master of Philosophy

Chronic musculoskeletal pain is a leading cause of disability worldwide yet the mechanisms of chronification and neural responses to effective treatment remain elusive. Non-invasive imaging techniques are useful for investigating brain alterations associated with health and disease. Thus the overall goal of this dissertation was to investigate the white (WM) and grey matter (GM) structural differences in patients with musculoskeletal pain before and after psychotherapeutic intervention: cognitive behavioral therapy (CBT). To aid in the interpretation of clinical findings, we used a novel porcine model of low back pain-like pathophysiology and developed a post-mortem, in situ, neuroimaging approach to facilitate translational investigation. The first objective of this dissertation (Chapter 2) was to identify structural brain alterations in chronic pain patients compared to healthy controls. To achieve this, we examined GM volume and diffusivity as well as WM metrics of complexity, density, and connectivity. Consistent with the literature, we observed robust differences in GM volume across a number of brain regions in chronic pain patients, however, findings of increased GM volume in several regions are in contrast to previous reports. We also identified WM changes, with pain patients exhibiting reduced WM density in tracts that project to descending pain modulatory regions as well as increased connectivity to default mode network structures, and bidirectional alterations in complexity. These findings may reflect network level dysfunction in patients with chronic pain. The second aim (Chapter 3) was to investigate reversibility or neuroplasticity of structural alterations in the chronic pain brain following CBT compared to an active control group. Longitudinal evaluation was carried out at baseline, following 11-week intervention, and a four-month follow-up. Similarly, we conducted structural brain assessments including GM morphometry and WM complexity and connectivity. We did not observe GM volumetric or WM connectivity changes, but we did discover differences in WM complexity after therapy and at follow-up visits. To facilitate mechanistic investigation of pain related brain changes, we used a novel porcine model of low back pain-like pathophysiology (Chapter 6). This model replicates hallmarks of chronic pain, such as soft tissue injury and movement alteration. We also developed a novel protocol to perform translational post-mortem, in situ, neuroimaging in our porcine model to reproduce WM and GM findings observed in humans, followed by a unique perfusion and immersion fixation protocol to enable histological assessment (Chapter 4). In conclusion, our clinical data suggest robust structural brain alterations in patients with chronic pain as compared to healthy individuals and in response to therapeutic intervention. However, the mechanism of these brain changes remains unknown. Therefore, we propose to use a porcine model of musculoskeletal pain with a novel neuroimaging protocol to promote mechanistic investigation and expand our interpretation of clinical findings.

Certes malalties neurològiques estan associades amb problemes en els sistemes de neurotransmissió. Una aproximació a l'estudi d'aquests sistemes és la tomografia d'emissió SPECT (Single Photon Emission Computed Tomography) com a tècnica
no-invasiva que proporciona imatges funcionals representatives de l'activitat neuronal. Aquesta tècnica permet la visualització i l'anàlisi de diferents òrgans i teixits dins l'àmbit de la Medicina Nuclear.

Malgrat que la inspecció visual de la imatge a vegades és suficient per establir el diagnòstic, la quantificació dels paràmetres de la imatge reconstruida poden millorar la fiabilitat i exactitud del diagnòstic precoç de la malaltia. En particular, la quantificació d'estudis de neurotransmissors de dopamina pot ajudar a detectar els estadis inicials de malalties com el Parkinson. Així mateix, la quantificació permet un seguiment més acurat de l'evolució de la malaltia i una evaluació dels efectes de la terapèutica aplicada.

La quantificació es veu afectada pels efectes degradants de la imatge com són el soroll estadístic, la resposta del sistema col.limador/detector i l'efecte de dispersió i/o atenuació dels fotons en la seva interacció amb la matèria. Alguns d'aquests efectes poden ser corregits mitjançant l'ús d'algoritmes de reconstrucció iteratius.

L'objectiu d'aquesta tesi és aconseguir una quantificació tant absoluta com relativa dels valors numèrics de la imatge reconstruida de manera que reprodueixin la distribució d'activitat real del pacient en el moment de l'adquisició de l'estudi de SPECT. Per aconseguir-ho s'han desenvolupat diferents codis i algoritmes per millorar els mètodes de reconstrucció existents i validar-ne els seus resultats.

La validació i millora dels algoritmes s'ha basat en l'ús de tècniques de simulació Monte Carlo. S'han analitzat els diferents codis Monte Carlo disponibles en l'àmbit de la Medicina Nuclear i s'ha escollit SimSET. La interpretació dels resultats obtinguts i la comparació amb els resultats experimentals ens van dur a incorporar modificacions en el codi original. D'aquesta manera vam obtenir i validar SimSET com a generador d'estudis de SPECT a partir de pacients i objectes virtuals.

La millora dels algoritmes es va basar en la incorporació de models analítics de la resposta del sistema col.limador/detector. La modelització del sistema es va implementar per diferents configuracions i energies de la font amb la utilització del codi Monte Carlo PENELOPE. Així mateix es va dissenyar un nou algoritme iteratiu que incorporés l'efecte 3D del sistema i es va tenir en compte la valoració de la imatge en tot el seu volum.

Finalment, es va proposar una correcció de l'scattering utilitzant el simulador SimSET modificat per tal d'accelerar el procés de reconstrucció. Els valors reconstruits de la imatge ens han permès recuperar més d'un 95\% dels valors originals, permetent per tant la quantificació absoluta de les imatges de SPECT.
Many forms of brain diseases are associated with problems in the neurotransmission systems. One approach to the assessment of such systems is the use of Single Photon Emission Computed Tomography (SPECT) brain imaging. Neurotransmission SPECT has become an important tool in neuroimaging and is today regarded as a useful method in both clinical and basic research. SPECT is able to non-invasively visualize and analyze different organs and tissues functions or properties in Nuclear Medicine.

Although visual inspection is often sufficient to assess neurotransmission imaging, quantification might improve the diagnostic accuracy of SPECT studies of the dopaminergic system. In particular, quantification of neurotransmission SPECT studies in Parkinson Disease could help us to diagnose this illness in the early pre-clinical stages. One of the main research topics in SPECT is to achieve early diagnosis, indeed preclinical diagnosis in neurodegenerative illnesses. In this field detailed analysis of shapes and values of the region of interest (ROIs) of the image is important, thus quantification is needed. Moreover, quantification allows a follow-up of the progression of disease and to assess the effects of potential neuroprotective treatment strategies. Therefore, the aim of this thesis is to achieve quantification of both the absolute activity values and the relative values of the reconstructed SPECT images.

Quantification is affected by the degradation of the image introduced by statistical noise, attenuation, collimator/detector response and scattering effects. Some of these degradations may be corrected by using iterative reconstruction algorithms, which thus enable a more reliable quantification. The importance of correcting degradations in reconstruction algorithms to improve quantification accuracy of brain SPECT studies has been proved.

Monte Carlo simulations are the --gold standard' for testing reconstruction algorithms in Nuclear Medicine. We analyzed the available Monte Carlo codes and we chose SimSET as a virtual phantom simulator. A new stopping criteria in SimSET was established in order to reduce the simulation time. The modified SimSET version was validated as a virtual phantom simulator which reproduces realistic projection data sets in SPECT studies.

Iterative algorithms permit modelling of the projection process, allowing for correction of spatially variant collimator response and the photon crosstalk effect between transaxial slices. Thus, our work was focused on the modelling of the collimator/detector response for the parallel and fan beam configurations using the Monte Carlo code PENELOPE. Moreover, a full 3D reconstruction with OS-EM algorithms was developed.

Finally, scattering has recognized to be one of the most significant degradation effects in SPECT quantification. Nowadays this subject is an intensive field of research in SPECT techniques. Monte Carlo techniques appear to be the most reliable way to include this correction. The use of the modified SimSET simulator accelerates the forward projection process although the computational burden is already a challenge for this technique.

Full 3D reconstruction simultaneously applied with Monte Carlo-based scattering correction and the 3D evaluation procedure is a major upgrade technique in order to obtain valuable, absolute quantitative estimates of the reconstructed images. Once all the degrading effects were corrected, the obtained values were 95\% of the theoretical values. Thus, the absolute quantification was achieved.

Brain perfusion weighted images acquired using dynamic contrast studies have an important clinical role in acute stroke diagnosis and treatment decisions. The purpose of my PhD research is to develop novel methodologies for improving the efficiency and quality of brain perfusion-imaging analysis so that clinical decisions can be made more accurately and in a shorter time. This thesis consists of three parts: My research investigates the possibility that parallel computing brings to make perfusion-imaging analysis faster in order to deliver results that are used in stroke diagnosis earlier. Brain perfusion analysis using local Arterial Input Functions (AIF) techniques takes a long time to execute due to its heavy computational load. As time is vitally important in the case of acute stroke, reducing analysis time and therefore diagnosis time can reduce the number of brain cells damaged and improve the chances for patient recovery. We present the implementation of a deconvolution algorithm for brain perfusion quantification on GPGPU (General Purpose computing on Graphics Processing Units) using the CUDA programming model. Our method aims to accelerate the process without any quality loss. Specific features of perfusion source images are also used to reduce noise impact, which consequently improves the accuracy of hemodynamic maps. The majority of existing approaches for denoising CT images are optimized for 3D (spatial) information, including spatial decimation (spatially weighted mean filters) and techniques based on wavelet and curvelet transforms. However, perfusion imaging data is 4D as it also contains temporal information. Our approach using Gaussian process regression (GPR) makes use of the temporal information in the perfusion source imges to reduce the noise level. Over the entire image, our noise reduction method based on Gaussian process regression gains a 99% contrast-to-noise ratio improvement over the raw image and also improves the quality of hemodynamic maps, allowing a better identification of edges and detailed information. At the level of individual voxels, GPR provides a stable baseline, helps identify key parameters from tissue time-concentration curves and reduces the oscillations in the curves. Furthermore, the results show that GPR is superior to the alternative techniques compared in this study. My research also explores automatic segmentation of perfusion images into potentially healthy areas and lesion areas, which can be used as additional information that assists in clinical diagnosis. Since perfusion source images contain more information than hemodynamic maps, good utilisation of source images leads to better understanding than the hemodynamic maps alone. Correlation coefficient tests are used to measure the similarities between the expected tissue time-concentration curves (from reference tissue) and the measured time-concentration curves (from target tissue). This information is then used to distinguish tissues at risk and dead tissues from healthy tissues. A correlation coefficient based signal analysis method that directly spots suspected lesion areas from perfusion source images is presented. Our method delivers a clear automatic segmentation of healthy tissue, tissue at risk and dead tissue. From our segmentation maps, it is easier to identify lesion boundaries than using traditional hemodynamic maps.

In this thesis, the challenges of establishing the very first human in vivo 23Na magnetic resonance imaging in the United Kingdom are presented. A comprehensive framework for quantitative human in vivo studies is established and translated for clinical research imaging. Quantitative measures to obtain the sodium bioscale using external calibrants are discussed and results from a scan re-scan reproducibility study on five healthy volunteers are presented. The protocol was subsequently used in a clinical research study and published by a clinical collaborator in a high impact journal. Improvements in acquisition are achieved by implementation and 23Na adapt- ation of the state of the art 3D-Cones pulse sequence. For evaluation, it is compared against more established 3D-radial k-space sampling schemes featuring cylindrical stack-of-stars (SOS) and 3D-spokes kooshball trajectories on five healthy volunteers in a clinical setting and numerical phantoms. Signal- to-noise ratio (SNR) as a measurement of sequence performance was compared between the sequences and the results are presented. The results were published in a special issue on X-nuclei imaging in the journal of Magnetic Resonance Materials in Physics, Biology and Medicine. The work was subsequently shortlisted and presented for the Young Investigator Awards at the annual meeting of the European Society for Magnetic Resonance in Medicine and Biology. Reconstruction improvements by means of sophisticated k-space weighting schemes are presented on numerical and in vivo data and its effects on image appearance, SNR and total tissue sodium concentration estimates are discussed. The work is currently in peer review for journal publication. A protocol for clinically feasible in vivo 23Na relaxometry measurements of the transverse relaxation time constant T2 is established and results for a range of anatomical white and grey matter locations is presented using both a bi-exponential two-component fit and an unrestricted continuous distribution model. The implications of the results on the underlying tissue sodium environment are discussed. This work has subsequently been presented at international conferences of the International Society for Magnetic Resonance in Medicine and European Society for Magnetic Resonance in Medicine and Biology and has been submitted for peer review as a journal publication. As a conclusion I discuss how the methods presented here can be used to obtain unprecedented spatial and temporal resolution in in vivo 23Na imaging at 3T. Preliminary results are presented.

Thesis (M.A.)--Boston University PLEASE NOTE: Boston University Libraries did not receive an Authorization To Manage form for this thesis or dissertation. It is therefore not openly accessible, though it may be available by request. If you are the author or principal advisor of this work and would like to request open access for it, please contact us at open-help@bu.edu. Thank you.
Alzheimer's disease is the sixth leading cause of death in the United States. While the pathology of the disease is not fully understood, it is becoming increasingly apparent that it involves a complex homeostatic system involving multiple metals, including zinc, copper, and iron. There is also growing evidence that demonstrates developmental lead exposure may also have a role in the pathogenesis of the disease. Understanding the role of these elements in Alzheimer's disease and other metal dyshomeostasis related maladies is key in the development of treatments and possible cures. The development of metallomic imaging using systems like Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS) shows great promise in tracking the distribution of individual elements in physiological tissues. However, the process is both time- and resource-consuming. In an effort to alleviate these issues, we developed a method for creating calibration standards for both LA-ICP-MS and LA-ICP-OES (Laser Ablation Inductively Coupled Plasma Optical Emission Spectrometry) and a method for creating 60µm sections for laser ablation. In addition, we also explored the capabilities and sensitivity of a LA-ICP-OES system for metallomic imaging using murine brains. While imaging of the 60µm sections will require additional calibration and fine-tuning, we were able to successfully image and identify physiological areas of interest in the murine brain by elemental distribution. Continued development of this technology will lead to better optical emission spectrometry image resolution, while freeing up the LAICP-MS for ultra-trace elemental and isotopic analysis.

Thesis: Ph. D. in Medical Engineering and Medical Physics, Harvard-MIT Program in Health Sciences and Technology, 2019
Cataloged from PDF version of thesis.
Includes bibliographical references.
An intricate interplay of signaling molecules underlies brain activity, yet studying these molecular events in living whole organisms remains a challenge. Magnetic resonance imaging (MRI) is the most promising imaging modality for development of molecular signaling sensors with deeper tissue penetration than optical imaging, and better spatial resolution and more dynamic potential in sensor design, compared to radioactive probes. MRI molecular sensors, however, have largely required micromolar concentrations to achieve detectable signals. In order to detect signaling molecules in the brain at their native low nanomolar concentrations, an improvement in MRI molecular sensors is necessary. Here we introduce a new in vivo imaging paradigm that uses vasoactive probes (vasoprobes) that couple molecular signals to vascular responses. We apply the vasoprobes to detect molecular targets at nanomolar concentrations in living rodent brains, thus satisfying the sensitivity requirement for imaging endogenous signaling events. Even with more sensitive probes, molecular imaging of the brain is further complicated by the presence of the blood-brain barrier (BBB), designed by nature to protect this most vital of organs. We have therefore implemented a means to permit noninvasive delivery of imaging agents following ultrasonic BBB opening. We use the ultrasound technique to deliver another potent class of contrast agents, superparamagnetic iron oxides, and we show that effective permeation of brain tissue is achieved using this approach. We have also designed ultrasensitive vasoprobe variants designed to permeate the brain completely noninvasively, using endogenous transporter-mediated mechanisms. We present preliminary results based on this approach and discuss future directions.
by Agata E. Wiśniowska.
Ph. D. in Medical Engineering and Medical Physics
Ph.D.inMedicalEngineeringandMedicalPhysics Harvard-MIT Program in Health Sciences and Technology

The brain activities are characterized by spontaneous and persistent irregular fluctuations in space and time. Criticality theory from statistical physics has been proposed as a principle to explain the variability in normal brain spontaneous activity and has suggested the functional benefits of variability, such as maximized dynamic range of response to stimuli and information capacity. In parallel, the brains show variability in other aspects, such as the structural heterogeneity across brain regions, the intra-individual variability across experimental trials, and the behavior difference across groups and individuals. The associations between the variability of spontaneous activities and these different types of structural, intra and inter-individual variabilities remain elusive. My doctoral study thus aimed to bridge the brain variability and the above-mentioned variations based on criticality theory and analysis of empirical data. As a preparatory analysis, we first collected evidence to prove criticality in human functional magnetic resonance imaging (fMRI) data. The advanced statistical criteria were used to exclude potential artefacts that can induce power-law scaling without the mechanism of criticality. In the first part of the study, we addressed methodological issue and tested whether several measures of either spatial or temporal complexity due to experimental limitations could be reliable proxy of spatiotemporal variability (related to criticality) in vivo. The high spatiotemporal resolutions of whole-cortex optical voltage imaging in mice brain during the waking up from anesthesia enabled simultaneous investigation of functional connectivity (FC), Multi-Scale Entropy (MSE, measure of temporal variability), Regional Entropy (RE, quantity of spatiotemporal variability) and the interdependency among them under different brain states. The results suggested that MSE and FC could be effective measures to capture spatiotemporal variability under limitation of imaging modalities applicable to human subjects. This study also lays methodological basis for the third study in this thesis. In the second study, we explored the interaction between spontaneous activity and evoked activity from mice brain imaging under whisker stimulus. The whisker stimulus will first evoke the local activation in sensory cortex and then trigger whole-cortex activity with variable patterns in different experimental trials. This trial-to-trial variability in the cortical evoked component was then attributed to the changes of ongoing activity state at stimulus onset. The study links ongoing activity variability and evoked activity variability, which further consolidates the association between ongoing activity and brain functions. In the third study, we measured the signal variability of the whole brain from resting state fMRI, and developed the multivariate pattern of cortical entropy, called entropy profile, as reliable and interpretable biomarker of individual difference in cognitive ability. We showed that the whole cortical entropy profile from resting- state fMRI is a robust personalized measure. We tested the predictive power for general and specific cognitive abilities based on cortical entropy profiles with out- of-sample prediction. Furthermore, we revealed the anatomical features underlying cross-region and cross-individual variations in cortical entropy profiles. This study provides new potential biomarker based on brain spontaneous variability which could benefit the applications in psychology and psychiatry studies. The whole study laid a foundation for brain criticality-/variability-based studies and applications and broadened our understanding of the associations between neural structures, functional dynamics and cognitive ability

Thesis (M. S.)--Biomedical Engineering, Georgia Institute of Technology, 2009.
Committee Chair: James K. Rilling; Committee Chair: Xiaoping Hu; Committee Member: Shella Keilholz; Committee Member: Todd M. Preuss.

Magnetic resonance imaging (MRI) is a valuable imaging technique. It provides excellent soft tissue contrast and multi-parametric non-invasive imaging protocols. Among those various techniques, diffusion MRI measures the water diffusion properties of biological tissue. It is a useful tool in characterizing various brain tissue microstructures quantitatively. With its rapid development, it is emerging that subtle changes can be probed by diffusion tensor imaging (DTI) quantitation. The objectives of this doctoral work are to access the subtle microstructural alterations in rodent brains due to hemodynamic changes, fear conditioning and sleep deprivation using in vivo DTI. With the improved reproducibility and specificity achieved by using advanced post-processing and animal preparation procedures, in vivo DTI is expected to be useful to explore the underlying biological mechanisms for posttraumatic stress disorder and memory deficit following sleep loss in human. Firstly, as DTI could be influenced by the presence of water molecules in brain vasculature, for better understand the reproducibility and sensitivity of in vivo DTI measurements, conventional DTI protocol was applied to a well-controlled rat model of hypercapnia. Our data demonstrated that diffusivities increased in whole brain, gray and white matter regions in response to hypercapnia. These results indicate that in vivo DTI quantitation in brain can be interfered by vascular factors on the order of few percents. Cautions should be taken in designing and interpreting quantitative DTI studies as all DTI indices can be potentially confounded by physiologic conditions, cerebrovascular and hemodynamic characteristics. Secondly, recent DTI studies have shown detection of long-term neural plasticity weeks to months following relatively extensive periods of training in animals. However, rapid plasticity within a short period (24 hours) after learning is important for observing the time course of training-evoked changes and narrow down candidate mechanisms. Fear conditioning (FC), which typically occurs over a short timescale (in minutes), was selected as a paradigm for investigation. Using voxel-wise repeated measures analysis, FA was found to increase in amygdala and decrease in hippocampus 1-hour post-FC, and it reversed in both regions 1-day post-FC. Results indicate that DTI can detect rapid microstructural changes in brain regions known to mediate fear conditioning in vivo. DTI indices could be explored as a translational tool to capture potential early biological changes in individuals at risk for developing post-traumatic stress disorder. Thirdly, in vivo DTI was employed to access regional specific microstructural changes following rapid eye movement sleep deprivation (SD), and explore possible temporal differentiation of these changes. With voxel-base analysis, MD is found to decrease in post-SD time points in bilateral hippocampi and cerebral cortex. The distributions of these clusters exhibited differentiable layer specific patterns, which were pointing to dentate gyrus and CA1 layer in hippocampus, and parietal cortex and barrel field layers in cerebral cortex. Results indicate that in vivo DTI is capable to detect microstructural changes in specific layers and reveal temporal distinction between them. These specific layers are possibly more susceptible to sleep loss, and the temporal distinction of changes between these layers might underlie learning and memory decline after SD.
published_or_final_version
Electrical and Electronic Engineering
Doctoral
Doctor of Philosophy

Intraoperative ultrasound strain imaging of the brain visualizes brain tissue deformation as an image. The hypothesis is that strain and elastographic values can be used to complement conventional B-mode image in the task of brain tumor segmentation. A key asumption is that the natural pulsation of the cerebral arteries causes deformation in the brain tissue that is measurable with ultrasound.Strain values are found with a least-squares technique that estimates the spatial derivative of axial velocity, which in turn is mea- sured using a phase-based velocity estimator. A correlation coefficient is calculated for each estimate, giving an indicator of estimation accuracy. Additionally a method for hiding estimates of bad quality based on correlation coefficient thresholding is demonstrated. More- over, a novel elastographic processing technique suitable for cineloop display is introduced. This method extracts a stiffness parameter from a series of strain images, producing an elastogram. A graphical user interface allowing the user to change parameters and see the corresponding result in real-time, minimizing the time needed for parameter optimization, has been developed.The method has been tested using an elasticity phantom. The phantom elastogram cineloop shows a live image that visualizes the difference between stiff and soft tissue well, portraying information not found in the B-mode image. The conclusion is that the proposed elastographic technique, combined with correlation coefficient thresholding, produces elastograms that are suitable for real-time display. This technique is not limited to imaging of the brain, and could, with different parameters, be used for imaging other parts of the body as well.Clinical data sets from two brain tumor patients have been studied as well, where the estimated velocity, strain and elastographic values is discussed in detail. In both patients the tissue movement due to arterial pulsation was measurable with ultrasound. For one patient, a correlation was found between tissue pathology and estimated strain and elastographic values. For the second patient the strain and elastographic processing broke down, and no similar correlation was found.

In the first section of this thesis structural brain imaging is presented as a powerful tool for investigating the biological correlates of schizophrenia, thereby informing the ongoing evolution of the concept. The major pathophysiological hypotheses of schizophrenia are discussed from the perspective of structural brain imaging. The contemporary literature is then reviewed before three potential strategies for future structural imaging research are presented: (1) imaginative clinical study design employing carefully selected sub-groups from within the schizophrenic population; (2) adoption of newly developed approaches to MRI data acquisition and analysis; (3) utilisation of novel structural imaging techniques. In the second part of this thesis three studies are presented, each illustrating one of the strategies described above. Study 1: Structural magnetic resonance imaging (MRI) of the brain in presumed carriers of gene(s) for schizophrenia, their affected and unaffected siblings. The aim of this study was to establish if the gene(s) for schizophrenia are associated with specific abnormalities of brain structure. Six sib-ships from multiple affected families were recruited. Each sib-ship consisted of one patient with schizophrenia, one ‘obligate carrier’ without the disorder but with an affected child and one ‘non-affected non-carrier’. MRI was conducted with a semi-automated region of interest analysis. Between-group comparisons were tested by repeated measures analysis of variance. Reductions in volumes of cortical structures and of whole brain were found only in schizophrenics and therefore appear to be associated with phenotype. In contrast, reduced volume of the amygdalo-hippocampal complex (AHC) was found in both schizophrenics and obligates and therefore appears to be associated with genetic risk for the disorder even in the absence of disease. Reduced AHC volume may therefore represent an endophenotypic marker for schizophrenia. Study 2: A voxel based morphometry (VBM) and region of interest (ROI) analysis of the genotypic and phenotypic neuroanatomy of schizophrenia. The aim of this study was to explore the likely impact of a new method of data analysis upon future structural brain imaging research. The MRI data set from study 1 (above) was analysed using a voxel-based statistical technique (VBM) and a conventional ROI approach. The results obtained by the two methods were then compared. Overall, the results obtained by VBM were compatible with those obtained by ROI. However, the extent of the overlap varied according to the statistical methods employed. Reassuringly, maximal agreement was found when the ‘optimal’, most methodologically appropriate, VBM analysis was compared with the ‘optimal’ ROI analysis (from study 1). Study 3: Diffusion tensor imaging (DTI) and proton magnetic resonance spectroscopy (¹H-MRS) in schizophrenic subjects and normal controls. This study was included to illustrate the challenges inherent in the adoption of novel imaging techniques. The study was designed to identify anatomical correlates of functional dysconnectivity between the pre-frontal and temporal regions in schizophrenia.

Motion sickness (MS) has been experienced for thousands of years, yet much is still unknown about this disorder. For example, the purpose of MS is not yet understood, nor is the underlying neuroanatomy and neurophysiology of the disorder known.A recent theory states that MS is a mechanism that limits inappropriate, self-generated motor strategies that can cause inadvertent changes in the function of the vestibular system and therefore lead to disordered postural, locomotor, and gaze control (Watt et al., 1992). In light of this theory, MS may best be studied under actively- rather than passively-generated conditions. For that reason, self-generated coriolis stimulation (CS) was used to induce MS in susceptible subjects in this thesis. An important feature of CS is that for a given head movement, the pattern of vestibular stimulation depends on the direction of whole-body of rotation. The thesis consists of three parts. First, a method had to be devised to reposition subjects accurately within the positron emission tomography (PET) scanner after they performed the active, MS-inducing stimulus. Secondly, the effects of CS were assessed in a functional brain imaging study. Positron emission tomography was used to determine which brain areas are active when a person experiences the signs and symptoms of, and emotional reactions to, MS. Thirdly, the consequences of the direction-specific vestibular stimulation patterns of CS were studied by determining the effect of direction of rotation on adaptation to CS.As a result of these experiments, a safe and effective head holder was developed, some of the brain structures involved in MS were revealed, and a unique method for producing MS in a laboratory setting was further characterized.
Le mal des transports, bien qu'expérimenté depuis des milliers d'années, est somme toute majoritairement méconnu. Entre autre, la raison d'être de ce trouble n'est pas encore comprise ainsi que sa neuroanatomie et sa neurophysiologie sous-jacentes.Une récente théorie stipule que le mal des transports serait le résultats d'un mécanisme qui limiterait certaine activités moteurs volontaires inappropriées pouvant causer des changements involontaires de la fonction vestibulaire et donc, mener à une distorsion de la posture, des patrons moteurs et du contrôle visuel (Watt et al., 1992). À la lumière de cette théorie, le mal des transports peut probablement être mieux étudié dans des conditions de mouvements actifs plutôt que passifs. Pour cette raison, dans cette thèse, la stimulation coriolis (SC) autogénérée a été utilisée afin d'induire le mal des transports chez des sujets susceptibles. Une caractéristique importante de la SC est que pour un mouvement de tête donné, le patron de la stimulation vestibulaire dépend de la direction globale de la rotation du corps.Cette thèse consiste en trois parties. Premièrement, une méthode a dû être conçue afin de pouvoir repositionner les sujets de façon precise à l'intérieur du scannographe de tomographie par émission de positons après qu'ils aient effectué la stimulation active du mal des transports. Deuxièment, les effets de la SC furent évalués par une étude d'imagerie cérébrale fonctionnelle. La tomographie par émission de positons a été utilisé pour déterminer quelle partie du cerveau sont actives lorsqu'une personne éprouve les signes, symptômes et réactions émotionnelles du mal des transports. Troisièment, les conséquences des patrons spécifiques à la direction de la stimulation vestibulaire par la SC ont été étudiés par la détermination des effets du sens de la rotation sur l'adaptation à la SC.À la suite de ces expériences, un support pour la tête sécuritaire et efficace fut développé, quelques structures du cerveau impliquées dans le mal des transports ont été révélées et une méthode unique générant le mal des transports en laboratoire a plus amplement été caractérisée.

Brain operates on a network level. Magnetic resonance imaging (MRI) provides structural and functional images noninvasively with large field of view and at high spatial resolution and thus assumes an extremely valuable role in studying brain networks. The objectives of this doctoral work were to develop and apply novel MRI methods on human and rodent brains, for in vivo and global assessments of functional brain networks at resting and task-evoked states. Firstly, the feasibility of passband balanced steady-state free precession (bSSFP) imaging for distortion-free and high-resolution resting-state fMRI (rsfMRI) was investigated. Resting-state networks (RSNs) derived from bSSFP images were shown spatially and spectrally comparable to those derived from conventional gradient-echo echo-planar imaging (GE-EPI) with considerable intra- and inter-subject reproducibility. High-resolution bSSFP corresponded well to the anatomical images, with RSNs exquisitely co-localized to gray matter. Furthermore, RSNs at areas of severe susceptibility were proved accessible including human anterior prefrontal cortex and rat piriform cortex. These findings demonstrated for the first time that passband bSSFP approach can be a promising alternative to GE-EPI for rsfMRI. It offers distortion-free and high-resolution RSNs and is potentially suited for high field studies. Secondly, to examine the macrovascular contributions to the spatial and spectral prosperities of resting-state networks, spin-echo echo-planar imaging (SE-EPI) with moderate diffusion weighting (DW) was proposed for rsfMRI. SE and DW suppressed the extravascular and intravascular contributions from macrovessels respectively. Significantly lower functional connectivity strength was observed in the posterior cingulate cortex of the default mode network derived from DW SE-EPI data comparing to that derived from SE-EPI, suggesting a confounding role played by the intravascular component from large veins, whereas no significant spectral difference was detected. Therefore, the DW SEEPI approach for rsfMRI may assist in better identifying and interpreting largescale brain networks with future improvement in temporal resolution by acceleration techniques and in sensitivity at higher field. Thirdly, rsfMRI was performed to evaluate the intrinsic functional networks in the corresponding anatomical visual brain connections traced by Mn-enhanced MRI (MEMRI). Strengths of resting-state functional connectivity appeared to couple with structural connectivity in MEMRI, demonstrating the sensitivity of these structural and functional connectivity MRI techniques for assessing the neuroarchitecture, neurophysiology and structural-functional relationships in the visual brain in vivo. Fourthly, the hypothesis that a regional activation identified via general linear model analysis of fMRI data reflects the summation of multiple distinct networks that carry different functional purposes was tested. Overlapping frontoparietal networks engaged in a simple single-digit multiplication task were found and their functional roles were evaluated through independent components analysis and contributive source analysis. Future studies incorporating different arithmetic tasks and resting state will shed more light upon how brain accomplishes arithmetic and more complex tasks in general. Lastly, benefiting from higher SNR, better spatial and temporal resolution at higher field, exploratory fMRI studies were conducted on rats at 7 T for in vivo assessments of 1) the effect of dark-rearing on postnatal visual development, 2) sound amplitude modulations and 3) sound frequency modulation sweep direction selectivity in auditory system. (
published_or_final_version
Electrical and Electronic Engineering
Doctoral
Doctor of Philosophy

The work described in this thesis was performed by the author, except where indicated. All the studies were accomplished on the 3 Tesla system within the Magnetic Resonance Centre at the University of Nottingham, and the Wellcome Trust MEG Laboratory at the Aston University during the period between October 1999 and June 2005. Functional Magnetic Resonance Imaging (fMRI) and Magnetoencephalography (MEG) are two promising brain function research modalities, sensitive to the hemodynamic and electrophysiological responses respectively during brain activites. The feasibility of joint employment of both modalities was examined in both spatial and temporal domains. A somatosensory tactile stimulus was adopted to induce simple functional reaction. It was shown that a reasonable spatial correspondence between fMRI and MEG can be established. Attempts were made on MEG recordings to extract suitable aspects for temporal features matching fMRI with a method reflecting the physical principles. It was shown that the this method is capable of exposing the nature of neural electric activities, although further development is required to perfect the strategy.

High resolution brain magnetic resonance (MR) images acquired at multiple time points across the treatment of a patient allow the quantification of localised changes brought about by disease progression. The aim of this thesis is to address the challenge of performing automatic longitudinal analysis of magnetic resonance imaging (MRI) in paediatric brain tumours. The first contribution in this thesis is the validation of a semi-automated segmentation technique. This technique was applied to intra-operative MR images acquired during the surgical resection of hypothalamic tumours in children, in order to assess the volume of tumour resected at different stages of the surgical procedure. The second contribution in this thesis is the quantification of a rare condition known as hypertrophic olivary degeneration (HOD) in lobes within the brain known as inferior olivary nucleii (ION) in relation to the development of posterior fossa syndrome (PFS) following tumour resection in the hind brain. The change in grey-level intensity over time in the left ION has been identified as a suitable biomarker that correlates with the occurrence of posterior fossa syndrome following tumour resection surgery. This study demonstrates the application of machine learning techniques to T2 brain MR images. The third contribution presents a novel approach to longitudinal brain MR analysis, focusing on the cerebellum and brain stem. This contribution presents a technique developed to interpolate multi-slice 2D MR image slices of the brain stem and cerebellum both to infill gaps between slices as well as longitudinally over time, that is, in four-dimensional space. This study also investigates the application of machine learning techniques directly to the MR images. Another novel method developed in this study is the Jacobian of deformations in the brain over time, and its use as an imaging feature. Unlike the previous contribution chapter, the third contribution is not hypothesis-driven, and automatically detects six potential biomarkers that are related to the development of PFS following tumour resection in the posterior fossa. The limited number of patients considered in each study posed a major challenge. This has prompted the use of multiple validation techniques in order to provide accurate results despite the small dataset. These techniques are presented in the second and third contribution chapters.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2010.
Cataloged from PDF version of thesis.
Includes bibliographical references (p. 119-137).
Localizing sources of activity from electroencephalography (EEG) and magnetoencephalography (MEG) measurements involves solving an ill-posed inverse problem, where infinitely many source distribution patterns can give rise to identical measurements. This thesis aims to improve the accuracy of source localization by incorporating spatio-temporal models into the reconstruction procedure. First, we introduce a novel method for current source estimation, which we call the l₁l₂-norm source estimator. The underlying model captures the sparseness of the active areas in space while encouraging smooth temporal dynamics. We compute the current source estimates efficiently by solving a second-order cone programming problem. By considering all time points simultaneously, we achieve accurate and stable results as confirmed by the experiments using simulated and human MEG data. Although the l₁l₂-norm estimator enables accurate source estimation, it still faces challenges when the current sources are close to each other in space. To alleviate problems caused by the limited spatial resolution of EEG/MEG measurements, we introduce a new method to incorporate information from functional magnetic resonance imaging (fMRI) into the estimation algorithm.
(cont.) Whereas EEG/MEG record neural activity, fMRI reflects hemodynamic activity in the brain with high spatial resolution. We examine empirically the neurovascular coupling in simultaneously recorded MEG and diffuse optical imaging (DOI) data, which also reflects hemodynamic activity and is compatible with MEG recordings. Our results suggest that the neural activity and hemodynamic responses are aligned in space. However, the relationship between the temporal dynamics of the two types of signals is non-linear and varies from region to region. Based on these findings, we develop the fMRI-informed regional EEG/MEG source estimator (FIRE). This method is based on a generative model that encourages similar spatial patterns but allows for differences in time courses across imaging modalities. Our experiments with both Monte Carlo simulation and human fMRI-EEG/MEG data demonstrate that FIRE significantly reduces ambiguities in source localization and accurately captures the timing of activation in adjacent functional regions.
by Wanmei Ou.
Ph.D.

Over the past years early detection and intervention in schizophrenia have become a major objective in psychiatry. Early intervention strategies are intended to identify and treat psychosis prior to fulfilling diagnostic criteria for the disorder. To this aim, reliable early diagnostic biomarkers are needed in order to identify a high-risk state for psychosis and also predict transition to frank psychosis in those high-risk individuals destined to develop the disorder. Recently, machine learning methods have been successfully applied in the diagnostic classification of schizophrenia and in predicting transition to psychosis at an individual level based on magnetic resonance imaging (MRI) data and also neurocognitive variables. This work investigates the application of machine learning methods for the early identification of schizophrenia in subjects at high risk for developing the disorder. The dataset used in this work involves data from the Edinburgh High Risk Study (EHRS), which examined individuals at a heightened risk for developing schizophrenia for familial reasons, and the FePsy (Fruherkennung von Psychosen) study that was conducted in Basel and involves subjects at a clinical high-risk state for psychosis. The overriding aim of this thesis was to use machine learning, and specifically Support Vector Machine (SVM), in order to identify predictors of transition to psychosis in high-risk individuals, using baseline structural MRI data. There are three aims pertaining to this main one. (i) Firstly, our aim was to examine the feasibility of distinguishing at baseline those individuals who later developed schizophrenia from those who did not, yet had psychotic symptoms using SVM and baseline data from the EHRS study. (ii) Secondly, we intended to examine if our classification approach could generalize to clinical high-risk cohorts, using neuroanatomical data from the FePsy study. (iii) In a more exploratory context, we have also examined the diagnostic performance of our classifier by pooling the two datasets together. With regards to the first aim, our findings suggest that the early prediction of schizophrenia is feasible using a MRI-based linear SVM classifier operating at the single-subject level. Additionally, we have shown that the combination of baseline neuroanatomical data with measures of neurocognitive functioning and schizotypal cognition can improve predictive performance. The application of our pattern classification approach to baseline structural MRI data from the FePsy study highly replicated our previous findings. Our classification method identified spatially distributed networks that discriminate at baseline between subjects that later developed schizophrenia and other related psychoses and those that did not. Finally, a preliminary classification analysis using pooled datasets from the EHRS and the FePsy study supports the existence of a neuroanatomical pattern that differentiates between groups of high-risk subjects that develop psychosis against those who do not across research sites and despite any between-sites differences. Taken together, our findings suggest that machine learning is capable of distinguishing between cohorts of high risk subjects that later convert to psychosis and those that do not based on patterns of structural abnormalities that are present before disease onset. Our findings have some clinical implications in that machine learning-based approaches could advise or complement clinical decision-making in early intervention strategies in schizophrenia and related psychoses. Future work will be, however, required to tackle issues of reproducibility of early diagnostic biomarkers across research sites, where different assessment criteria and imaging equipment and protocols are used. In addition, future projects may also examine the diagnostic and prognostic value of multimodal neuroimaging data, possibly combined with other clinical, neurocognitive, genetic information.

The present thesis comprises three main parts: one theoretical and two experimental. The first part, composed of two chapters, will introduce the clinical and neuropathological features underlying parkinsonian disorders, namely in Parkinson’s disease (PD) as well as in atypical parkinsonisms — multiple system atrophy (MSA) and progressive supranuclear palsy (PSP) (Chapter 1). In this regard, PD and MSA are defined as synucleinopathies due to the presence of synuclein aggregates; while PSP that is characterized by tau protein accumulations, is part of tauopathies. Further, Chapter 2 will provide an overview of the cognitive dysfunctions characterizing these disorders, as well as evidence on the biological mechanisms and structural changes underlying cognitive alterations. The second and third parts are composed by studies I conducted during my doctoral research. Namely, in Chapter 3, I report results of my studies on cognitive screening instruments most sensitive in detecting cognitive alterations in atypical parkinsonisms compared to PD. In the following study, I characterized the progression of cognitive decline in these disorders (Chapter 4). Finally, I investigated with magnetic resonance imaging the structural changes underlying cognitive alterations in PD (Chapter 5), and MSA (Chapter 6). I conclude this thesis by discussing the clinical consequences of these cognitive and imaging findings (Chapter 7). PART I - Theoretical background Chapter 1: Parkinsonian disorders Parkinsonian disorders are characterized by different underlying pathologies. In PD and MSA there are synuclein aggregates respectively in dopamine neurons or in glial cells, while PSP patients present pathological aggregation of the tau-protein, resulting in neurofibrillary tangles formation (Daniel, de Bruin, & Lees, 1995; Dickson, 1999). Clinical manifestations depend by the characteristics of protein aggregation and by the extent of disease spread to cortical and subcortical regions (Halliday, Holton, Revesz, & Dickson, 2011). Thus, the present chapter will overview the underlying pathology of PD, MSA and PSP; and it will describe the different clinical features; and lastly review the most recent diagnostic criteria (e.g., Gelb, Oliver, & Gilman, 1999; Gilman et al., 2008; Höglinger et al., 2017). Chapter 2: Cognitive features and their underlying mechanisms in parkinsonian disorders Non-motor symptoms represent a crucial part of the parkinsonian disorders spectrum; and cognitive dysfunctions, including dementia, are probably the most relevant, since they affect functional independence of patients, increase caregiver burden as well as wield a considerable socioeconomic impact (Keranen et al., 2003; McCrone et al., 2011; Vossius, Larsen, Janvin, & Aarsland, 2011). The first part of this chapter will provide an overview on cognitive dysfunctions in PD, MSA, and PSP. Moreover, the clinical criteria for the diagnosis of mild cognitive impairment and dementia in PD will be reported (Dubois et al., 2007; Emre et al., 2007; Litvan et al., 2012), while so far there are no available criteria to assess cognitive syndromes in PSP and MSA. Lastly, the second and third parts of this chapter will review the evidence on biological mechanisms and structural changes underlying cognitive alterations in these disorders. PART II - Studies on cognitive manifestations in parkinsonian disorders Chapter 3: Montreal Cognitive Assessment and Mini-Mental State Examination performance in progressive supranuclear palsy, multiple system atrophy and Parkinson’s disease There is general agreement that cognitive dysfunctions are common in PD as well as in other parkinsonian disorders (Aarsland et al., 2017; Brown et al., 2010; Gerstenecker, 2017). Brief screening cognitive scales can be adopted in routine care, to support the clinician in the diagnostic process (Marras, Troster, Kulisevsky, & Stebbins, 2014). The Mini-Mental State Examination (MMSE) is the most widely used (Folstein, Folstein, & McHugh, 1975) although MMSE is relatively insensitive in detecting cognitive deficits in parkinsonian disorders mainly because it does not investigate the fronto-executive domain (Hoops et al., 2009). Conversely, the Montreal Cognitive Assessment (MoCA), another brief cognitive screening tool widely used with PD patients (Nasreddine et al., 2005), showed high sensitivity and specificity in the assessment of cognitive dysfunctions in PD (Gill, Freshman, Blender, & Ravina, 2008; Hoops et al., 2009; Zadikoff et al., 2008), as well as also in several neurodegenerative conditions such as Alzheimer’s disease, dementia with Lewy bodies (DLB) and Huntington’s disease (Biundo et al., 2016b; Hoops et al., 2009; Nasreddine et al., 2005; Videnovic et al., 2010). However, MoCA has been poorly investigated in atypical parkinsonisms — especially in PSP and MSA (Kawahara et al., 2015). Thus, this study’s main aim was to determine if MoCA is more sensitive than the commonly used MMSE in detecting cognitive abnormalities in patients with probable PSP and MSA, compared to PD. In this multicenter study across three European institutions, MMSE and MoCA were administered to 130 patients: 35 MSA, 30 PSP and 65 age, and education and sex matched-PD. We assessed between-group differences for MMSE, MoCA, and their subitems and calculated receiver operating characteristic (ROC) curves. Our results show that the mean MMSE is higher than the mean MoCA score in each patient group: MSA (27.7 ± 2.4 vs. 22.9 ± 3.0, p<0.0001), PSP (26.0 ± 2.9 vs. 18.2 ± 3.9, p<0.0001), and PD (27.3 ± 2.0 vs. 22.3 ± 3.5, p<0.0001). Furthermore, MoCA total score as well as its letter fluency subitem differentiates PSP from MSA and PD with high specificity and moderate sensitivity. Namely, a cut-off score of seven words or less per minute would support a diagnosis of PSP (PSP vs. PD: 86% specificity, 70% sensitivity; PSP vs. MSA: 71% specificity, 70% sensitivity). On the contrary, MMSE presented a ceiling effect for most subitems, except for the ‘bisecting pentagons’, with PSP performing worse than MSA and PD patients. These findings suggest that PSP and MSA, similar to PD patients, may present normal performance on MMSE, but reduced performance on MoCA. To conclude, MoCA is more sensitive than MMSE in detecting cognitive dysfunctions in atypical parkinsonisms, and together with its verbal fluency subitem can be a valuable test to support PSP diagnosis. Chapter 4: Prospective assessment of cognitive dysfunctions in parkinsonian disorders Clinical and research evidence suggests cognitive impairments in parkinsonian disorders are progressive. However, there are only a few longitudinal studies in the literature that investigated cognitive progression in PSP and MSA compared to PD (Dubois & Pillon, 2005; Rittman et al., 2013; Soliveri, 2000). In addition, previous studies are based on brief cognitive screening scales or on neuropsychological assessments that do not extensively investigate the full spectrum of cognitive abilities across the five cognitive domains (i.e., attention/working-memory, executive, memory, visuospatial and language). Furthermore, even though clinical criteria for mild cognitive impairment (MCI) and dementia in PD have been formulated (Dubois et al., 2007; Litvan et al., 2012), it remains to be investigated whether similar criteria might be applied also for atypical parkinsonisms (Marras et al., 2014). Based on these observations, the aims of the present study were to: i) assess the severity of cognitive dysfunctions in PSP and MSA patients using PD-criteria for cognitive statuses (i.e., MCI or dementia); ii) investigate the sensitivity of two widely used cognitive screening instruments, the MMSE and MoCA, in differentiating MSA, PSP and PD global cognitive profile; iii) characterize the progression of cognitive decline on the five cognitive domains and behavioral features; and to compare the 15-month follow-up profile across the parkinsonian diseases. Our sample included 18 patients with PSP, 12 MSA; and 30 PD patients, matched for age, education and sex. They were evaluated at baseline and at a mean of 15-month follow-up. Demographic and clinical variables were collected. From the cognitive standpoint, I selected a comprehensive neuropsychological battery specifically designed to target cognitive deficits in PD, according to Level II criteria (Dubois et al., 2007; Litvan et al., 2012; Marras et al., 2014). Thus, I applied these criteria also to MSA and PSP since there are no published criteria for atypical parkinsonisms. Statistical non-parametric analyses were used. I found PSP patients had more severe cognitive decline compared to PD and MSA. Namely, after 15-month follow-up, we noted a marked decline in the executive and language domains in the PSP group. Baseline and follow-up evaluations agreed, showing that PSP had a worse performance than PD and MSA patients: especially, in the Stroop test, verbal fluencies (semantic and phonemic) and MoCA. Assessing the severity of cognitive deficits, I found different percentages of cognitive status (i.e., normal cognition vs. MCI vs. dementia) among the three groups. In particular, the percentage of patients with dementia was higher in PSP compared to MSA (33% vs. no patients with dementia) even if disease duration was similar. Among MSA and PSP patients with multidomain MCI at baseline only PSP converted to dementia at follow-up. Then, although the disease duration was longer for PD patients compared with PSP, the proportion of patients who converted to dementia was lower in the PD group compared to PSP (7% vs. 16%), despite both groups having had similar baseline severity of MCI. Overall, these results suggest more rapid and severe cognitive decline in PSP while MSA patients generally have milder deficits. MoCA showed higher sensitivity than MMSE in detecting cognitive changes, especially in PSP. But MoCA was less sensitive than MMSE in detecting cognitive decline at 15-month in PD, suggesting that MMSE is better if one wants to track cognitive changes in PD. Neuropsychiatric features are more common in PSP than PD patients, especially apathy with accompanying low levels of anxiety and depression. Lastly, analysis of subitems revealed that PSP patients had a ‘clinically significant’ worsening after 15-month in the attentive/executive subitems (Trial Making Test part B and Clock drawing). But it has been observed that some patients also improved in specific subtasks at the follow-up. This improvement could be related to their higher medication dose (although the dopaminergic treatment was not significantly different between the baseline and follow-up). However, noteworthy alterations in performance have been seen for subitems sensitive to motor conditions (such as drawing figures and linking circles with a pen), which could affect cognitive outcome, leading to higher performance at follow-up. These limits of MoCA and MMSE scale have already been reported in PD patients (Biundo et al., 2016b; Hu et al., 2014), and maybe are more pronounced in atypical parkinsonisms. Taken together, these findings show that PSP patients were markedly impaired in comparison to the other parkinsonian disorders (MSA and PD) and six years after first symptoms, 33 percent of patients have dementia. This severe progression is possibly associated with the distribution of tau pathology that involves also cortical structures. On the contrary, the pattern of cognitive impairment in MSA is less severe, possibly due to the predominance of subcortical pathology with cortical involvement occurring only secondary to these abnormalities. Thus, these findings recommend using cognitive assessment to help differential diagnosis in atypical parkinsonisms, and to monitor disease progression. PART III - Neuroimaging studies of synucleinopathies Chapter 5: Amyloid depositions affect cognitive and motor manifestations in Parkinson’s disease Cognitive deficits, particularly executive problems, can be observed early in PD (Aarsland, Bronnick, Larsen, Tysnes, & Alves, 2009). Dysfunction of the frontostriatal dopaminergic system may influence the presence of executive and attention problems (Bruck, Aalto, Nurmi, Bergman, & Rinne, 2005), but so far, evidence from dopamine transporter (DAT) imaging are inconsistent (Delgado-Alvarado, Gago, Navalpotro-Gomez, Jimenez-Urbieta, & Rodriguez-Oroz, 2016). In this regard, the neuropathology underlying cognitive impairment in PD is heterogeneous (Irwin, Lee, & Trojanowski, 2013; Kehagia, Barker, & Robbins, 2010) and amyloid deposit involvement with synuclein pathology remains poorly defined, particularly in the disease’s early stages. Thus, this study’s aims were to investigate the interplay between amyloid depositions in frontostriatal pathways, striatal dopaminergic deficit and brain atrophy rates; and their contribution to cognitive defects (i.e., fronto-executive functions) in early-PD. A multicenter cohort of 33 PD patients from the Parkinson's Progression Markers Initiative underwent [18F]florbetaben positron emission tomography (PET) amyloid, [123I]FP-CIT (see Abbreviations List) single-photon emission computed tomography (SPECT), structural magnetic resonance imaging (MRI), clinical and selective cognitive evaluations. Our results showed that high amyloid levels were associated with reduced dopaminergic deficits in the dorsal striatum (as compared to low amyloid levels), increased brain atrophy in frontal and occipital regions and a tendency to show more frequent cognitive impairment in global cognition (as assessed by MoCA) and fronto-executive tests. Of note, amyloid depositions in frontostriatal regions were inversely correlated with cognitive performance. Overall, our findings suggest that early-PD patients with amyloid burden have higher brain atrophy rates and may experience more cognitive dysfunctions (i.e., executive) and motor impairment as compared to amyloid negative subjects. In this regard, our results seem to be aligned with a recent neuropathological hypothesis that considers synaptic axonal damage and dysfunction as the PD key feature (Tagliaferro & Burke, 2016). Indeed, dopaminergic system neurons are particularly vulnerable to synuclein pathology due to their axonal characteristics — long, thin and unmyelinated. This is also confirmed by imaging studies with DAT-binding PET (Caminiti et al., 2017), suggesting that synuclein aggregations in PD can affect synaptic function, and thus signal transmission from the disease’s very early stages. Our findings suggested a possible interaction between synuclein and the coincident amyloid pathology, wherein amyloid burden may facilitate the spread of synuclein (i.e., Lewy bodies) (Toledo et al., 2016), and we speculate that this interaction can further contribute to axonal vulnerability. Thus, consistently with this hypothesis, we conclude that possibly amyloid depositions act synergistically with synuclein pathology and affect PD clinical manifestations. Chapter 6: Brain structural profile of multiple system atrophy patients with cognitive impairment In contrast to other synucleinopathies (e.g., PD and DLB), presence of dementia is considered a non-supporting feature for MSA diagnosis (Gilman et al., 2008), however there is growing evidence that MSA patients can experience cognitive impairment ranging from executive dysfunctions to multiple-domain cognitive deficits, and in a few cases, also dementia (Gerstenecker, 2017). MMSE is a commonly used global cognitive scale and recently a large multicenter study has suggested using a cutoff score below 27 to increase the MMSE sensitivity in identifying cognitive dysfunctions in MSA (Auzou et al., 2015). Underlying mechanisms of cognitive impairment in MSA are still not understood, and in this regard evidence from MRI studies suggested a discrete cortical and subcortical contribution to explain cognitive deficits (Kim et al., 2015; Lee et al., 2016a), even though these findings were based on a relatively small number of patients at various disease stages as well as being single-center. Thus, the aim of our multicenter study was to better characterize the anatomical changes associated with cognitive impairment in MSA and to further investigate the cortical and subcortical structural differences in comparison to a sample of healthy subjects. We examined retrospectively 72 probable MSA patients and based on the MMSE threshold below 27, we defined 50 MSA as cognitively normal (MSA-NC) and 22 with cognitive impairment (MSA-CI). We directly compared the MSA subgroup, and further compared them to 36 healthy subjects using gray- and white-matter voxel-based morphometry and fully automated subcortical segmentation. Compared to healthy subjects, MSA patients showed widespread cortical (i.e., bilateral frontal, occipito-temporal, and parietal areas), subcortical, and white matter alterations. However, the direct comparison MSA-CI showed only focal volume reduction in the left dorsolateral prefrontal cortex compared with MSA-NC. These findings suggest only a marginal contribution of cortical pathology to cognitive deficits in MSA. Hence, we suggest that cognitive alterations are driven by focal frontostriatal degeneration that is in line with the concept of ‘subcortical cognitive impairment’.
La presente tesi è formata da tre parti principali: la prima teorica mentre le due seguenti sono sperimentali. La prima parte, composta di due capitoli, introdurrà le caratteristiche cliniche e neuropatologiche sottostanti ai disturbi parkinsoniani, in particolare nella malattia di Parkinson (PD) e nei parkinsonismi atipici — atrofia multisistemica (MSA) e paralisi progressiva sopranucleare (PSP) (Capitolo 1). Nello specifico, PD ed MSA sono definite come sinucleinopatie per la presenza di aggregati di sinucleina, mentre la PSP che è caratterizzata dall’accumulo di proteina tau rientra a far parte delle tauopatie. Invece, il Capitolo 2 fornirà una panoramica delle disfunzioni cognitive che caratterizzano questi disturbi e fornirà inoltre evidenze circa i meccanismi biologici e i cambiamenti strutturali che sono alla base delle alterazioni cognitive. Nella seconda e la terza parte sono riportati alcuni studi che ho condotto durante il dottorato di ricerca. In particolare, nel Capitolo 3 riporto i risultati dei miei studi sugli strumenti di screening cognitivo più sensibili nel rilevare alterazioni cognitive nei parkinsonismi atipici rispetto ai pazienti con PD. Nel successivo studio invece ho investigato la progressione del declino cognitivo in questi disturbi (Capitolo 4). Infine, ho investigato con studi di risonanza magnetica i cambiamenti strutturali che sottendono le alterazioni cognitive nel PD (Capitolo 5) e nella MSA (Capitolo 6). Seguiranno le conclusioni generali, in cui discuto le conseguenze cliniche dei risultati ottenuti negli studi cognitivi e di imaging (Capitolo 7). PARTE I – Background teorico Capitolo 1: I disturbi parkinsoniani I disturbi parkinsoniani sono caratterizzati da una diversa patologia sottostante. Nel PD ed MSA ci sono aggregati di sinucleina rispettivamente nei neuroni dopaminergici o nelle cellule gliali, mentre i pazienti con PSP presentano delle aggregazioni di proteina tau che determina la formazione di ammassi neurofibrillari (Daniel, de Bruin, & Lees, 1995; Dickson, 1999). Le manifestazioni cliniche dipendono dalle caratteristiche di aggregati proteici e dall’entità di diffusione della malattia nelle regioni corticali e sottocorticali (Halliday, Holton, Revesz, & Dickson, 2011). Quindi, il presente capitolo illustrerà la patologia sottostante nel PD, MSA e PSP, saranno poi descritte le diverse caratteristiche cliniche ed infine, saranno presentati i più recenti criteri diagnostici di questi disturbi (e.g., Gelb, Oliver, & Gilman, 1999; Gilman et al., 2008; Höglinger et al., 2017). Capitolo 2: Caratteristiche cognitive e i sottostanti meccanismi nei disturbi parkinsoniani I sintomi non-motori rappresentano una parte cruciale dello spettro dei disturbi parkinsoniani, in particolare le disfunzioni cognitive, inclusa la demenza, sono probabilmente tra i sintomi non-motori più rilevanti, in quanto influenzano l'autonomia funzionale dei pazienti, incrementano il carico di gestione del caregiver ed hanno un notevole impatto socioeconomico (Keranen et al., 2003; McCrone et al., 2011; Vossius, Larsen, Janvin, & Aarsland, 2011). La prima parte di questo capitolo fornirà una panoramica sulle disfunzioni cognitive nel PD, MSA e PSP. Saranno inoltre riportati i criteri clinici per la diagnosi di declino cognitivo lieve e di demenza nel PD (Dubois et al., 2007; Emre et al., 2007; Litvan et al., 2012), al contrario invece non esistono al momento criteri disponibili per valutare le sindromi cognitive in PSP e MSA. Infine, la seconda e la terza parte di questo capitolo forniranno evidenze sui meccanismi biologici e sui cambiamenti strutturali sottostanti alle alterazioni cognitive in questi disturbi. PARTE II - Studi sulle manifestazioni cognitive nei disturbi parkinsoniani Capitolo 3: Performance al Montreal Cognitive Assessment e Mini-Mental State Examination nella paralisi sopranucleare progresiva, atrofia multisistemica e malattia di Parkinson Vi è un generale consenso nel riconoscere che le alterazioni cognitive siano frequenti nei PD e negli altri disturbi parkinsoniani (Aarsland et al., 2017; Brown et al., 2010; Gerstenecker, 2017). Pertanto, nella pratica clinica possono essere adottate delle scale brevi di screening cognitivo, per supportare il clinico nel processo diagnostico (Marras, Troster, Kulisevsky, & Stebbins, 2014). Il Mini-Mental State Examination (MMSE) è la scala più utilizzata (Folstein, Folstein, & McHugh, 1975), anche se MMSE è relativamente insensibile nell’identificare rilevare disfunzioni cognitive nei disturbi parkinsoniani principalmente perché non indaga il dominio fronto-esecutivo (Hoops et al., 2009). Al contrario, il Montreal Cognitive Assessment (MoCA), un altro strumento di screening cognitivo ampiamente utilizzato nei pazienti con PD (Nasreddine et al., 2005), ha mostrato un’elevata sensibilità e specificità nell’identificazione di alterazioni cognitive nei PD (Gill, Freshman, Blender, & Ravina, 2008; Hoops et al., 2009; Zadikoff et al., 2008), come anche in altre malattie neurodegenerative come l’Alzheimer, la demenza da corpi di Lewy (DLB) e la malattia di Huntington (Biundo et al., 2016b; Hoops et al., 2009; Nasreddine et al., 2005; Videnovic et al., 2010). Tuttavia, vi sono poche evidenze sull’uso del MoCA nei parkinsonismi atipici, in particolare nella PSP ed MSA (Kawahara et al., 2015). Pertanto, lo scopo del presente studio era di determinare se il MoCA fosse più sensibile del comunemente utilizzato MMSE nel rilevare alterazioni cognitive nei pazienti con probabile PSP e MSA, rispetto al PD. In questo studio multicentrico, che ha coinvolto altri tre centri europei, sono state somministrate le scale MMSE e MoCA a 130 pazienti: 35 MSA, 30 PSP e 65 pazienti PD appaiati per età, scolarità e sesso. Sono state valutate le differenze tra i gruppi per MMSE, MoCA, e i loro subitem; infine sono state calcolate le curve ROC (Receiver-Operating Characteristic). Dai risultati emerge che la media del MMSE è superiore al punteggio medio del MoCA in ogni gruppo di pazienti: MSA (27.7 ± 2.4 vs. 22.9 ± 3.0, p<0.0001), PSP (26.0 ± 2.9 vs. 18.2 ± 3.9, p<0.0001), e PD (27.3 ± 2.0 vs. 22.3 ± 3.5, p<0.0001). Inoltre, il punteggio totale MoCA così come il suo subitem di fluenza fonemica è in grado di differenziare la PSP da MSA e PD con un’alta specificità e moderata sensibilità. Specificamente, un punteggio uguale o inferiore a sette parole al minuto sembra supportare una diagnosi di PSP (PSP vs PD: 86% specificità, sensibilità al 70%, PSP vs MSA: 71% specificità, sensibilità al 70%). Al contrario, nel MMSE è stato possibile osservare un ‘effetto-soffitto’ per la maggior parte dei subitem, ad eccezione del subitem dei ‘due pentagoni’, in cui i pazienti con PSP hanno una prestazione peggiore rispetto a MSA e PD. I nostri risultati suggeriscono che PSP ed MSA, similmente al PD, possono presentare una prestazione normale al MMSE ma deficitaria al MoCA. In conclusione, il MoCA è più sensibile del MMSE nel rilevare disfunzioni cognitive nei parkinsonismi atipici ed insieme al suo subitem di fluenza verbale sembra essere un valido test per supportare una diagnosi di PSP. Capitolo 4: Valutazione prospettica delle disfunzioni cognitive nei disturbi parkinsoniani Evidenze in ambito clinico e di ricerca suggeriscono che le disfunzioni cognitive nei disturbi parkinsoniani siano progressive. Tuttavia, in letteratura vi sono pochi studi longitudinali che indagano la progressione cognitiva in pazienti con PSP ed MSA rispetto a pazienti PD (Dubois & Pillon, 2005; Rittman et al., 2013; Soliveri, 2000). In particolare, i precedenti studi si basano solo su scale globali di screening cognitivo, oppure su valutazioni neuropsicologiche parziali che non esaminano l'intero spettro delle abilità cognitive nei cinque domini (i.e., attenzione/memoria di lavoro, esecutivo, mnesico, visuospaziale e del linguaggio). Inoltre, sebbene siano stati formulati criteri clinici per la diagnosi di declino cognitivo lieve (MCI) e di demenza in pazienti PD (Dubois et al., 2007; Litvan et al., 2012), rimane ancora da investigare se tali criteri possano essere applicati anche nei parkinsonismi atipici (Marras et al., 2014). Date tali premesse, gli obiettivi del presente studio sono stati: i) valutare la severità delle alterazioni cognitive in pazienti PSP ed MSA utilizzando i criteri validati nei pazienti PD, per identificare gli stati cognitivi (i.e., MCI o demenza); ii) esaminare la sensibilità di due strumenti di screening cognitivo ampiamente utilizzati, (i.e., MMSE e MoCA), nel differenziare il profilo cognitivo globale di pazienti MSA, PSP e PD; iii) caratterizzare la progressione del declino cognitivo nei cinque domini, il profilo comportamentale e infine confrontare il profilo cognitivo al follow-up tra i vari disturbi parkinsoniani. Il nostro campione includeva 18 pazienti con PSP, 12 MSA e 30 pazienti con PD appaiati per età, scolarità e sesso, che sono stati valutati alla baseline e al follow-up a 15 mesi. Sono stati raccolti dati demografici e clinici; inoltre dal punto di vista cognitivo è stata selezionata una batteria di test neuropsicologici completa, specifica per l’identificazione di deficit cognitivi in pazienti PD, secondo i criteri pubblicati di ‘Livello II’ (Dubois et al., 2007; Litvan et al., 2012; Marras et al., 2014). Abbiamo quindi applicato tali criteri anche a pazienti MSA e PSP, dato che non esistono criteri pubblicati per i parkinsonismi atipici. Infine, sono state utilizzate analisi statistiche di tipo non-parametrico. Dai nostri risultati emerge che i pazienti con PSP hanno un declino cognitivo più severo rispetto a pazienti PD ed MSA. Nello specifico, al follow-up è stato possibile osservare un marcato declino a carico del dominio esecutivo e del linguaggio nel gruppo con PSP. Le valutazioni cognitive alla baseline e al follow-up erano concordanti, ed entrambe confermano che i pazienti PSP hanno una prestazione peggiore rispetto ai pazienti PD ed MSA: in particolare, nello Stroop test, nelle fluenze verbali (semantica e fonematica) e nel MoCA. Valutando la severità dei deficit cognitivi, abbiamo inoltre trovato diverse percentuali di diagnosi cognitive (i.e., profilo nella norma, MCI vs. demenza) tra i tre gruppi. In particolare, la percentuale più elevata di pazienti con demenza era nel gruppo con PSP rispetto ai pazienti MSA (i.e., 33% vs. nessun paziente con demenza), anche se la durata di malattia era simile. Inoltre, tra i pazienti MSA e PSP con un profilo MCI-multidominio alla baseline, solo pazienti con PSP passano ad una diagnosi di demenza al follow-up. Infine nel gruppo di pazienti PD, nonostante avessero una durata di malattia più lunga, la percentuale di soggetti che passano ad una diagnosi di demenza era inferiore rispetto al gruppo con PSP (7% vs. 16%), nonostante entrambi i gruppi avessero una gravità di MCI simile alla baseline. Complessivamente questi risultati suggeriscono un più rapido e severo declino cognitivo in soggetti PSP, mentre i pazienti MSA mostrano generalmente deficit più limitati. La scala globale MoCA sembra essere maggiormente sensibile, rispetto al MMSE, nel rilevare cambiamenti cognitivi, in particolare nella PSP. Tuttavia il MoCA mostra una sensibilità inferiore rispetto al MMSE nell’identificare un declino cognitivo al follow-up in pazienti PD; quindi il MMSE sembra essere uno strumento migliore per monitorare longitudinalmente cambiamenti cognitivi in pazienti PD. Riguardo al profilo comportamentale, i pazienti PSP riportano più comunemente rispetto ai pazienti PD: apatia, ansia e depressione. Infine, l'analisi dei subitem rivela che i pazienti PSP mostrano un peggioramento ‘clinicamente significativo’ dopo 15 mesi soprattutto nei subitem attentivo-esecutivi (Trial Making Test parte B e il disegno di un orologio). Tuttavia è stato possibile osservare che alcuni pazienti hanno anche un miglioramento in specifici subitem al follow-up. Questo miglioramento potrebbe essere attribuibile ad una più elevata dose farmacologica (nonostante il trattamento dopaminergico alla baseline non fosse significativamente diverso al follow-up). Tuttavia, è importante notare che tali alterazioni erano presenti soprattutto in subitem sensibili alle problematiche motorie (i.e., disegno di figure e collegamento di cerchi con una penna) che quindi potrebbero aver alterato la performance. Questi limiti della scala MoCA e MMSE sono già stati osservati in precedenza nei pazienti con PD (Biundo et al., 2016b; Hu et al., 2014), e possibilmente sono ancora più pronunciati nei parkinsonismi atipici. In conclusione i nostri risultati rivelano che i pazienti PSP hanno una performance notevolmente alterata rispetto agli altri disturbi parkinsoniani (MSA e PD), e dopo circa 6 anni di durata di malattia, il 33% dei pazienti PSP ha una diagnosi di demenza. Questa severa progressione è probabilmente associata ad una diffusione di aggregati tau che coinvolge anche strutture corticali. Al contrario, il pattern di compromissione cognitiva in pazienti con MSA è meno severo, e probabilmente è associato ad una predominanza sottocorticale della patologia, con un coinvolgimento corticale solo secondario alle alterazioni sottocorticali. Pertanto, i nostri risultati suggeriscono che la valutazione neuropsicologica può essere utile nella differenziazione dei profili cognitivi nei parkinsonismi atipici e per monitorare la progressione della malattia. PARTE III – Studi di neuroimmagine sulle sinucleinopatie Capitolo 5: Effetti dei depositi di amiloide sulle manifestazioni cognitive e motorie nella malattia di Parkinson Alterazioni cognitive, in particolare deficit esecutivi, possono essere osservati anche nelle prime fasi del PD (Aarsland, Bronnick, Larsen, Tysnes & Alves, 2009). La disfunzione frontostriatale del sistema dopaminergico può influenzare la presenza di problemi esecutivi ed attentivi (Bruck, Aalto, Nurmi, Bergman, & Rinne, 2005), tuttavia al momento le evidenze relative al trasportatore striatale di dopamina (DAT) sono inconsistenti (Delgado-Alvarado, Gago, Navalpotro-Gomez, Jimenez-Urbieta, & Rodriguez-Oroz, 2016). I meccanismi neuropatologici che stanno alla base delle alterazioni cognitive nei PD sono eterogenei (Irwin, Lee, & Trojanowski, 2013; Kehagia, Barker & Robbins, 2010), ed il contributo del deposito di amiloide in aggiunta alla sinucleinopatia rimane ancora poco definito, soprattutto nelle prime fasi della malattia. Pertanto, lo scopo del presente studio è stato quello di indagare l'interazione tra depositi di amiloide nel circuito frontostriatale, deficit dopaminergico striatale, grado di atrofia cerebrale ed il loro contributo nelle alterazioni cognitive (i.e., funzioni fronto-esecutive) nelle prime fasi del PD. Una coorte multicentrica di 33 pazienti con PD ricavata dal ‘Parkinson's Progression Markers Initiative’ è stata sottoposta a una tomografia ad emissione di positroni (PET) con radiofarmaco [18F]florbetaben, tomografia ad emissione di fotone singolo (SPECT) con radiofarmaco [123I]FP-CIT, risonanza magnetica (MRI) strutturale, valutazione clinica e cognitiva. Dai nostri risultati emerge che elevati livelli di depositi di amiloide erano associati ad una riduzione del deficit dopaminergico nello striato dorsale (rispetto ai bassi livelli di depositi di amiloide), ad un aumento dell’atrofia cerebrale in regioni frontali ed occipitali, e ad una tendenza a manifestare più frequentemente alterazioni cognitive globali (come valutato dal MoCA), ed in test fronto-esecutivi. Inoltre, le deposizioni di amiloide nelle regioni frontostriatali erano inversamente correlate alla performance cognitiva. Nel complesso i nostri risultati suggeriscono che pazienti con PD in fase iniziale di malattia e amiloidosi hanno un più elevato grado di atrofia cerebrale e possono esperire maggiori deficit cognitivi (i.e., disfunzioni esecutive) e alterazioni motorie rispetto a soggetti negativi all’amiloide. I nostri risultati sembrano essere in linea con una recente ipotesi neuropatologica che considera il danno e disfunzione assonale a livello sinaptico come un elemento caratteristico del PD (Tagliaferro & Burke, 2016). Infatti, i neuroni del sistema dopaminergico sono particolarmente vulnerabili alla sinucleinopatia a causa delle loro caratteristiche assonali: gli assoni sono lunghi, sottili e non mielinizzati. Questa ipotesi è confermata anche da studi di neuroimmagine PET con traccianti che si legano al DAT (Caminiti et al., 2017), suggerendo che le aggregazioni di sinucleina nel PD possono influenzare la funzione sinaptica e la trasmissione di segnale sin dalle prime fasi della malattia. I nostri risultati suggeriscono quindi una possibile interazione tra depositi di amiloide e sinucleinopatia, in cui la presenza di amiloide può facilitare la diffusione di sinucleina (i.e., corpi di Lewy) (Toledo et al., 2016), pertanto questa interazione può contribuire ulteriormente alla vulnerabilità assonale. In linea con questa ipotesi, i nostri risultati sembrano confermare che le deposizioni di amiloide agiscono sinergicamente con la sinucleinopatia, influenzando le manifestazioni cliniche del PD. Capitolo 6: Profilo neurostrutturale dell’atrofia multisistemica con alterazioni cognitive A differenza di altre sinucleinopatie (e.g., PD e DLB), la presenza di demenza è considerata un criterio di esclusione nella diagnosi di MSA (Gilman et al., 2008), tuttavia vi è una crescente evidenza che pazienti affetti da MSA possano manifestare alterazioni cognitive, che includono disfunzioni esecutive ma anche deficit cognitivi multidominio, e in alcuni casi anche demenza (Gerstenecker, 2017). Il MMSE è una scala cognitiva globale comunemente utilizzata nella pratica clinica, e recentemente uno studio multicentrico ha suggerito l’utilizzo di un cutoff <27 per aumentare la sensibilità di tale scala nell'identificare alterazioni cognitive in pazienti MSA (Auzou et al., 2015). I meccanismi che sottendono le disfunzioni cognitive in soggetti MSA non sono ancora stati identificati ed evidenze da studi di MRI suggeriscono un discreto contributo corticale e sottocorticale per spiegare tali alterazioni cognitive (Kim et al., 2015; Lee et al., 2016a). Tuttavia questi risultati sono basati su un numero relativamente piccolo di pazienti e in vari stadi di malattia, inoltre sono studi basati su singoli centri. Pertanto, lo scopo del nostro studio multicentrico è stato quello caratterizzare i cambiamenti anatomici associati ad alterazioni cognitive in pazienti MSA e di investigare le differenze strutturali corticali e sottocorticali rispetto ad un campione di soggetti sani. Abbiamo quindi esaminato retrospettivamente 72 pazienti MSA, e definito 50 MSA come cognitivamente normali (MSA-NC) e 22 con alterazioni cognitive (MSA-CI) utilizzando il cutoff del MMSE <27. Abbiamo inoltre confrontato direttamente i due sottogruppi di MSA, e comparato l’intero gruppo di MSA ad un campione di 36 controlli sani (HC) utilizzando un’analisi di ‘morfometria basata sui voxel’ che analizzava la sostanza grigia e bianca. Inoltre, abbiamo applicato anche una segmentazione automatizzata dei volumi sottocorticali. Dai nostri risultati emerge che i pazienti MSA, rispetto a soggetti sani, hanno una diffusa atrofia corticale (i.e., che coinvolge bilateralmente aree frontali, occipito-temporali e parietali), sottocorticale ed alterazioni alla sostanza bianca. Tuttavia, nel confronto diretto, i soggetti MSA-CI mostrano solo una focale riduzione del volume a carico della corteccia prefrontale dorsolaterale sinistra rispetto a pazienti MSA-NC. Tali risultati suggeriscono che la patologia corticale abbia un effetto marginale sulle alterazioni cognitive nei pazienti MSA. Suggeriamo quindi che le alterazioni cognitive siano piuttosto determinate da una degenerazione frontostriatale focale, che sembra essere in linea con il concetto di ‘alterazioni cognitive sottocorticali’.

Nel campo delle reti neurali per il riconoscimento immagini, una delle più recenti e promettenti innovazioni è l’utilizzo delle Capsule Networks (CapsNet). Lo scopo di questo lavoro di tesi è studiare l'approccio CapsNet per l'analisi di immagini, in particolare per quelle neuroanatomiche. Le odierne tecniche di microscopia ottica, infatti, hanno posto sfide significative in termini di analisi dati, per l'elevata quantità di immagini disponibili e per la loro risoluzione sempre più fine. Con l'obiettivo di ottenere informazioni strutturali sulla corteccia cerebrale, nuove proposte di segmentazione possono rivelarsi molto utili. Fino a questo momento, gli approcci più utilizzati in questo campo sono basati sulla Convolutional Neural Network (CNN), architettura che raggiunge le performance migliori rappresentando lo stato dell'arte dei risultati di Deep Learning. Ci proponiamo, con questo studio, di aprire la strada ad un nuovo approccio che possa superare i limiti delle CNNs come, ad esempio, il numero di parametri utilizzati e l'accuratezza del risultato. L’applicazione in neuroscienze delle CapsNets, basate sull’idea di emulare il funzionamento della visione e dell’elaborazione immagini nel cervello umano, concretizza un paradigma di ricerca stimolante volto a superare i limiti della conoscenza della natura e i limiti della natura stessa.

Thesis (Ph. D.)--University of California, San Diego, 2006.
Title from first page of PDF file (viewed July 11, 2006). Available via ProQuest Digital Dissertations. Vita. Includes bibliographical references.

Attenuation and scatter are limiting factors in image quality and quantitation of organ function by single photon emission computed tomography (SPECT). To correct brain images for attenuation an effective water/tissue attenuation coefficient of 0.12 cm$ sp{-1}$ (at 140 keV) or larger has been recommended in order to compensate for the additional bone (skull) attenuation.
It has been determined that the reconstructed images are overcorrected in the centre by 5%, and the optimum correction occurs for a reduced coefficient of 0.09 cm$ sp{-1}$. The overcorrection is due to increased attenuation at the edges of all projections where the path length through the bone is greater, although the bone also increases the scatter at the projection edges.
A correction scheme which uses effective bone and water coefficients was developed to compensate for the attenuation. Alternatively, prior to attenuation correction, a common scatter correction was found to be effective in explicitly removing the bone and water scatter.

Mild traumatic brain injury (TBI) may be complicated by long term cognitive and affective symptoms. Conventional imaging findings often do not correlate with the clinical picture in these patients, and frequently underestimate the extent of damage. Quantitative MR imaging techniques are sensitive to microstructural damage in brain grey matter (GM) and white matter (WM) which appear uninjured on conventional MRI. Previous work has predominantly evaluated their use in acute TBI in moderate and severely injured patients, or in chronic TBI across the severity spectrum. This thesis explored the application of quantitative T1 (qT1) and quantitative T2 (qT2) relaxometry and diffusion tensor imaging (DTI) in the acute evaluation of 44 mild and 9 moderate TBI patients in whom neuropsychological assessment had been performed, and compared the results to those of 30 matched control subjects. By combining the scan data with results from the cognitive testing, this work sought to identify correlations between regions of detectable microstructural damage and the neurocognitive functions related to them. Differences between groups were observed in whole brain normal appearing GM in qT1, and in frontal lobe normal appearing GM and WM in qT1 and DTI measures. Differences were also observed in memory performance and executive function between patients and control subjects which correlated with injury severity. Significant negative correlations were revealed between whole brain WM qT1 time and executive function and negative correlations were shown between frontal and left temporal GM qT1 time and both memory performance and phonemic fluency. Also demonstrated were a positive correlation between frontal GM MD and phonemic fluency, and a negative correlation between frontal GM FA and both memory and executive function. Lastly, increases in WM FA in the corpus callosum, corona radiata, superior longitudinal fasciculus and cingulum were shown to negatively correlate with all components of verbal fluency. This work has demonstrated, using quantitative MR imaging, acute differences at a microstructural level between TBI patients and matched control subjects, in tissue appearing normal on conventional imaging. Furthermore, it has shown that these changes correlate with post-concussive cognitive deficits. It is likely that these changes represent damage as a result of traumatic brain injury in the regions responsible for the cognitive functions found to be impaired.

This thesis considers the registration of shapes, estimation of shape variability and the statistical modelling of human brain magnetic resonance images (MRI). Current shape registration techniques, such as Procrustes analysis, superimpose shapes in order to make inferences regarding the mean shape and shape variability. We apply Procrustes analysis to a subset of the landmarks and give distributional results for the Euclidean distance of a shape from a template. Procrustes analysis is then generalised to minimise a Mahalanobis norm, with respect to a symmetric, positive denite matrix, and the weighted Procrustes estimators for scaling, rotation and translation obtained. This weighted registration criterion is shown, through a simulation study, to reduce the bias and error in maximum likelihood estimates of the mean shape and covariance matrix compared to isotropic Procrustes. A Bayesian Markov chain Monte Carlo algorithm is also presented and shown to be less sensitive to prior information. We consider two MRI data sets in detail. We examine the first data set for large-scale shape dierences between two volunteer groups, healthy controls and schizophrenia patients. The images are registered to a template through modelling the voxel values and we maximise the likelihood over the transformation parameters. Using a suitable labelling and principal components analysis we show schizophrenia patients have less brain asymmetry than healthy controls. The second data set is a sequence of functional MRI scans of an individual's motor cortex taken while they repeatedly press a button. We construct a model with temporal correlations to estimate the trial-to-trial variability in the haemodynamic response using the Expectation-Maximisation algorithm. The response is shown to change with task and through time. For both data sets we compare our techniques with existing software packages and improvements to data pre-processing are suggested. We conclude by discussing potential areas for future research.

Introduction Temporal lobe epilepsy (TLE) is the most common cause of medically intractable partial epilepsy in adults. For many patients, anterior temporal lobe resection (ATLR) is an effective means of treatment, but can cause a significant decline in language or memory function, and visual field deficits. Diffusion tensor imaging and tractography is an MRI technique that can be used to probe white matter structure, and delineate the white matter tracts relevant to vision, language, and memory function. Aims We aimed to use diffusion MRI to increase understanding of the causes and consequences of TLE, and identify patients who are at risk of language, and visual impairment after surgery. Methods and Analysis Techniques Healthy controls, and patients with TLE were scanned pre- and post operatively using 3T MRI. All patients in the study underwent a comprehensive pre- and post-surgical evaluation including clinical, MRI, video-EEG, and neuropsychological assessment. Whole brain analysis of both pre-, and post-operative diffusion MRI was carried out. Tractography was used to assess white matter relevant to memory, language and vision. Correlation analysis of white matter data, and neuropsychological and clinical variables was carried out using the statistical software package, SPSS. Results and Discussion This thesis demonstrates the widespread changes in white matter microstructure present in patients with TLE, and the relationship between medial temporal lobe connections and memory function. It demonstrates how white matter microstructure changes after anterior temporal lobe resection, and how this information can be used to aid prediction of post-operative language deficits in patients. It concludes by showing that tractography can be used to predict postoperative visual field deficits. Conclusion Diffusion Mill can be used to increase our understanding of the causes and consequences of TLE, and to improve pre-surgical planning.

In this thesis electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) are used to investigate the temporal and spatial patterns of noxious-evoked brain activity in newborn infants. EEG was used to investigate responses to graded intensities of experimental noxious stimulation, and evoked brain activity was compared with behavioural and spinal cord activity constituting common surrogate measures of pain in infants. Nociceptive-specific brain activity was elicited in response to all forces of experimental noxious stimulation (applied forces: 32 - 128 mN). In addition, the magnitude of the noxious-evoked response was positively correlated with the magnitude of reflex leg withdrawal, and this relationship was observed in the absence of changes in facial expression. As fMRI had not previously been used to investigate nociceptive processing in infants at 3 Tesla, initial experiments were conducted to optimise the acquisition parameters. The results from optimisation showed that an echo time of approximately 50 ms should be used in future fMRI studies in infants. Experiments conducted alongside this optimisation used fMRI to investigate the cortical and subcortical structures activated by experimental noxious stimulation (applied forces: 32 - 128 mN) in newborn infants (0 - 11 days old). This was compared with noxious-evoked brain activity in adults (applied forces: 32 - 512 mN). Experimental noxious stimulation evoked a widespread pattern of brain activity in newborn infants that overlapped with the network of brain regions activated by nociceptive processing in adults.

Mestrado em Engenharia de Computadores e Telemática
A Imagiologia Cerebral (IC) está na fronteira entre a neurologia, engenharia e física. écnicas de imagens médicas multimodais, tais como a Ressonância Magnética (MRI e fMRI) e Espectroscopia (MRS), Tomografia Computadorizada por Emissão de Fotões/Positrões (SPECT/PET), entre outros, são emergentes ferramentas de pesquisa médica que pode fornecer informações valiosas para o diagnóstico de doenças do cérebro. Eletroencefalograma de alta resolução (HR-EEG), técnicas para sincronizar e fundir seus resultados de análise e várias técnicas de imagem são também parte de IC. Em Portugal, dado o facto que a maioria das áreas relacionadas com IC (por exemplo, medicina, engenharia ou física) são assuntos de investigação em muitos grupos de P&D, um consórcio de universidades de Aveiro, Coimbra, Minho e Porto criou a Rede Nacional de Imagiologia Funcional Cerebral (RNIFC). A RNIFC é uma associação sem fins lucrativos que foi formalizada e assinada em fevereiro de 2009. Actualmente, com o suporte de sistemas digitais para armazenar imagens médicas, é possível partilhar dados entre essas instituições para melhorar o diagnóstico, e permitir investigações entre a comunidade médica de diferentes instituições. O principal objectivo desta dissertação é descrever a implementação dos serviços de sistemas de informação essenciais para a Brain Imaging Network (BIN) que suportam actualmente o RNIFC acessível através do Portal BIN, o principal ponto de entrada para a BING. O Portal BIN permite aos pesquisadores na comunidade BING espalhadas pelo país e no estrangeiro, quer para solicitar o acesso a instrumentos científicos ou para recuperar os seus casos e executar as suas análises. ABSTRACT: Brain Imaging is in the frontier between neurology, engineering and physics. Multimodal medical imaging techniques, such as Magnetic Resonance Imaging (MRI and fMRI) and Spectroscopy (MRS), Single Photon/Positron Emitting Tomography (SPECT/PET) among others, are emergent medical research tools that can provide valuable information for diagnosis of brain diseases. High-resolution electroencephalogram (HR-EEG), techniques for synchronizing and fuse its analysis results and several imaging techniques are also part of BI. In Portugal, given fact that most of the BI related areas (e.g. medical, engineering or physics) are subjects of research in many R&D groups, a consortium of the universities of Aveiro, Coimbra, Minho and Porto created the National Functional Brain Imaging Network (RNIFC). The RNIFC is a non-profitable association that was formalized and signed in February 2009. Currently, with the support of digital systems to store medical images, it is possible to share data among these institutions to improve diagnosis, and allow investigations by the medical community among different institutions. The main objective of this thesis is to describe the implementation of the essential Brain Imaging Network (BIN) information systems services that currently support the RNIFC accessible through the BIN Portal, the main entry point for the BING. BIN Portal enables researchers in the BING community scattered along the country and abroad either to apply for access to the scientific instruments or to retrieve their cases and run their analysis.

Functional magnetic resonance imaging (fMRI) measuring a blood-oxygen-level dependent (BOLD) signal is the most commonly used neuroimaging tool to understand brain function in humans. As mouse models are one of the most commonly used neuroscience experimental models, and with the advent of transgenic mouse models of neurodegenerative pathologies, there has been an increasing push in recent years to apply fMRI techniques to the mouse brain. This thesis focuses on the development and implementation of mouse brain fMRI techniques, in particular to describe the mouse visual system. Multiple studies in the literature have noted several technical challenges in mouse fMRI. In this work I have developed methods which go some way to reducing the impact of these issues, and I record robust and reliable haemodynamic-driven signal responses to visual stimuli in mouse brain regions specific to visual processing. I then developed increasingly complex visual stimuli, approaching the level of complexity used in electrophysiology studies of the mouse visual system, despite the geometric and magnetic field constraints of using a 9.4T pre-clinical MRI scanner. I have also applied a novel technique for measuring high-temporal resolution BOLD responses in the mouse superior colliculus, and I used this data to improve statistical parametric mapping of mouse brain BOLD responses. I also describe the first application of dynamic causal modelling to mouse fMRI data, characterising effective connectivity in the mouse brain visual system. This thesis makes significant contributions to the reverse translation of fMRI to the mouse brain, closing the gap between invasive electrophysiological measurements in the mouse brain and non-invasive fMRI measurements in the human brain.

Electrical brain imaging (EBI) refers to a set of techniques that exploit either the spontaneous electrical activity of the central nervous system, as in electroencephalographic (EEG) source reconstruction, or make use of external current injections, as in electrical impedance tomography (EIT) , to image the structure or function of the brain. When compared to other brain imaging methods used in research or in the clinical setting, such as computed tomography (CT), magnetic resonance imaging (MRI), functional MRI (fMRI), positron emission tomography (PET) and single photon emission computed tomography (SPECT), EIT and EEG source localization instrumentation offer the advantages of portability, low cost, high temporal resolution [ms] and quick setup. The downsides are a low spatial resolution [cm], high computational cost of the image reconstruction process and high sensitivity to imperfections of the electrical model of the head. In this work, a new special purpose reconstruction algorithm for EIT is presented and validated wth experimental measurements performed on a cylindrical phantom and on a simulated human head. The algorithm focuses on the quick detection of compact conductivity contrasts in imperfectly known in 3D domains. The performance of the proposed algorithm is then compared to the one of a benchmark reconstruction method in the EIT field, Tikhonov regularized reconstruction, with stroke detection and classification as a case study. Moreover, the possible application of EIT imaging to the detection of epileptic foci with intracranial deep electrodes (stereoelectroencephalography or SEEG) is explored. Finally, EEG source reconstruction algorithms are implemented on a heterogeneous multi-CPU and multi-GPU computing system to significantly reduce the reconstruction time.

This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Thesis: Ph. D., Harvard-MIT Program in Health Sciences and Technology, 2020
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 171-185).
Magnetic Particle Imaging (MPI) is an emerging tracer-based imaging modality that uniquely images the nonlinear magnetization of superparamagnetic iron oxide nanoparticles (SPIOs). MPI boasts high sensitivity, zero background signal, positive contrast, fast temporal resolution, and quantitative detection. The field of MPI is currently preclinical, and this work aims to scale MPI to human sizes by developing and validating it for two clinical applications: tumor detection and imaging for intraoperative margin assessment during breast-conserving surgery (BCS), and functional neuroimaging. For margin assessment in BCS, a hand-held Magnetic Particle detector and a small-bore MPI imager are assessed for intraoperative use along with an injected SPIO agent. The goal is to detect positive margins during surgery and thus reduce the need for future reexcision. Both hardware systems are validated using clinically relevant phantoms. For functional Magnetic Particle Imaging (fMPI) of the brain, a continuous time-series MPI imager is developed and validated for imaging of cerebral blood volume (CBV) changes during functional activation. The goal is improved sensitivity beyond the capabilities of current functional imaging modalities. We present initial results of in vivo rodent fMPI in a small-bore imager, and the design of a human head-sized system, with implementation underway. Through the collective development of these MPI hardware systems and validation of their potential for these two clinical applications, this work aims to catalyze the expansion of MPI into the clinical setting.
by Erica Ellis Mason.
Ph. D.
Ph.D. Harvard-MIT Program in Health Sciences and Technology

Thesis (MSc (Radiography))--Cape Peninsula University of Technology, 2019
Introduction: Traumatic brain injury (TBI) is frequently associated with mortality and morbidity in low-income countries. Computed Tomography Brain (CTB) imaging aid in the management of patients by accurately exploring primary and secondary brain injuries following trauma. However, there is controversy among researchers on the benefits of follow-up CTB imaging (FCTBI) amongst patients presenting with TBI showing a normal baseline scan. As such, in an attempt to address the contention, the primary focus of this research study was to explore the role of FCTBI with regards to the clinical status of such patients. The secondary focus was to determine the timing of performing FCTBI post TBI. Method: A retrospective cross sectional quantitative design was conducted for this research study. A total sampling strategy was employed on medical records of 85 patients treated at the research site in Zimbabwe. Data were collected over a two year period. Adult patients between the ages of 18 and 75, with TBI and who had a normal first CTBI1 (primary scan done upon hospital admission) were included in this research study. The evolution of different types of brain pathology diagnosed on FCTBI in affected patients were recorded on data collection sheets. An analysis then followed to establish whether the sample patients had developed any neurological complications. Results: The study showed that in 85 patients with TBI, 36% recorded abnormal radiological findings on FCTBI with subdural haematoma (19%) being the most common intracranial lesion followed by intracerebral haemorrhage (8%), subarachnoid haemorrhage (6%) and lastly, pneumocephalus and epidural haematoma (1% respectively). The most frequent causal mechanism of trauma was road traffic accidents (RTAs) at 58%. Males with TBI comprised a higher proportion (53%) than did females (47%). The performance of CTBI1 at 8 hours post trauma occurrence, within a recommended hospital observation period of 20 hours post trauma occurrence, may provide sufficient time for lesions to evolve and thus determine the appropriate patient management. The young adult age group of 26-35 years was found to be more susceptible to TBI. Conclusion: FCTBI was found to be of value in timely detection of evolving intracranial lesions which enabled appropriate management of patients. The current study recommends that patients who exhibit a declining Glasgow Coma Scale (GCS) score and deteriorating neurological status undergo a FCTBI.