Ceren Tozlu

Background: Pregnancy is a distinct neuroendocrine and immunological state in women with multiple sclerosis (MS). Relapse activity typically decreases during gestation and increases after delivery, but the relationship between peri-pregnancy brain volumetric measures and disability remains unclear. Methods: This retrospective longitudinal pilot study included 14 women with relapsing-remitting MS who underwent MRI examinations before pregnancy and after delivery. Thirty-eight MRI scans were analyzed, with a mean of 2.71 ± 0.99 timepoints per patient. Global and regional brain volumes were measured using the longitudinal FreeSurfer pipeline. Choroid plexus (CP) volume was evaluated as an exploratory neuroimmune imaging marker. Associations between volumetric measures and Expanded Disability Status Scale (EDSS) scores were assessed using linear mixed-effects modeling with forward stepwise selection. Results: The final EDSS model retained bilateral ventral diencephalon (ventral DC) volumes, left amygdala volume, right pallidum volume, left nucleus accumbens volume, and cerebral cortex volume. The most prominent association involved the left ventral DC, where lower volume was associated with higher EDSS scores (p = 0.0008). CP volume showed no significant longitudinal change across peri-pregnancy MRI timepoints (p = 0.21) and was not retained in the final EDSS model. Conclusion: Regional brain volumetric measures, particularly ventral DC volumes, were associated with disability across the peri-pregnancy period despite low EDSS levels. To our knowledge, this is the first longitudinal evaluation of CP volume across the peri-pregnancy period in women with MS; no significant change was observed over time. These exploratory findings require confirmation in larger prospective studies using standardized serial MRI protocols.

Traumatic brain injury is a major cause of long‐term cognitive impairment, yet the mechanisms underlying recovery remain poorly understood. Neuroimaging methods such as diffusion magnetic resonance imaging (MRI), functional MRI (fMRI), and positron emission tomography (PET) provide insight into micro‐ and macro‐scale changes post‐traumatic brain injury (TBI), but the relationships between regional cellular and functional alterations remain unclear. In this exploratory study, we conducted a longitudinal, multimodal neuroimaging analysis quantifying TBI‐related pathologies in four biomarkers, namely flumazenil PET derived binding potential, diffusion MRI (dMRI)‐derived structural connectivity, and resting‐state fMRI‐derived functional connectivity and fractional amplitude of low‐frequency fluctuations in individuals with complicated mild‐to‐severe brain injury at the subacute (4–6 months post‐injury) and chronic (1‐year post‐injury) stages. The TBI sample consisted of 41 fMRI, 40 dMRI, and nine PET subjects, with 16 fMRI and dMRI and seven PET longitudinal measurements. The control sample consisted of 14 dMRI and fMRI and 19 PET subjects scanned at a single time point for comparison with TBI subjects at both time points. Most of the PET and MRI subjects are overlapping in both TBI and control groups. Brain injury related regional pathologies, and their changes over time in TBI subjects, were correlated across the four biomarkers. Our results reveal complex, dynamic changes over time. We found that flumazenil‐PET binding potential was significantly reduced in frontal and thalamic regions in brain‐injured subjects, consistent with neural loss and dysfunction, with partial recovery over time. Functional hyperconnectivity was observed in brain injured subjects initially but declined while remaining elevated compared to non‐injured controls, whereas cortical structural hypoconnectivity persisted. Importantly, we observed that brain injury‐related alterations across MRI modalities became more strongly correlated with flumazenil‐PET at the chronic stage. Regions with chronic reductions in flumazenil‐PET binding also showed weaker structural node strength and lower amplitude of low‐frequency fluctuations, a relationship that was not found at the subacute stage. This observation could suggest a progressive convergence of structural and functional disruptions with neuronal dysfunction and loss over time. Additionally, regions with declining structural node strength also exhibited decreases in functional node strength, while these same regions showed increased amplitude of low‐frequency fluctuations over time. This pattern suggests that heightened intrinsic regional activity may serve as a compensatory mechanism in regions increasingly disconnected due to progressive axonal degradation. Altogether, these findings advance our understanding of how multimodal neuroimaging captures the evolving interplay between neuronal integrity, structural connectivity, and functional dynamics after brain injury. Given the exploratory nature of this study, stemming from the modest sample size, future work in larger cohorts will be essential to validate and refine these preliminary associations as well as the inclusion of multiple measures of healthy controls. Clarifying these interrelationships could inform prognostic models and enhance knowledge of degenerative, compensatory, and recovery mechanisms in traumatic brain injury.

Classification et modélisation statistique intégrant des données cliniques et d’imagerie par résonance magnétique conventionnelle et avancée: L'accident vasculaire cérébral et la sclérose en plaques figurent parmi les maladies neurologiques les plus destructrices du système nerveux central. L'accident vasculaire cérébral est la deuxième cause de décès et la principale cause de handicap chez l'adulte dans le monde alors que la sclérose en plaques est la maladie neurologique non traumatique la plus fréquente chez l'adulte jeune. L'imagerie par résonance magnétique est un outil important pour distinguer le tissu cérébral sain du tissu pathologique à des fins de diagnostic, de suivi de la maladie, et de prise de décision pour un traitement personnalisé des patients atteints d'accident vasculaire cérébral ou de sclérose en plaques. La prédiction de l'évolution individuelle de la maladie chez les patients atteints d'accident vasculaire cérébral ou de sclérose en plaques constitue un défi pour les cliniciens avant de donner un traitement individuel approprié. Cette prédiction est possible avec des approches statistiques appropriées basées sur des informations cliniques et d'imagerie. Toutefois, l'étiologie, la physiopathologie, les symptômes et l'évolution dans l'accident vasculaire cérébral et la sclérose en plaques sont très différents. Par conséquent, dans cette thèse, les méthodes statistiques utilisées pour ces deux maladies neurologiques sont différentes. Le premier objectif était l'identification du tissu à risque d'infarctus chez les patients atteints d'accident vasculaire cérébral. Pour cet objectif, les méthodes de classification (dont les méthodes de machine learning) ont été utilisées sur des données d'imagerie mesurées à l'admission pour prédire le risque d'infarctus à un mois. Les performances des méthodes de classification ont été ensuite comparées dans un contexte d'identification de tissu à haut risque d'infarctus à partir de données humaines codées voxel par voxel. Le deuxième objectif était de regrouper les patients atteints de sclérose en plaques avec une méthode non supervisée basée sur des trajectoires individuelles cliniques et d'imagerie tracées sur cinq ans. Les groupes de trajectoires aideraient à identifier les patients menacés d'importantes progressions et donc à leur donner des médicaments plus efficaces. Le troisième et dernier objectif de la thèse était de développer un modèle prédictif pour l'évolution du handicap individuel des patients atteints de sclérose en plaques sur la base de données démographiques, cliniques et d'imagerie obtenues a l'inclusion. L'hétérogénéité des évolutions du handicap chez les patients atteints de sclérose en plaques est un important défi pour les cliniciens qui cherchent à prévoir l'évolution individuelle du handicap. Le modèle mixte linéaire à classes latentes a été utilisé donc pour prendre en compte la variabilité individuelle et la variabilité inobservée entre sous-groupes de sclérose en plaques

The energetic and entropic organization of the brain's functional activity in mild cognitive impairment (MCI) has yet to be fully characterized. Network Control Theory (NCT) is a multi‐modal approach that captures alterations in the brain's energetic landscape by combining the brain's functional activity and the structural connectome. Entropy is another complementary metric that can quantify the complexity and predictability in a neural time series, offering insights into the brain's dynamic functional activity. Our study aims to explore the differences in the brain's energetic and entropic landscape between people with MCI and healthy controls (HC). Four hundred ninety‐nine HC and 55 MCI patients were included. First, k ‐means clustering was applied to functional MRI (fMRI) time series to identify commonly recurring brain activity states. Second, NCT was used to calculate the minimum energy required to transition between these brain activity states, otherwise known as transition energy (TE). The entropy of the fMRI time series as well as PET‐derived amyloid beta (Aβ) and tau deposition were measured for each brain region. The TE and entropy were compared between MCI and HC at the network, regional, and global levels using linear models where age, sex, and intracranial volume were added as covariates. The association of TE and entropy with Aβ and tau deposition was investigated in MCI patients using linear models where age, sex, and intracranial volume were controlled. Commonly recurring brain activity states included those with high (+) and low (‐) amplitude activity in visual (+/‐), default mode (+/‐), and dorsal attention (+/‐) networks. Compared to HC, MCI patients required lower transition energy in the limbic network (adjusted p = 0.028). Decreased global entropy was observed in MCI patients compared to HC ( p = 7.29e‐7). There was a positive association between TE and entropy in the frontoparietal network ( p = 7.03e‐3). Increased global Aβ was associated with higher global entropy in MCI patients ( ρ = 0.632, p = 0.041). Lower TE in the limbic network in MCI patients may indicate either neurodegeneration‐related neural loss and atrophy or a potential functional upregulation mechanism in this early stage of cognitive impairment. Future studies that include people with Alzheimer's Disease (AD) are needed to better characterize the changes in the energetic landscape in the later stages of cognitive impairment.

BACKGROUND AND OBJECTIVES: Parkinson disease (PD) patients with motor complications are often considered for deep brain stimulation (DBS) surgery. Predicting symptom improvement to separate DBS responders and nonresponders remains an unmet need. Currently, DBS candidacy is evaluated using the levodopa challenge test (LCT) to confirm dopamine responsiveness and diagnosis. However, prediction of DBS success by measuring presurgical symptom improvement associated with levodopa dosage changes is highly problematic. Quantitative susceptibility mapping (QSM) is a recently developed MRI method that depicts brain iron distribution. As the substantia nigra and subthalamic nuclei are well visualized, QSM has been used in presurgical planning of DBS. Spatial features resulting from iron distribution in these nuclei have been previously linked with disease progression and motor symptom severity. Given its clear target depiction and prior findings regarding susceptibility and PD, this study demonstrates the technical feasibility of predicting DBS outcomes from presurgical QSM. METHODS: A novel presurgical QSM radiomics approach using a regression model is presented to predict DBS outcome according to spatial features in QSM deep gray nuclei. To overcome limited and noisy training data, data augmentation using label noise injection or “compensation” was used to improve outcome prediction of the regression model. The QSM radiomics model was evaluated on 67 patients with PD who underwent DBS at 2 medical centers. RESULTS: The QSM radiomics model predicted DBS improvement in the Unified Parkinson Disease Rating Scale at Center 1 and Center 2 with Pearson correlation , ( ) and , ( ), respectively. LCT failed to predict DBS improvement at Center 1 and Center 2 with Pearson correlation ( ) and ( ), respectively. CONCLUSION: QSM radiomics has potential to accurately predict DBS outcome in treating patients with PD, offering a valuable alternative to the time-consuming and low-accuracy LCT.

Cocaine use disorder (CUD) detrimentally impacts personal health, social relationships, and economic opportunity. Here, we assess CUD-associated shifts in brain dynamics using Network Control Theory and examine how they align with previously identified changes in neurological systems and behavioral profiles of people with CUD. The SUDMEX CONN dataset consists of multi-modal MRI, cocaine use metrics, behavioral measures, and demographics of individuals with CUD (N=132, 71 CUD). We identified recurring brain activity states and used NCT to calculate the transition energy (TE) between pairs of states. ANCOVAs examined global and regional TE associations with drug use group (CUD vs controls (NC)), years of CUD, and risk-taking behaviors. We identified potential mechanisms driving the differences by correlating CUD-related regional TE effects with neurotransmitter/receptor systems. People with CUD had significantly lower global TE and default mode, dorsal attention and limbic network TE compared to non-user controls, particularly in regions enriched for noradrenaline and mu opioid receptors. Longer duration of CUD was associated with more decreased global TE, top-down TE, default mode, control and ventral attention network TE, and regional TE enriched for excitatory neurotransmitters and receptors. People with CUD needed to expend more global and top-down TE to perform better on a risk-taking task (the Iowa Gambling task), an effect which was not found in NCs. Our analysis of whole-brain activity dynamics provides a link between the effects of upstream glutamatergic excitotoxicity and/or opioid receptor dysfunction, and downstream weakening of inhibitory control that is central to CUD.

Introduction: Patients with mild cognitive impairment (MCI) have shown disruptions in both brain structure and function, often studied separately. However, understanding the relationship between brain structure and function can provide valuable insights into this early stage of cognitive decline for better treatment strategies to avoid its progression. Network Control Theory (NCT) is a multi-modal approach that captures the alterations in the brain’s energetic landscape by combining the brain’s functional activity and the structural connectome. Our study aims to explore the differences in the brain’s energetic landscape between people with MCI and healthy controls (HC). Methods: Four hundred ninety-nine HC and 55 MCI patients were included. First, k-means was applied to functional MRI (fMRI) time series to identify commonly recurring brain activity states. Second, NCT was used to calculate the minimum energy required to transition between these brain activity states, otherwise known as transition energy (TE). The entropy of the fMRI time series as well as PET-derived amyloid beta (Aβ) and tau deposition were measured for each brain region. The TE and entropy were compared between MCI and HC at the network, regional, and global levels using linear models where age, sex, and intracranial volume were added as covariates. The association of TE and entropy with Aβ and tau deposition was investigated in MCI patients using linear models where age, sex, and intracranial volume were controlled. Results: Commonly recurring brain activity states included those with high and low amplitude activity in visual (+/−), default mode (+/−), and dorsal attention (+/−) networks. Compared to HC, MCI patients required lower transition energy in the limbic network (adjusted p = 0.028). Decreased global entropy was observed in MCI patients compared to HC ( p = 7.29e-7). There was a positive association between TE and entropy in the frontoparietal network ( p = 7.03e-3). Increased global Aβ was associated with higher global entropy in MCI patients (ρ = 0.632, p = 0.041). Conclusion: Lower TE in the limbic network in MCI patients may indicate either neurodegeneration-related neural loss and atrophy or a potential functional upregulation mechanism in this early stage of cognitive impairment. Future studies that include people with AD are needed to better characterize the changes in the energetic landscape in the later stages of cognitive impairment.

Quantitative susceptibility maps from magnetic resonance images can provide both prognostic and diagnostic information in multiple sclerosis, a neurodegenerative disease characterized by the formation of lesions in white matter brain tissue. In particular, susceptibility maps provide adequate contrast to distinguish between "rim" lesions, surrounded by deposited paramagnetic iron, and "non-rim" lesion types. These paramagnetic rim lesions (PRLs) are an emerging biomarker in multiple sclerosis. Much effort has been devoted to both detection and segmentation of such lesions to monitor longitudinal change. As paramagnetic rim lesions are rare, addressing this problem requires confronting the class imbalance between rim and non-rim lesions. We produce synthetic quantitative susceptibility maps of paramagnetic rim lesions and show that inclusion of such synthetic data improves classifier performance and provide a multi-channel extension to generate accompanying contrasts and probabilistic segmentation maps. We exploit the projection capability of our trained generative network to demonstrate a novel denoising approach that allows us to train on ambiguous rim cases and substantially increase the minority class. We show that both synthetic lesion synthesis and our proposed rim lesion label denoising method best approximate the unseen rim lesion distribution and improve detection in a clinically interpretable manner. We release our code and generated data at https://github.com/agr78/PRLx-GAN upon publication.

Our prior work demonstrated that estimated structural and functional connectomes (eSC and eFC) generated using multiple sclerosis (MS) lesion masks and artificial intelligence (AI) models can predict disability as effectively as SC and FC derived from diffusion and functional MRI in MS. The goal of this study was to assess the ability of eSC and eFC in predicting baseline and 4-year follow-up cognition in MS patients. The Network Modification tool was performed to estimate SC from the clinical MRI-derived lesion masks. The eSC was then used as an input to Krakencoder, an encoder-decoder model, to estimate FC. The highest accuracy was obtained when predicting the follow-up Symbol Digit Modalities Test (SDMT) using regional eSC or eFC with a median Spearman's correlation of 0.58 for eSC and 0.56 for eFC, which is higher or similar to other studies that predicted cognition in healthy and diseased cohorts. Decreased eSC and eFC in the cerebellum and increased eFC in the default mode network were associated with lower follow-up SDMT scores. Our findings demonstrate that eSC and eFC derived from clinically acquired MRI and AI models can effectively predict cognition. The use of lesion-based estimates of connectome disruptions may potentially improve cognition-related individualized treatment planning.