Since the development of magnetic resonance imaging (MRI), there have been many novel advances in our understanding of brain structure and function. More recently, functional MRI has revealed networks of spatially isolated brain regions with temporally correlated activity, forming resting state networks. Research has long shown that mindfulness can produce psychological improvements. A new wave of research is demonstrating how mindfulness is associated with alterations in these brain networks. The current thesis examined changes in patterns of functional connectivity associated with scores from a commonly used mindfulness questionnaire in three resting state networks: the default mode network, the central executive network, and the salience network. Independent component analysis data from 32 healthy participants revealed that mindfulness is associated with altered patterns of functional connectivity in all three networks. For example, decreased connectivity was observed in the precuneus in two of the networks, a region associated with mind wandering. This suggests that mindfulness has a physiological influence on the resting state functional connectivity of the brain that coincides with the underlying principles of mindfulness. / May 2017

Clinical care for major depressive disorder (MDD) would be greatly improved if we had reliable clinical predictors of individual antidepressant treatment outcome. While, at the present time, no biomarkers have sufficiently proven utility to be ready for clinical application, several neuroimaging modalities have shown promise for such development. Attempts to combine the recently developed modality of resting-state functional Magnetic Resonance Imaging (rs-fMRI) with pharmacological challenges to explore the impact of antidepressants on resting-state brain connectivity have just begun (McCabe et al., 2011a, McCabe et al., 2011b). The aim of the current study was to investigate the effects of a single dose of the SSRI (selective serotonin reuptake inhibitor) escitalopram on resting-state functional connectivity in health.

Schizophrenia (SCZ) is a complex psychiatric disorder including various symptoms. Resting-state fMRI investigations mostly focused on functional connectivity alterations in SCZ reflecting the spontaneous activity’s spatial structure. Complementing its spatial structure, the brain’s spontaneous activity can be characterized by a complex temporal structure such as scale-free dynamics or long- range temporal correlations (LRTCs). However, it remains an open question whether the temporal structure of spontaneous brain activity, as indexed by the power-law exponent (PLE), can provide biomarkers specific to SCZ as distinguished from other psychiatric disorders like major depressive disorder (MDD) and bipolar disorder (BP). Here, we studied a large-scale cohort (n = 244) of two independent schizophrenic data sets (n = 45), MDD (n = 28), and BP patients (n = 73, in manic, depressed, and euthymic phases) and 98 healthy controls. We found significant PLE reduction in specifically the medial prefrontal cortex (mPFC) in SCZ. This was replicated in an independent sample and was shown to be specific when compared to MDD and different phases of BP. Due to its disease-specific nature, the mPFC PLE reduction may eventually serve as a biomarker for SCZ.

Background: There is growing evidence of abnormalities in neurocognition, neuroanatomy, and functional connectivity in posttraumatic stress disorder (PTSD) and major depressive disorder (MDD). However, there has been less work on individuals who suffer with comorbid PTSD and MDD. It is important to investigate the neurobiology of this overlap because of its prevalence, its associated morbidity, and the hope that it may shed more light on the mechanisms involved in each disorder, including the role of the prefrontal regions. This dissertation tests the hypothesis that women with PTSD and MDD display distinct patterns of neurocognitive impairment and associated brain dysfunction, relative to healthy controls, and these effects will be amplified in patients with both disorders. Methods: This dissertation was undertaken within the Drakenstein Child Health Study, a study exploring child health determinants in mother-infant dyads from the Drakenstein district, Western Cape. Mothers (between 18 and 50 years) were recruited and divided into 4 groups: PTSD, MDD, PTSD with MDD, and healthy controls. Participants were assessed using the computerised NIH Toolbox, and paper and pencil neurocognitive tests. Domains assessed included memory, learning, and processing speed, and with particular focus on executive function and attention domains. Participants underwent resting-state functional imaging as well as structural brain imaging. Functional connectivity within and between cognitive control networks (salience network, dorsal attention network, and frontoparietal networks) and a default mode network were compared across the 4 groups. Neuroanatomical indices (cortical thickness, volume, and surface area) of 10 frontal cortical regions from the Desikan-Killiany atlas in Freesurfer 6 were analysed across the 4 groups. Results: All three clinical groups demonstrated no group differences on measures of attention and executive function, diagnoses of PTSD and MDD were associated with more intrusive thoughts and delayed recall impairment, respectively. However, neurocognitive findings indicate that PTSD with comorbid MDD is not associated with greater neurocognitive dysfunction relative to mono-diagnostic groups. Abnormal resting-state connectivity was observed for the MDD group in the default mode network, and for both comorbid and MDD patient groups within frontoparietal networks. Abnormal salience network connectivity for the comorbid group was observed when examining performance on the Pattern Comparison Processing Speed test. No between-network connectivity group differences were observed. Surface area and volume reductions of prefrontal regions were evident for PTSD and MDD, however, no volumetric and surface area differences were observed for the comorbid group. Conclusion: In this sample of mothers from a low-middle income region, distinct patterns of neurocognitive dysfunction and impairment in PTSD, MDD, and PTSD with MDD were observed. However, contrary to hypotheses, comorbidity is not associated with greater dysfunction and impairment and the associations of PTSD and comorbid MDD are not amplified in this sample. These findings have implications for the development of treatment plans for patients diagnosed with PTSD, MDD, and PTSD with comorbid MDD, so that interventions are tailored in a way that is responsive to differences between these groups in the presentation of neurocognitive profile, brain function, and structure.

Indiana University-Purdue University Indianapolis (IUPUI) / Negative urgency, defined as a tendency to act rashly under extreme negative emotion, is strongly associated with tobacco use. Despite the robust cross-sectional and experimental evidence linking negative urgency and tobacco use, neural correlates of negative urgency in tobacco use have not been studied. The purpose of the current study was to 1) identify neural circuits that differ between tobacco users and non-tobacco users and 2) explore the relationship between resting-state seed-based functional connectivity (rsFC) and negative urgency, both in the overall group and between tobacco users and non-tobacco users. Using negative urgency-related brain regions as seed regions (voxel level p = .005, cluster-level a < .05), compared to non-tobacco users (n = 21; mean age = 36.57, 62% female, 76% white), tobacco users (n = 22; mean age = 37.50, 64% female, 77% white) had stronger rsFC strengths in the right amygdala – left medial orbitofrontal cortex/ventromedial prefrontal cortex circuit and the right nucleus accumbens – right temporoparietal junction circuit. Additionally, rsFC in the bilateral temporal pole – left supramarginal gyrus circuits was positively correlated with negative urgency (Left temporal pole: r = .55, p < .001; Right temporal pole: r = .51, p < .001). The current study extends previous neuroimaging findings, which have mainly focused on how negative urgency is related to brain responses in localized, segregated brain regions, by examining the network-level interactions between different brain regions. This study provides prime preliminary data for future neuroimaging studies of negative urgency by providing potential target networks that would aid the development of novel intervention strategies for negative urgency-based maladaptive behaviors.

Complex network models of functional connectivity have emerged as a paradigm shift in brain mapping over the past decade. Despite significant attention within the neuroimaging and cognitive neuroscience communities, these approaches have hitherto not been extensively explored in neurosurgery. The aim of this thesis is to investigate how the field of connectomics can contribute to understanding the effects of focal brain lesions and to functional brain mapping in neurosurgery. This datasets for this thesis include a clinical population with focal brain tumours and a cohort focused on healthy adolescent brain development. Multiple network analyses of increasing complexity are performed based upon resting state functional MRI. In patients with focal brain tumours, the full complement of resting state networks were apparent, while also suggesting putative patterns of network plasticity. Connectome analysis was able to identify potential signatures of node robustness and connections at risk that could be used to individually plan surgery. Focal lesions induced the formation of new hubs while down regulating previously established hubs. Overall these data are consistent with a dynamic rather than a static response to the presence of focal lesions. Adolescent brain development demonstrated discrete dynamics with distinct gender specific and age-gender interactions. Network architecture also became more robust, particularly to random removal of nodes and edges. Overall these data provide evidence for the early vulnerability rather than enhanced plasticity of brain networks. In summary, this thesis presents a combined analysis of pathological and healthy development datasets focused on understanding the functional effects of focal brain lesions at a network level. The coda serves as an introduction to a forthcoming study, known as Connectomics and Electrical Stimulation for Augmenting Resection (CAESAR), which is an evolution of the results and methods herein.

Background: Parkinson's disease (PD) is a chronic, progressive, neurodegenerative disorder with currently no known cure. PD has a significant impact on quality of life of the patients, as well as, the caregivers and family members. It is the second most common cause of chronic neurological disability in US and Europe. According to National Parkinson's Foundation, there are almost 1 million patients in the Unites States and 50,000 to 60,000 new cases of PD are diagnosed each year. The total number of cases of PD is predicted to double by 2030. The annual cost associated with this disease is estimated to be $10.8 billion in the United States, including the cost of treatment and the cost of the disability. Although it is primarily thought of as a movement-disorder and is clinically diagnosed based on motor symptoms, non-motor symptoms such as cognitive and emotional deficits are thought to precede the clinical diagnosis by almost 20 years. By the time of clinical diagnosis, there is 80% loss in the dopamine content in the striatum and 50% degeneration of the substantia nigra dopamine cells. The research presented in this thesis was an attempt to develop an animal model of PD in its pre-motor stages. Such a model would allow us to develop pre-clinical markers for PD, and facilitate the development and testing of potential treatment strategies for the non-motor symptoms of the disorder. Specific Aims: There were five specific aims for this research: * The first specific aim dealt with development of a rat model of PD with slow, progressive onset of motor deficits, determination of timeline for future studies, and quantification the dopamine depletion in this model at a pre-motor stage. * The second and the third specific aims focused on testing for emotional (aversion) deficits and cognitive (executive functioning) deficits in this rat model at the 3 week timepoint determined during specific aim 1. * The fourth specific aim was to determine the brain network changes associated with the behavioral changes observed our rat model using resting state connectivity as a measure. * The fifth and the final specific aim was to test sodium butyrate, a drug from the histone deacetylase inhibitor family, as a potential treatment option for cognitive deficits in PD. Results: The 6-hydroxy dopamine based stepwise striatal lesion model of pre-motor PD, developed during this research, exhibits delayed onset of Parkinsonian gait like symptoms by week 4 after the lesions. At 3 weeks post lesion (3WKPD), the rats exhibit 27% reduction in striatal dopamine and 23%reduction in substantia nigra dopamine cells, with lack of any apparent motor deficits. The 3WKPD rats also exhibited changes in aversion. The fMRI study with the aversive scent pointed towards possible amygdala dysfunction sub-serving the aversion deficits. The executive function deficits tested using a rat analog of the Wisconsin card sorting test, divulged an extra-dimensional set shifting deficit in the 3WKPD rats similar to those reported in PD patients. The resting state connectivity study indicated significant changes in the 3WKPD rats compared to age matched controls. We observed increased overall connectivity of the motor cortex and increased CPu connectivity with prefrontal cortex, cingulate cortex, and hypothalamus in the 3WKPD rats compared to the controls. These observations parallel the observations in unmedicated early-stage PD patients. We also observed negative correlation between amygdala and prefrontal cortex as reported in humans. This negative correlation was lost in 3WKPD rats. Sodium butyrate treatment, tested in the cognitive deficit study, was able to ameliorate the extra-dimensional set shifting deficit observed in this model. This treatment also improved the attentional set formation. Conclusion: Taken together, our observations indicate that, the model of pre-motor stage PD developed during this research is a very high face validity rat model of late Braak stage 2 or early Braak stage 3 PD. Sodium butyrate was able to alleviate the cognitive deficits observed in our rat model. Hence, along with the prior reports of anti-depressant and neuroprotective effects of this drug, our results point towards a possible treatment strategy for the non-motor deficits of PD.

Older bilingual adults typically perform better than monolinguals in tasks of executive control, and are diagnosed later with dementia. Studies have also shown structural and functional brain differences between bilinguals and monolinguals. However, it remains poorly understood how language history influences the functional organization of the aging brain. The current study investigated; 1) differences in resting-state functional connectivity between monolinguals and bilinguals in the Default Mode Network (DMN), Frontoparietal Network (FPN), Executive Control Network (ECN), Language Network (LANG), and a network consisting of structures associated with tasks of executive control coined the Bilingual Control Network (BCN); 2) the relationship of cognitive performance with functional connectivity of the BCN; and 3) whether proficiency, age of second language acquisition, degree of second language exposure, and frequency of language use predicts the network’s functional connectivity. Healthy older bilinguals (N=10) were matched pairwise for age, sex and education to healthy older monolinguals (N=10). All participants completed a battery of cognitive tests, a language history questionnaire, and a 6-minute functional scan during rest. Results showed that groups did not differ in cognitive performance, or in the functional connectivity of the FPN, ECN, LANG, or BCN. However, monolinguals had significantly stronger functional connectivity in the DMN compared to bilinguals. Later age of second language acquisition and lower proficiency were also associated with greater DMN functional connectivity. None of these variables predicted BCN’s functional connectivity. However, bilinguals showed stronger functional connectivity with other structures outside of the canonical networks compared to monolinguals. Finally, vocabulary scores, local switch cost accuracy and reaction time were negatively correlated with BCN’s functional connectivity. Overall, these findings illustrate differences in functional brain organization associated with language experience in the DMN, while challenging the “bilingual advantage” hypothesis. The results also suggest a possible neural mechanism by which bilingualism might mediate cognitive reserve.

This thesis gives an overview on how to estimate changes in functional brain networks using graph theoretical measures. It explains the assessment and definition of functional brain networks derived from fMRI data. More explicitly, this thesis provides examples and newly developed methods on the measurement and visualization of changes due to pathology, external electrical stimulation or ongoing internal thought processes. These changes can occur on long as well as on short time scales and might be a key to understanding brain pathologies and their development. Furthermore, this thesis describes new methods to investigate and visualize these changes on both time scales and provides a more complete picture of the brain as a dynamic and constantly changing network.

In this study, we used resting-state functional magnetic resonance imaging (fMRI) scans from subjects with early mild cognitive impairment (EMCI) and control subjects to study functional network connectivity. The scans were acquired by the Alzheimer's Disease Neuroscience Initiative (ADNI). We used genetic data from the ADNI database to further subdivide the EMCI and control groups into genotype groups with or without the ApoE4 allele. ROI-to-ROI resting-state functional connectivity was measured using Freesurfer and the Functional Connectivity Toolbox for Matlab (CONN). In our analysis, we compared whole-brain ROI connectivity strength and ROI-to-ROI functional network connectivity strength between EMCI, control and genotype subject groups. We found that the ROI network properties were disrupted in EMCI and ApoE4-containing groups. Notably, we show that (1) EMCI disrupts functional connectivity strength in many areas; (2) the ApoE4 allele disrupts functional connectivity strength in similar areas to EMCI; and (3) the differences in functional connectivity between groups shows a multifactor contribution to functional network dysfunction along the trajectory leading to dementia.