Bharat B. Biswal, Ph.D. - US grants

DESCRIPTION (provided by applicant): The enormous potential of fMRI as a tool for understanding the influences of TBI (Traumatic Brain Injury) on cognitive and motor functioning in humans has begun to be realized as a body of neurocognitive TBI literature has started to emerge (McAllister et al., 1999;2001, Christodoulou et al., 2001). While the consistency in these findings has been encouraging, the validity of conclusions regarding the influences of TBI on brain activation patterns measured by fMRI will be dependent upon better characterizing the wide variability in brain activation patterns observed (Hillary, Steffener, Biswal et al., 2002). The aim of our proposal is to systematically examine and quantify the vascular and neural factors contributing to trauma- related changes in the fMRI signal. To the best of our knowledge, this is the first attempt to comprehensively examine the biophysical, neural, and cognitive contributions to TBI-related differences in fMRI activation. The techniques we propose to develop and utilize here can be implemented by other investigators to accurately isolate hemodynamic changes due to trauma-related (or, indeed, any population-related) differences in neural activity. Because of this, the current proposal provides the opportunity to establish new standards for applying BOLD fMRI to clinical samples.. The present project intends to combine basic science, engineering, and computational issues to specifically elucidate mechanisms (neuronal vs vascular) that results in TBI subjects having altered brain activation in comparison to healthy controls. Results obtained from the noninvasive technique (fMRI) would provide ways to measure a number of relevant physiological factors and characterize them biophysically to understand human brain function with TBI. Methods and techniques developed can also be used to study between two or more different groups.

DESCRIPTION (provided by applicant): In the recent years, a number of studies using functional MRI (fMRI) have shown substantial differences between the activation pattern of older subjects (> 50 years of age) and younger subjects (21-40 years) while performing a number of different sensorimotor and cognitive tasks. It has been concluded that the contrast observed is due to differences in neuronal activity in the older subjects. However, early hypothesis that normal aging involves widespread loss of neurons have been revised in light of accumulating evidence that in most regions of the brain, the number of neurons is stable throughout adulthood and senescence. In addition to direct effects on neuronal function, factors contributing to cerebrovascular reactivity is known to be altered in older people that could give rise to altered hemodynamic responses. Since the signal observed using fMRI could be modulated both by hemodynamics and oxygenation changes resulting from neuronal changes, these two factors must be separated to gain a better understanding about age related changes in the activation pattern obtained using fMRI. The present project proposed intends to combine basic science, engineering, and computational issues to specifically elucidate mechanisms (neuronal vs vascular) that results in older subjects having altered brain activation in comparison to young subjects. Results obtained from the noninvasive technique (fMRI) would provide ways to measure a number of relevant physiological factors and characterize them biophysically to understand human brain function with aging. Methods and techniques developed can also be used to study between two or more different groups. PUBLIC HEALTH RELEVANCE: The present project will help determine biophysical aspects of aging using non-invasive functional Magnetic Resonance Imaging (fMRI). As longitudinal studies are very important to follow individuals through different stages of their life span, fMRI techniques would become crucial in obtaining valuable biomarkers in studies of aging. FMRI presents many caveats in determining the actual physiological indicators that influence signal response in young and old subjects. This project is designed to address certain caveats by effectively testing and quantifying the neural and hemodynamic components that may modulate signal response in young and old subjects. This study will significantly gain information regarding the underlying nature and necessary corrections in fMRI signals. Such a correction is necessary to accurately determine the progression and determinants of change across all segments of the life span that affect cognitive effects and brain function.

DESCRIPTION (provided by applicant): Once a distant goal, discovery science for human brain function is now a reality. Capitalizing on the ease of data-sharing with resting state fMRI (R-fMRI), the 1000 Functional Connectomes Project (FCP) has invigorated the neuroimaging community by aggregating datasets independently collected by labs around the world, and making them publicly available without restriction. As of Dec 11, 2009, researchers who once struggled to obtain 20 - 30 datasets for analyses suddenly had access to over 1200 datasets. Most importantly, feasibility analyses performed by the founding members of the FCP demonstrated the ability to carry out discovery science with the aggregate dataset. In order for discovery science to take hold in the neuroimaging field, researchers need continued access to large-scale imaging datasets that will enable both data mining and replication studies. To address this need, the FCP launched the International Neuroimaging Data-sharing Initiative (INDI). Designed to encourage the open sharing of more detailed phenotypic data with imaging datasets, and to establish a model for sharing data prospectively (i.e., prepublication), INDI is making significant strides in this regard. The present proposal seeks to extend this impact by overcoming a second key obstacle faced by prospective researchers - namely, scientists need access to appropriate tools to facilitate data exploration. This applies particularly to investigators who are inexperienced with the nuances of fMRI image analysis, or lack the programming support or resources necessary for handling and analyzing large- scale datasets. The aim of the proposed work is to create an open-source user interface for flexible, automated processing of R-fMRI datasets by both novice and expert users. An additional aim is to provide benchmark results so that users can calibrate their local results. Attaining these aims within the 24 months of the project will substantially accelerate the trajectory of discovery science of human brain function. ! PUBLIC HEALTH RELEVANCE: The 1000 Functional Connectomes Project (FCP) has invigorated the neuroimaging community by aggregating datasets from around the world and making them publicly available for researchers to use without restriction. By promoting a culture of open data-sharing, the FCP and similar efforts can rapidly accelerate the pace of neuroscientific and psychiatric discovery, though many statisticians, mathematicians, engineers and neuroscientists are deterred from participating due to limited experience with neuroimaging analysis. The present proposal aims to overcome this obstacle by creating a software tool capable of automated processing of datasets in the FCP by both novice and expert users.

DESCRIPTION (provided by applicant): The brain is the major organ of humans and animals for their interaction with nature. Despite huge research efforts in the field of neurosciences a thorough understanding of the principles and dynamics of the nervous system is still in its infancies. Computational neuroscience is one of the key methodologies for a better understanding of brain function. However, explore and model aspects of brain function and the interplay of neurons and brain regions, as well as to check the validity of models specific boundary conditions evolving from in vivo experimental data and their analysis are needed. A powerful method to gain in vivo functional-metabolic information is non-invasive imaging, specifically represented by the recent advances in combined positron emission tomography and magnetic resonance imaging (PET/MRI), which reveals multiple temporal linked in vivo parameters. Thus, PET/MRI and computational neuroscience complement each other in a perfect way. In this proposal two world leading institutions in the fields of multimodality imaging (University of Tuebingen) and brain connectivity mapping (New Jersey Institute of Technology, NIJT) join forces to explore so far uncharted terrains of metabolic brain connectivity. Many questions regarding large scale networks in the brain, which are even active during resting conditions, are so far unanswered. Their exact origin, interpretation and also their demand on energy consumption are so far not understood. In the last three years, resting state, functional connectivity (RSFC) functional magnetic resonance imaging (fMRI) often also termed functional connectivity fMRI (fc-fMRI) has seen a tremendous increase in interest in applications ranging from basic brain imaging to clinical applications in brain surgery planning or neurodegenerative disease. fcMRI was first developed by the PI from the USA and colleagues who observed that fluctuations in fMRI signals during behavioral rest are temporally correlated within functionally related cortical networks such as motor cortex but not between functionally unrelated networks. Despite its widespread use of this technology its physiological and metabolic basics are not understood. Recently the PI from Germany and colleagues also found that useful information about brain connectivity appears to be hidden in dynamic as well as static positron emission tomography (PET) data. Combined PET/MR imaging in small animals, offers the ability to investigate these metabolic and neurophysiological basics of brain connectivity. One strength of this technology is that PET and fMRI data can be acquired simultaneously, therefore minimizing confounding factors such as changes in temperature, respiration rate or animal position. However, the wealth of data generated by PET/MRI and their complex origin requests for advanced computational analysis methods, but in turn these data provide a novel input for mathematical models. By using novel data in conjunction with computational models, we propose to take an important step in determining the metabolic basis of brain connectivity. For this we want to acquire so far unique, combined PET/MR data using a variety of PET-tracers investigating glucose metabolism, blood flow, the serotonergic and the dopaminergic system of the rat brain during rest and stimulation. This data will be acquired in combination with fc-fMRI, and will hence allow a direct comparison of fc-PET and fc-fMRI. We will further develop specific computational neuroscience methods for the analysis of fc- PET data, based on independent component analysis (ICA) as well as graph based network measures. Such methods have so far not been presented. Our combined data acquisition and data analysis approach will then be utilized to investigate if fc-PET and fc-fMRI information are redundant or complimentary. We also envision that the quantitative nature of PET data gives novel insights into the energetic budget used by large-scale brain networks. This US-German research proposal has the potential of a tremendous impact on science but also society in general. The novel and unique brain connectivity data might yield detailed insights into brain networks, based on specific transporter systems in the brain - this is of fundamental interest for basic research, neurophysiology but also for computational neuroscientists who aim to implement novel networks in their modeling and theoretical framework. The proposed analysis methods will be useful for medical imaging scientists, since it derives so far unutilized information from PET imaging data. Moreover, also clinicians can apply the developed fc-PET techniques in a variety of neurological diseases ranging from brain tumors to Alzheimer`s or Parkinson disease. It is especially the metabolic basis of these diseases, which can potentially earlier be identified using fc-PET and fc-fMRI methods developed in this proposal. This would especially in aging societies have a tremendous effect not only on the economical burden of such diseases, but might also due to a better treatment monitoring give new hope to millions of patients. Therefore our project can be seen as the basis of a framework that can be applied to a huge variety of basic research as well as clinical challenges involving fc networks.

? DESCRIPTION (provided by applicant): In a recent survey in South Africa, 95% of HIV-positive pregnant women received antiretroviral therapy (ART), and 68% of HIV-exposed infants received ART. Among the estimated 330,000 HIV-infected children in South Africa, 58% receive ART. Research suggests that there are long-term effects associated with in utero exposure to HIV and ART, warranting further study of this burgeoning population of HIV-exposed uninfected (HEU) children. In particular, very little is known about the long-term effects of HIV infection an ART treatment on brain development in pediatric populations, which we will investigate in the proposed study. We have access to an on-going longitudinal neuroimaging study combining several complementary, non-invasive magnetic resonance imaging (MRI) data sets with clinical and neuropsychological measures. The existing study (funded from R21 MH096559 and R01 HD071664) has been following a cohort of children carefully monitored since birth as part of the Children with HIV early antiretroviral therapy (CHER) trial (part of Comprehensive International Program for Research in AIDS (CIPRA-SA), sponsored by the Division of AIDS, National Institute of Allergy and Infectious Diseases). At 7 weeks, HIV-infected children were randomly assigned to either one of two early ART arms or to a deferred ART arm. In the cohort, the HIV-uninfected control group includes both HIV-unexposed, uninfected (HUU) and HEU children. At 5 and 7 years, children received clinical, neuropsychological and neuroimaging (structural, diffusion and spectroscopic; resting state functional MRI (RS-FMRI) at 7 years only) assessment. At 9 years (beginning June 2014), the same multimodal imaging and evaluations will be repeated. This project seeks to expand the current neuroimaging measures (under the existing R01) to include MR spectroscopic lactate and lipid concentrations, and to develop new multimodal markers involving brain network connectivity measures. We hypothesize that these additional multimodal biomarkers will be sensitive to the effects of HIV infection, timing and long-term usage of ART, and in utero ART/HIV exposure on brain development. This project extends the existing collaboration between Drs Meintjes (University of Cape Town), Laughton (Stellenbosch University), and van der Kouwe (Massachusetts General Hospital) to include Dr Bharat Biswal from New Jersey Institute of Technology (NJIT) who is a leading authority on RS-FMRI analyses, as well as two UCT-based early stage investigators, Drs Holmes and Taylor.

Dynamic brain signals provide key information that can be deciphered for a better understanding of brain function. Functional Magnetic Resonance Imaging (fMRI) has been developed to map not only the activity pattern but also functional connectivity in the whole brain. Although it has been over twenty years since the development of fMRI and the resting-state fMRI, the key challenge of fMRI-based signal interpretation remains to be the temporal and spatial resolution limit of the hemodynamic signal detected by fMRI. In most fMRI studies, the size of these voxels pre-determines the limits of the basic biological conclusion. With recent advancement leading to dramatic improvement in spatiotemporal resolution of MRI, the dynamic signal feature can be better clarified, which will significantly improve our understanding of brain function. In this proposal, a merged effort from three research groups is made to study the neural and vascular correlates of laminar-specific resting state fMRI signal fluctuation, which underlies the functional connectivity mapping in both human and animal models at varied brain states. Since the first report of large scale spatial signal correlation in fMRI images (Biswal et al, 1995), the drastically improved spatiotemporal resolution of high-field fMRI has revealed a number of key networks in the brain relevant to the default mode, attention, cognition, and sensorimotor connections. However, the millimeter size of voxels for resting-state and task results in fMRI signal dominated by large venous vessels. Although numerous studies have been designed to exclude signal contribution from veins, the neuronal correlates of resting-state and task fMRI signal has been heavily linked to vascular contributions. It remains a challenge to disentangle the distinct neuronal and vascular contribution to fMRI signal fluctuation given the limited spatial and temporal resolution. As connectivity measures are increasingly being used, a number of groups are beginning to focus on mind body interactions. In particular, connectivity measures, both local and global are being used in brain regions including insula and thalamus to better understand more complex brain behavior interactions.