Zhongming Liu, Ph.D. - US grants

DESCRIPTION (provided by applicant): Dynamic interactions among large sets of brain regions produce all human perception, cognition and behavior. It is increasingly recognized that most mental disorders are caused by disruptions of distributed neural circuits, the structure and function of which still remain poorly known. Therefore, mapping the anatomy and dynamics of human brain networks is critical for us to understand the mechanisms underlying a variety of human behaviors and mental illness. However, significant progress in this area is hindered by technical limitations of existing neural recording and imaging techniques. To date, there is no single non-invasive neuroimaging technique ca- pable of providing a complete spatiotemporal pattern of whole-brain neuronal interactions. There is a critical need to establish new non-invasive imaging methods with high spatial and temporal resolution to uncover neural circuit dynamics in normal vs. diseased brains. To meet this critical need, we propose to establish and validate a novel multimodal hyperspectral imaging (MHI) technique, based on simultaneous acquisition and joint analysis of functional magnetic resonance imaging (fMRI) and electroencephalography (EEG), to permit high- resolution mapping of brain activity and connectivity at specific frequencies over the full spectrum of brain dynamics. This unique technique combined with diffusion MRI (dMRI) will be immediately usable to create a significantly enriched human brain connectome that will not only depict detailed connections among anatomically specific brain regions, but also assign to each region and each connection color-coded spectral signatures indicating their differential degrees of involvement in distributed network activities over various neuronal time scales across whole-brain neural circuits. To achieve this objective, we propose to accomplish three specific aims. 1) We will develop and optimize MHI through realistic computation simulations based on the virtual brain (TVB), a neuroinformatic platform to simulate the whole-brain network dynamics. 2) We will combine MHI and dMRI tractography to create a spectrally color-coded human connectome that entails both structural and functional connectivity. 3) We will validate the cortical activity and connectivity imaged with MHI against those directly measured with electrocorticography (ECoG) from the same group of epilepsy patients undergoing neuro- surgical evaluation with implanted subdural grids. The outcome from the proposed research will provide a new imaging tool to uncover the network basis for accurately assessing brain functions and identifying biomarkers for diagnosis of mental disorders. This project will have a significantly positive impact in delineating the brain's structural and functional connectivity, paving the way for better understanding and diagnosis of mental health, and significantly aid treatment and prevention of mental disorders.

This project aims to advance scientific knowledge about human vision and use neuroscience to enhance artificial intelligence for computer vision. Vision is central to how humans see and explore the world. About a dozen brain regions work together to process visual information within a fraction of a second. It is hypothesized that these brain regions actively predict one's visual surroundings and use errors of prediction to update their internal representations and guide actions. However, it is not clear how the brain performs computations for recognition and prediction, and whether it is possible for a machine to mimic the brain and recognize and predict visual input in complex, noisy, and uncertain circumstances. This project will address these questions from computational, psychological, and neuroscientific perspectives and deliver new models, data, and tools that promote the synergy between artificial intelligence and neuroscience.

Investigators will design a model based on predictive coding in the brain, and test its ability to perform computer vision tasks and explain human behaviors and brain responses to naturalistic visual stimuli. The investigators will first develop a deep neural network referred to as the predictive coding network. Unlike existing feedforward neural networks, the currently predominant vision models, the predictive coding network has several defining features relevant to neural processing in the brain. It is bi-directional, processing information both bottom-up and top-down. It is recurrent, utilizing the same architecture for dynamic computation. It is parallel, allowing information processing to occur in parallel both within and across different layers. It is both discriminative and generative, reconciling image recognition and synthesis in a single framework. The predictive coding network will be evaluated against benchmark data sets. It is hypothesized to reach competitive performance with many fewer parameters than the state of the art. Then, the investigators will test the model's behaviors given naturalistic images degraded in various ways and/or presented for various durations. It is hypothesized that the model will be more robust and accurate after running for increasingly longer times and reach a time-accuracy tradeoff like human perception under similar conditions. To test this hypothesis, the investigators will perform human behavioral experiments and compare the model's behaviors against human behaviors. Further, the investigators will test the model's ability to explain brain responses to naturalistic images and videos, measured with functional magnetic resonance imaging and intracranial electroencephalography. The model is hypothesized to be able to predict the brain's dynamic activity and representation given naturalistic stimuli. The successful completion of this project is expected to deliver a brain-inspired vision model learnable and computable end-to-end. This model will empower machines with adaptive and robust vision and provide a tool for understanding the computational basis of biological vision.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Project Summary Perhaps the most notable example of ?interoception? is the ?gut feeling?. The stomach can affect intuition, emotion and cognition; the brain can regulate food ingestion and digestion. The stomach contains its own enteric nervous system, or the ?little brain? in the gut. It connects directly to the central nervous system via the vagus. The vagus nerves provide a bi-directional ? afferent and efferent ? neural pathway for rapid interactions between the stomach and the brain. The stomach-vagus-brain connectome is central to human health and has significant health implications at dysfunction. However, this connectome has not been mapped or characterized in detail. It is unclear where and how the brain monitors and regulates the function of the stomach in terms of its electrical rhythm, mechanical contraction, and nutrient handling. It is also not exactly clear how the vagus nerves relay sensory information from the stomach to the brain and convey motor control from the brain to the stomach. To fill these gaps, this project is aimed to characterize the central and peripheral neural circuits of stomach-brain interoception in rats. For the central component, we will use functional magnetic resonance imaging in awake animals to map the central gastric network and characterize its activity and connectivity with respect to gastric electrical rhythm, mechanical contraction, and nutrient handling. To verify the central gastric network, we will use neuroanatomical tracing with pseudorabies virus and herpes simplex virus type-1. For the peripheral component, we will use the vagus nerve and nodose ganglion electrophysiology to characterize the afferent signaling from the stomach to the brain and the efferent signaling from the brain to the stomach. To elucidate the causal interaction between the stomach and the brain, we will use cell-type specific chemogenetics to perturb the central gastric network and assess the resulting effect on the stomach and use vagotomy to perturb the vagal circuitry and assess the resulting effect on the brain. This project has 4 specific aims for mapping the central gastric network (Aim 1) and characterizing the central and peripheral neural circuits for stomach-brain interoception related to gastric electrophysiology (Aim 2), motility (Aim 3), and ingestion of nutrients (Aim 4). To accomplish these aims, we form a collaborative and interdisciplinary team of experts with leading and complementary expertise in magnetic resonance imaging, gastroenterology, neuromodulation and electrophysiology. Upon its successful completion, this project will have integrated cutting-edge technologies into a unique platform for comprehensive assessment of the central and peripheral functional neural circuits underlying stomach-brain interoception. As the immediate outcome, we will have established the central gastric network in the rat brain, disentangled its functional roles, and elucidated the causal, rather than correlational, interactions between the stomach and the brain. These outcomes will lay both mechanistic and technical foundations for better understanding of stomach-brain interoception and its profound implications to mental illnesses (e.g., stress and anxiety), neurological disorders (e.g., Parkinson?s diseases) and gastric disorders (e.g., functional dyspepsia), and the co-occurrence of both brain and gastric disorders.

PROJECT ABSTRACT Delayed gastric emptying is a common and serious complication among patients with long standing and poorly controlled diabetes. Current prokinetic therapies are limited and elicit serious side effects. An improved understanding of the pathophysiology of diabetic gastroparesis is critical to the development of new approaches in the treatment of this difficult disorder. Gastric emptying is a complex process that is tightly coordinated. Fundic accommodation, peristaltic and tonic antral contractions, and antral-pyloric coordination all play important roles in regulating gastric emptying. Because of technical limitations, we still do not know to what extent gastric emptying is produced by each of these gastric motor components. To address these deficiencies, we developed a robust strategy to image and characterize gastric motility and emptying in rats and humans based on contrast-enhanced magnetic resonance imaging (MRI) and computer-assisted image processing. This novel technology not only shows gastric anatomy, but also captures and quantifies stomach emptying, intestinal filling, antral contractions and pylorus opening with fully automated image processing. Based on our pilot investigations and studies derived from computational modeling and simulations of gastric flow, we hypothesize that proper coordination of gastric motor function is required for optimal emptying. Fundic motor events, antral contractions, and opening of the pylorus are closely coordinated. Abnormalities in these events can result in delayed gastric emptying. To test this hypothesis, we plan to perform gastric MRI studies in healthy and diabetic rats and humans. We have 3 Specific Aims: Aim 1: Develop and perform gastric MRI to examine gastric motor events under postprandial conditions in rats and healthy subjects. This will define normal gastric MRI profiles and elucidate how each component of the gastric motor events contribute to emptying in health. Aim 2: Apply MRI technology to study gastroparesis in STZ-induced diabetes in rats and investigate how vagal stimulation might improve antral duodenal coordination and enhance gastric emptying. Aim 3: Employ MRI technology to study gastric motility and emptying in diabetic patients with gastroparesis. We will examine how abnormalities of different components of gastric motor function contribute to delayed gastric emptying. In separate studies, we will investigate the mechanisms by which prucalopride improves gastric motility and emptying. Gastric MRI profiles will be compared between prucalopride responders and non- responders to elucidate which components of the gastric motor function are modified by prucalopride resulting in improved emptying. These gastric MRI studies may provide novel information to identify new strategies to improve gastric emptying in diabetic gastroparesis.