Research Mohit Adhikari

Research team

• Bio-Imaging lab

Expertise

My main expertise and current research interest is in statistical and mathematical analysis and modeling of resting state neuro-imaging data, especially in brain diseases. I use network models that incorporate anatomical connectivity along with a mathematical model for neural activity of local brain regions to simulate whole-brain functional connectivity (FC) that is then compared with empirically measured FC. I have applied this approach primarily to data from stroke patients for two projects. In the first project, I analyzed resting state data from 8-9 year old children who had suffered pre or perinatal stroke and found evidence in favor of normal functional architecture in contrast with a compensatory reorganization. In the second project, I analyzed data from first time stroke patients with cortical lesions in the acute phase. Here, after finding an optimal, individualized computational model, two model-based measures - a graph theoretical one of network integration and an information theoretical one of variability in the model network responses to hundreds of topographically random stimuli - were obtained. We found that both measures were reduced in patients, especially at the level of resting state networks (RSNs) and they correlated with two robust physiological markers of acute stroke, namely, reduced inter-hemispheric FC between homotopic regions and increased ipsilesional FC between task positive and task negative RSNs. I have also worked on two other projects involving resting state data from monkeys in which we used novel analysis methods. In one such project I applied a well-defined statistical framework to find evidence for dynamic (non-stationary) FC patterns. Several recent studies have investigated dynamic FC during the resting state but many of them either don’t do any statistical testing or use an ill-defined null hypothesis to generate incorrect surrogate distributions of test statistic. We found evidence for dynamic FC in these data using both linear and nonlinear test statistics but only when their values were averaged across 25 sessions thereby increasing the power of the test. In the second project, we found significant evidence for preferential activation of the default mode network by ripple events in the hippocampus of monkey brains. This preferential activation was not observed after other neural events in the hippocampus. Recently I contributed to an interesting project on brain development that explains the physiological rationale for a peculiar mode of migration of cortical interneurons. When this mode, consisting of sharp transitions between pauses and high amplitude movements, was altered by knock-down of a protein, it was found that a higher number of interneurons arrived at the cortex prematurely thereby derailing normal brain development. My contribution was to connect these two key observations. I demonstrated using a statistical analysis that the changes in the cellular motion properties measured in vitro result in higher number of cells arriving at the cortex observed in vivo. The goal of my current research is application of new measures of temporal fluctuations of resting state FC in animal models to improve diagnosis neurodegenerative disorders, namely, the Alzheimer’s and the Huntington’s disease. I use measures of dynamic FC such as the quasi-periodic and co-activation patterns and model-based effective connectivity in combination with supervised learning methods to classify transgenic animals from wild type littermates. My aim is to identify disease biomarkers and understand their longitudinal evolution in parallel to disease progression.