My research lies mainly in the multidisciplinary field of human brain mapping. Due to recent advances in technology, enormous human brain data in various modalities have been produced in many fields including biology, neurology, neuroscience, psychology, and psychiatry. The human brain data, one of the most commonly cited forms of big data, raise new challenges to data analysts, due to its unique properties that are different from conventional data. To address these challenges, I have collaborated closely with scientists outside statistics to develop new statistical models and computational algorithms for human brain research. Specifically, I have worked on and will continue to work on two major multidisciplinary projects.

(1) I have collaborated with psychologists for more than seven years to study the brain's emotion function and its effect on the human's social behavior. In our collaborative projects, I have developed various statistical methods to analyze multi-subject, stimulus-evoked functional magnetic resonance imaging (fMRI) data. (2) I have recently initiated a multidisciplinary epilepsy project, which builds on the internationally recognized F. E. Dreifuss Comprehensive Epilepsy Program (CEP) at UVA. The project studies abundant epileptic patients' brain data resources accumulated for many years and new data collected using advanced high-frequency recording technology by the CEP. We aim to develop data-driven approaches to accurately map normal and abnormal epileptic brain networks and use the developed quantitative tools to improve epilepsy diagnosis and treatment and thus, alleviate the burden over clinicians who otherwise have to examine epileptic patients' enormous brain data manually. For illustrative purpose, my research projects on human brain are depicted by arrows in the following figure.

(1) Spatial-Temporal Analysis of Multi-Subject Neuroimaging Data for Human Emotion Studies.

We have examined population-wide brain responses to designed negative emotional stimuli under different social contact conditions and located the anterior portion of dACC (adACC) to be responsive to the stimuli. adACC plays an important role in appraising and expressing negative emotion and is involved in regulating amygdala's affective response to fear. Our analysis confirms heterogeneous emotional functions of dACC subregions.

We also detected responses to negative emotional stimuli in the white matter corpus callosum, in line with a growing body of evidence on fMRI activation in white matter, especially in the corpus callosum. Our method potentially has captured the white matter activity in a data-driven manner and has potential to be used for further investigations of white matter and gray matter brain functions.

(2) Brain Network Studies of Epileptic Seizures Using Intracranial EEG Data.

Epilepsy is characterized by repeated seizures caused by abnormal, excessive, or synchronous neuronal activity. Focal seizures arise from a seizure onset zone (SOZ) and propagate to otherwise healthy brain regions. The neuronal systems that comprise the SOZ and the propagation zone create an epileptic network. The directional connectivity that forms the epileptic network is not well understood. This project aims to understand the directional connectivity of epileptic brain networks through statistical modeling of intracranial electroencephalographic (iEEG) data obtained from within the brains of patients with epilepsy. Understanding seizures as a network phenomenon is critical in the care and the treatment of patients who suffer from seizures that do not respond to medication, so-called drug-resistant epilepsy (DRE).

We have constructed several high-dimensional dynamic models of directional connectivity for iEEG data. Our analysis revealed that from the seizure initiation to seizure propagation, more and more regions become connected with the SOZ and that the SOZ has connectivity properties that stand out from other normal regions in the brain network.