www.psych.uic.edu

Articles

Runa Bhaumik, PhD

 

blank-picture-frame
CONTACT INFORMATION
Biostatistical Research Center
Department of Psychiatry
University of Illinois at Chicago
1601 W Taylor St, (M/C 912) RM # 437,
Chicago, IL 60612
Office Phone: 312-413-4455
E-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

Runa Bhaumik Ph.D.

Visiting Research Assistant Professor

Dr. Bhaumik is a visiting research assistant professor at Biostatistical Research Center, in the Department of Psychiatry. Her research focuses on Longitudinal Data Analysis, Multivariate Statistical analysis, Graph Theory and applications of Machine Learning Algorithms to Neuroimaging data.
Before joining UIC, she was an active research member at Data Mining and Predictive Analytics (DaMPA) Center in the Department of Computer Science at DePaul University. She earned her doctorate degree in computer science from DePaul University in 2011. During her doctorate program, her research focused on applied multivariate statistical analysis, machine learning and information retrieval. Her projects have been funded by the National Science Foundation (NSF). She is the author or a co-author of numerous papers in international conferences, workshops, and journals in the areas of data mining, Web personalization, and Recommender Systems. She had several years of teaching experience in Computer Science and Data Analysis at DePaul and several other universities.

Projects:

• Dimensional RDoC Modeling across the Range of Negative Mood Dysfunction (PI Scott Langenecker)

The goal of this project is to use advanced modeling and stratification techniques, with our expertise in biostatistics and statistical machine learning, to devise across-diagnosis subgroups that share core dimensions of dysfunction, which will link the neurophysiological abnormalities to disease risk and potentially provide more refined treatment targets.  We will pursue, an iterative strategy of (a) cross-modality scale development and consolidation for each subdomain (e.g., PCA, ICA), (b) illness vs well characterization via standard and novel techniques (e.g., support vector), (c) subtyping using scale and disease characteristics for each subdomain (cluster, ICA).

• Multimodal Fusion of Brain Imaging Data

The complex interactions between different brain structures can be studied by brain imaging methods, as the anatomical and functional properties of these structures can be measured simultaneously. Both diffusion tensor imaging (DTI) and functional magnetic resonance imaging (fMRI) gained popularity as non-invasive imaging tools for the study of the brain and the effects of neurological diseases. The primary goal of this project is to develop statistical approaches for combining structural and functional information in the context of connectivity studies. Further we plan to use machine learning algorithms to explore the ability of the network-based measures to provide diagnostic information for neurological diseases.  

• Pattern Recognition and Functional Neuroimaging 

Currently there are no known biological measures that can accurately predict future development of psychiatric disorders in individual at-risk. Our goal is to investigate whether machine learning and fMRI/DTI could help to differentiate healthy people from psychiatric disorder people.  

• Examining the capacity of publicly-funded after school programs to promote children's mental health in urban, poor communities (PI Stacy Frazier).

The primary goal of this study is to apply Principal Component Analysis and IRT modeling to evaluate Organizational Social Context survey items.  

• Links to Learning (PI Marc Atkins)

The primary goal of this study is to see whether the experimental intervention will lead to improvement in children's academic performance, behavior at home, and behavior at school relative to the treatment-as-usual in a longitudinal study.  

• Predicting and Detecting Drug Name Confusion Errors (PI Bruce Lambert)

The objective of this study is to develop three different types of prediction limits based on (i) mixed-effects logistic regression models, (ii) Poisson Distribution, and (iii) empirical Bayes estimates with a high assurance.  

• Toll-like Receptors and Cytokines in Depression and Suicide Brain (PI G. Pandey).

The goal of this project is to study Toll-like receptors, cytokines and neuroimmune genes in the postmortem brain of normal controls, suicides victims and depressed subjects to determine if abnormalities of the immune function are involved in the pathophysiology of depression and suicide. 

• Zero-Inflated Mixed Models for Mental Health Service Use

The overall goal of the proposed study is to develop new statistical estimation and testing procedures that reduce the considerable time, money, and subject burden now encountered in mental health services research. We hope that the new biostatistical approaches developed in this project will enable researchers to conduct higher-quality, more accurate studies that further our nation’s public health goal of recovery for significant proportions of citizens diagnosed with mental health disorders.  

You are here: Home