Dr. Tülay Adali is the director of the Machine Learning for Signal Processing Lab (MLSP-Lab) at UMBC and has been working on diagnosing schizophrenia by analyzing functional MRI and other medical imaging data.
Dr. Tülay Adali, professor, specializes in the field of statistical signal processing. An IEEE and AIMBE Fellow, Dr. Adali recently received the 2010 IEEE Signal Processing Society Best Paper Award for the paper entitled “Complex ICA using Nonlinear Functions,” which discusses a new framework for complex-valued signal processing and demonstrates how it can be used for developing new methods for independent component analysis (ICA). At the moment, Dr. Adali is pursuing research that focuses on understanding how the brain functions and diagnosing mental illness by analyzing functional magnetic resonance imaging (fMRI) and other medical imaging data.
“What we try to do is minimize the assumptions traditionally invoked in data analysis as much as possible, since most end up being not realistic anyway,” says Dr. Adali of their research approach. “So, our emphasis is on data-driven approaches rather than those that are model driven. That is what machine learning is about.” The focus of Dr. Adali’s lab, the Machine Learning for Signal Processing Lab (MLSP-Lab) at UMBC, is the development of data-driven methods such as blind source separation, ICA and independent vector analysis. Dr. Adali has been the director of MLSP since she came to UMBC in 1992, and says she values most the interactions with those in her lab group and also the in-depth conversations, which often get very lively, during the weekly lab meetings.
Dr. Adali’s main research application area has been studying brain function, both while the brain is in action and while it is at rest. She has been collaborating with Dr. Vince Calhoun, her former Ph.D. student who is now a professor and the director of the imaging lab at the University of New Mexico and Chief Technology Officer at the Mind Research Network for Neurodiagnostic Discovery. Drs. Adali, Calhoun and their collaborators, among other projects, have been working on diagnosing schizophrenia by comparing the fMRI data of healthy and afflicted patients. Their goal is to develop an automated system for diagnosing a patient with the disorder, allowing for better treatment of schizophrenic patients. This research was recently highlighted in an IEEE Spectrum article entitled “The Psychiatrist in the Machine.”
Two recent research results Dr. Adali is especially excited about are related to complex-valued optimization and data fusion. About the latter, “normally, when people talk about fusion, they use one domain to constrain the other,” says Adali. “But we developed a method that can use all three modalities at the same time to extract the information.” The team recently combined data from functional magnetic resonance imaging (fMRI), structural MRI (sMRI), and electroencephalography (EEG) data and showed how a fully joint approach improves both the sensitivity and specificity of the identification biomarkers for schizophrenia. Adali says that this is the first case that she knows of where information from three different modalities has been fused together.
On the theoretical side, she is particularly excited about the very fundamental nature of the work that was recently recognized with the Best Paper Award. “This work builds on Dr. Hualiang Li’s Ph.D. work, currently a Research Assistant Professor at UMBC, and opens a completely new array of possibilities whenever data are best represented as complex valued,” she says. Besides signal processing, complex valued representations are essential to communications, radar, sonar, geophysics, oceanography, biomedical imaging, optics, and other applied sciences. She notes that all these areas may benefit greatly by working within the new framework their group has developed.
Since she began her research on brain function in 2000, Dr. Adali and her collaborators have made notable strides in the field. Nevertheless, Dr. Adali enjoys the research process as much as she enjoys the results. When asked about the ultimate aim of her research pursuits, Dr. Adali responded: “Just making sure that we do enjoy ourselves along the way, have fun, and learn something new every day.”