Phone: (412) 268-7836
Email: genovese@andrew.cmu.edu

Individual Website: http://www.stat.cmu.edu/~genovese

Ph.D., University of California

Research Interests

A major focus of my research involves developing statistical methods for making effective inferences from data in complex scientific problems. I have tackled statistical problems in a wide variety of disciplines, including astrophysics, solid-state physics, and neuroscience. Devising new statistical techniques that can help advance the science requires a strong emphasis on inter-disciplinary collaboration, where I can gain an appreciation for the critical scientific issues and can in turn communicate the critical statistical issues involved in the problem. Most of my work on applied problems involves capturing very complex physical processes with high-dimensional, often nonlinear, statistical models. Building effective models in these cases calls for a wide range of statistical and mathematical tools. It is this combination of relevance to scientific applications and pure mathematical fun that drew me to statistics initially.

In recent work, I have developed new methods for statistical inference from functional Magnetic Resonance Imaging (fMRI) data. For example, I have built a nonlinear, Bayesian model for the complex temporal structure of fMRI data. My model allows investigators to infer the shape and magnitude of task-related signal changes and provides accurate assessments of the uncertainty in these inferences. Inferences about these changes can help cognitive neuroscientists understand how the brain subserves the tasks being studied. The model incorporates the best available information about the underlying physical processes generating the data, but it is also designed so that it can evolve as new research elucidates the key components of the system. Inferences based on model fits offer improved precision, and more importantly, the method makes it possible to use fMRI data to address scientific questions of interest that were inaccessible to previous methods. I have encoded this method in the BRAIN (Bayesian Response Analysis and Inference for Neuroimaging) software package, which is publicly available (bundled with FIASCO) and under continuing development. I am currently working to extend the model in several directions.

Publications

• Genovese CR, Sweeney JA: Functional connectivity in the cortical regions subserving eye movements (with discussion), in Case Studies in Bayesian Statistics, Vol. 4, eds. Kass RE, Carlin BP, Carriquiry AL, Gatsonis C, Gelman A, Verdinelli I, West M: Springer Verlag (to appear).
• Genovese CR: Statistical inference in functional magnetic resonance imaging, currently under review (also Technical Report 674, Carnegie Mellon Department of Statistics), 1997.
• Genovese CR, Noll DC, Eddy WF: (1997). Estimating test-retest reliability in fMRI I: Statistical methodology. Magn Reson Med 38: 497-507, 1997.
• Noll DC, Genovese CR, Nystrom L, Forman S, Eddy WF, Cohen J: Estimating test-retest reliability in fMRI II: Application to sensory-motor and cognitive activation. Magn Reson Med 38, 508-517, 1997.
• Genovese CR, Stark PB: Data reduction and statistical inconsistency in linear inverse problems. Physics of the Earth and Planetary Interiors 98: 143-162, 1996.
• Genovese CR, Stark PB, Thompson MJ: Uncertainties for two dimensional models of solar rotation from helioseismic eigenfrequency splitting. Astrophys J 443: 843-854,