Our sensation of the world, such as light hitting the eye or sound hitting the ear, is often very different than our perception, or our conscious experience of these sensations.  The difference between sensation and perception is determined both by the nature of low-level process, such as how orientation and contrast are represented in early visual areas, and by how these early processes are influenced by higher-level cognition such as attention, short-term and long-term memory.  My goal is to understand how these low- and high-level processes interact to produce our perception of the world.

I use a combination of experimental approaches to study perception and cognition, including behavioral and neuroimaging (fMRI) techniques. There are many differences between these approaches, however, they have one thing in common: Both the thing that generated the data (the brain) and the noise in the data are extremely complex.  Learning about the brain requires that we be able to see though the veil of noise, in order to observe the true structure in the data that reflects the cognitive processes.  To do so, I develop advanced statistical analysis tools designed specifically for separating the noise in reaction time, accuracy and fMRI measurements, from the interesting parts of the data. I often develop hierarchical Bayesian model estimation techniques for doing so, as this rapidly-evolving statistical framework allows us to fit models that have so far been impossible using more standard methods.  These are often non-linear models that have thousands of parameters, and their development combined with cutting edge technologies like fMRI, is providing for an unprecedented ability to understand the processes that underlie our perception of the world.