Brain Imaging Seminar Series: Peter Bandettini 3pm Georgia State University, Urban Life Bldg, Room 1199

Dynamics, Fluctuations and Patterns: Ongoing work towards separating physiologic from neuronal information in fMRI

Peter Bandettini, Section on Functional Imaging Methods and the Functional MRI Core Facility, National Institutes of Health, Bethesda, MD

In the past 18 years, functional MRI (fMRI) has experienced, by all measures, an explosive growth. The growth of the field is characterized by a continued advancement in the sophistication of the technology, methodology, and depth of interpretation of the signal. This talk will specifically outline some work from my research group which focuses on advances in how much more information we can extract from the time series fMRI signal and how cleanly we can separate neuronal information from the hemodynamics which transduce it. Specifically, my talk has three parts: Dynamics, fluctuations, and patterns. In the first part, I describe attempts using hemodynamic and neuronal modeling combined with Magnetoencephalography (MEG) to better understand the origin of the dynamic nonlinearities that are commonly observed in even-related fMRI. In the second part, I talk about the burgeoning subfield of fMRI fluctuations, and show that respiration and image acquisition itself may be a source of some of the time series correlations, but also show some promising new methods for “resting” state network extraction. Third, I talk about our groups’ work on “fMRI decoding.” I the utility of multi-voxel pattern extraction, and then demonstrate several advancements in this methodology for probing deeper into a more fine –grained spatial scale of brain organization within cortical regions. Overall, I would like to convey that fMRI methodology really does drive fMRI applications and the field of fMRI as a whole and that methodology advancement is showing every indication of increasing in rate of advancement.