Statistical Methods in Fast Functional MRI

The proposal focuses on developing statistical methods and related theory

for fast functional magnetic resonance imaging (fMRI), to sharply improve

the time-resolution of present techniques. Our objective is to improve the

time-resolution of functional magnetic resonance imaging by sampling only a

small fraction of the Fourier transform of the spin density, and using a

wavelet filter to approximately obtain, not the usual susceptibility map,

but instead an integral, representing the total activity, of the difference

in susceptibility between task and pre-task, over a prespecified region of

interest in the brain at successive time-points. This space/time trade-off

thus allows us to obtain, at high time-resolution, the total activity in a

specified region of the brain, believed to process the specific stimulus/task,

to learn or verify where the brain function takes place. An example of the

use of the technique is to learn where in the brain the 'rolodex' of memory

is located whereby one recognizes a familiar face in a fraction of a second.

To try to find where the rolodex is located by a method which requires more

than a second is a doubtful project; better time-resolution is required (but

high spatial resolution is not). Although some researchers claim faster

measurement capabilities they sacrifice signal/noise.

Statistical methods and related theory will be developed to sharply improve

time-resolution for the technique of functional magnetic resonance imaging (fMRI). Our objective is to improve the time-resolution of fMRI by acquiring data from only a small fraction of the MRI sampling space, to approximately obtain the total activity of the difference in susceptibility between task and pre-task, over a prespecified region of interest in the brain at successive time-points. For a typical region of the brain, say describing the hippocampus, believed to be involved in memory, our optimal choice of the sampling region should give a ten fold speed-up compared to the usual method of sampling. An example of the use of the technique is to learn where in the brain the 'rolodex' of memory is located whereby one recognizes a familiar face in a fraction of a second. This fast fMRI is expected to have profound and far-reaching consequences in the understanding of brain function, a problem of central scientific interest at the present time.