I am interested in know how feasible Positron Emission Tomography (PET)
scan can be used in real-time based systems. And how portable the unit
can be made (say in 10 years.)
I have formulated several rough schemes to train the brain to responde
to different regions which can then be used as a certain coding scheme.
While at Georgia Tech, I worked with Mike Furman using EEG, signal
processing, and Neural Net trying to locate the patterns. Due to the
summated neural response nature of EEG and the inherent attenuations of
the fields, the signals are not clean enough. Thus, I am looking at
several other alternative schems. One of which is using the blood-flow.
Case Western Reserve University
E-mail: ery2 at po.cwru.edu
PS.If you were following the thread on "fast MRI an lucid dream
proposals...", you may know that I've responded in several posts and
am aware of some of the limitations of PET.