No exposure to radiation or shots or other chemicals is necessary in
fMRI. It is non-invasive, accomplished by the powerful magnetic field
that interacts with radio waves and allows researchers to see the
changes in neural activity as a person views images or performs tasks.
The instrument measures blood flow, which is correlated with neural
activity. More oxygen is needed in areas where neural activity is high –
and blood with higher oxygen levels has different magnetic properties
than blood with lower oxygen levels.
That doesn't mean a researcher can read your mind once you slide into
the machine's bore, almost two feet in diameter, where the magnetic
resonance imaging is done. But that might not be too far off, given a
sketch of recent research described by Princeton's Norman, who was part
of the pioneering team that developed an fMRI data analysis method
called "multivariate pattern analysis" or MVPA.
"It's basically mind-reading based on fMRI data," said Norman,
professor of psychology at Princeton's Neuroscience Institute. "We take
brain scans and guess what you're thinking."
Specifically, his lab studies learning and memory by looking for
patterns in brain activity that are related to specific thoughts and
The fMRI provides enormous quantities of data, but using that data in
valuable ways requires more than observation. Higher neural activity
does not mean anything unless the reason for increased activity is
understood. If the study goes from a blank screen to a face, for
example, and higher neural activity is noted, it might mean only that
the brain was interested in seeing something other than a blank screen.
But seeing a difference in activity between images of a boot and a face
or a scene and an object can provide meaningful information about how
the brain processes specific information.
The MVPA method uses pattern classification algorithms to deduce what
a person was viewing when the data were collected – a shoe or a bottle,
for example. The overall accuracy of these computer analyses – matching
pixel to pixel – has reached about 96 percent, Norman said.
Norman also is exploring competing memories and how memory is
retrieved or inhibited – work that could one day help clinicians treat
such conditions as post-traumatic stress disorder (PTSD).
The capacity offered in CBBI goes well beyond brain science.
Dawn Elliott, professor and chair of biomedical engineering, said she
is "thrilled" to have MRI capacity for her lab's ongoing study of
intervertebral disc function and degeneration. Much of her study
involves human spine segments and also goat spines, whose disc size and
properties are similar to humans. In the goats she is gathering
preclinical data that measures how well a potential treatment is working
over time. If the treatment is successful, these data would be used as
part of approval by the U.S. Food and Drug Administration.
The new center is designed to accommodate animals, and Elliott said
it will offer much more statistical power for longitudinal studies.