To: fMRI Research Community
From: National fMRI Data Center
Re: Informational Memo                             July 15, 2000


As many of you know, a data repository for functional imaging studies is
being
established. From time to time informational memos will be sent to the
fMRI
Research community in order to provide background information about the
activities of the Center and to respond to questions and concerns that
have
been raised by those interested in this enterprise. This is the first of
such
memos.

General Background: Since the advent of technology for imaging the human
brain
in vivo over the last decade there has been a steady and predictable
rise in
the number of laboratories performing studies aiming to explore the
landscape
of cognitive function. Whereas early on, only a few, well funded medical

centers were in a position to actively pursue these studies, the list of
groups
now conducting neuroimaging experimentation is considerable. Despite the

popularity of neuroimaging, however, there remains little, if any,
sharing of
raw image data occurring amongst members of this widening community.
Presently,
heroic efforts are required by researchers to extract statistical values
from
the published literature in order to conduct meta-analytic assessment of

neuroimaging findings (e.g. Cabeza and Nyberg, J. Cog. Neurosci., 12,
1-47).
Previous pioneering work to design and establish an online repository
for the
results of imaging experiments helped to advance the concepts of more
dynamic
meta-analytic assessment of findings from across numerous reports (i.e.
BrainMap, http://ric.uthscsa.edu/projects/brainmap.html). However, these

efforts were undertaken prior to the newest generation of database tools
that
enables the storage, manipulation, and searching of extremely large
image data
sets such as those collected in fMRI.

The National fMRI Data Center (http://www.fmridc.org) endeavors to act
as the
curator for such large data sets used to obtain results published in the

leading peer reviewed journals, thereby enabling other researchers to
download
and reproduce experimental results, examine new data processing
methodologies,
and evaluate novel experimental designs. In this manner, fMRI can more
easily
draw on the knowledge-base of many disciplines, including Mathematics
(e.g.
image processing and statistics), Computer Science (e.g. data storage,
compression, and management), as well as Neuroscience (e.g.
interpretation of
brain function).

The storing of scientific data in online, readily accessible databases
is
certainly not a new idea and examples exist in many different areas of
study.
In each case, such repositories have been an integral element in
speeding the
progress in the relevant discipline, allowing researchers to compare and

confirm results, and build on previous work. Most famous among such
efforts is
the Human Genome Project database (see Science, 288, 2294-2295,
2304-2307)
which has, for some time, been continually updated and made available to
anyone
via the internet (http://www.ncbi.nlm.nih.gov), enabling researchers and
laymen
alike to inspect and examine genetic sequences. Companion to the Human
Genome
Project is the Protein Data Bank (http://www.rcsb.org/pdb) administered
by the
Research Collaboratory for Structural Bioinformatics
(http://www.rcsb.org/index.html). This is the single international
repository
for the processing and distribution of 3-D macromolecular structure
data,
primarily determined experimentally by X-ray crystallography and nuclear

magnetic resonance.

The astronomical sciences have also made great progress by using shared
data,
and moreover, the huge expense of the equipment used in this field (e.g.

telescopes, satellites) has necessitated such an approach. Particularly
impressive is the remote sensing images acquired from the Mars Global
Surveyor
mission (http://mars.jpl.nasa.gov/mgs) that are currently being made
available
online to anyone within hours after the data's 1.6 million mile journey
from
our neighboring planet. Recent examination of these images has revealed
evidence for the existence of water on Mars (Science, 288, 2330-2335).
Additionally, the National Space Science Data Center
(http://nssdc.gsfc.nasa.gov) makes available images and spacecraft
telemetry
data from a large number of NASA space missions. These data include
Earth
terrain and atmospheric measurements, solar activity data, as well as
high-resolution images taken during human lunar activity. Additional
representative databases for other fields can easily be located using
any
internet search engine.

Analogously, in order for cognitive neuroscience, via the use of
functional
neuroimaging, to reach the goal of understanding the principles of brain

function it is necessary to increase the number of neuroscientists and
other
researchers that can examine, consider, analyze and assess the brain
imaging
data that has already been collected and published. The ultimate goal of
The
National fMRI Data Center is to help facilitate this process by making
available the raw data from these studies with the hope of advancing the

understanding of the cognitive function and the neural substrates that
underlie
them. This publicly available database which is the National Data Center
hopes
to serve this function for the scientific community.

The fMRI database project in many ways is more complex than the several
other
database systems just reviewed. The complexities of human
experimentation
require a demanding set of associated experimental information be linked
to the
brain images.  The type of stimuli used and the variety of experimental
paradigms will all have to be part of each set of data.  Added to this
is the
reality that different laboratories use different storage formats for
their
experimentation.  All of these aspects of fMRI research will be
addressed by
the Data Center and each author will be assisted in presenting the data
to the
Center.

It is widely recognized that sometimes more information is collected in
the
process of a typical MRI study than is originally reported. In order to
allow
the authors more time to analyze their data, the data center will be
following
certain guidelines. In this regard, in a letter to the Office of
Management and
Budget, the President of the National Academy of Sciences stated that
"permitting the researcher who actually collected the data to be the
first to
analyze and publish conclusions concerning the data is an essential
motivational aspect of research"
(http://www.nas.edu/includes/letter.htm). The
National Data Center concurs with this sentiment and also with reports
from the
National Academy of Sciences, issued through the National Research
Council,
supporting the concept of data sharing
(http://www.nas.edu/includes/letter.htm). Therefore, the Data Center
will offer
authors the option to place their data in a "data hold" for a fixed
period of
time where visitors to the data center may view only the published
statistical
data and not the underlying functional and anatomical data. This
procedure has
already been implemented in other communities, namely the Protein Data
Bank
(http://www.rcsb.org/pdb/nih-policy.html). As the Data Center continues
to
evolve, these policies will continue to be reviewed and refined with
regard to
the best interests of the fMRI community. In addition, our external
advisory
board, chaired by Dr. Marcus Raichle will continue to assess our
activity and
achievements.

We look forward to working with the fMRI community. If you wish to be
apprised
of new developments at the center please visit us at
http://www.fmridc.org.

Sincerely,
The National fMRI Data Center




%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%