|
Notes:
This extensible database management system has been developed and
implemented to address the problems associated with managing the
increasingly large and diverse datasets collected as part of the
MBIRN and FBIRN collaboratories, and throughout clinical imaging
communities at large. The Human Imaging Database at a particular
site can be extended to contain relevant information concerning the
research subjects used in an experiment, subject assessments, the
experimental data collected, the experimental protocols used and
any annotations or statistics (metadata) normally included with an
experiment.
The Human Clinical and Imaging Database has been designed so that
it can be customized and extended to contain relevant information
from any particular sites needs without requiring modification to
the schema itself. This information consists of data concerning the
research subjects used in an experiment, subject assessments and
demographics, the experimental data collected, the experimental
protocols used and any annotations or statistics (metadata)
normally included with an experiment or study. Additionally, the
database architecture uses row level fine-grain database security
allowing for tight control over which individual records and fields
are visible to a particular user.
The major contributions from the development of the Human Imaging
Database are as follows:
•The database is composed of an extendible database and structured
core. The core database contains a hierarchical description of the
experiment (Figure 1). It defines the structure of the experiment
and how experimental protocols relate to this hierarchy. This
structure allows for the storage of an experiment in a rigorous
framework. Each descriptor in the database consists of a "base
tuple" which defines the minimum informational requirements of that
descriptor. For example, the base description of an experimental
event (i.e. stimulus) contains the base information required to
describe when that event occurs during an experimental protocol.
However, the actual information regarding various stimuli will need
to be stored in the database. Therefore, this tuple can be extended
for various experiments. For example, a researcher may collect
information regarding the actual properties (e.g. frequency and
strength) of a tone used during an fMRI experiment. These extended
tuples can be re-used and/or modified for other experiments. A
comprehensive printout of the data
schema is included in the Appendix.
•An extensible framework for the definition and storage of subject
assessment data. A dedicated section of the schema has been added
to handle the storage of assessment data from a wide variety of
assessments. Each assessment is defined through entries to parts of
the schema allowing for a very flexible system that allows for the
storage of virtually any assessment. Currently, some of the
assessments implemented include: Anxiety Status Inventory (ASI),
Beck Depression Inventory (BDI), North American Adult Reading Test
(NAART), Self-rating Anxiety Scale (SAS), SCID non-patient
(overview and screening module) modified by Stanford, quick mood
scale, race, ethnicity, education, and occupation.
•The incorporation of information regarding the history (i.e.
provenance) of how derived data was generated. All information
regarding the software used to generate derived data (e.g. software
name, version, hardware platform) are stored within the Human
Imaging Database to allow the full reconstruction of the processing
pipeline that was used to generate the data.
Changes:
|
