Frequenty Asked Questions

What data formats does the data center support?

For submissions, the fMRIDC database will accept all data formats from authors wishing to submit their data to the Center. Currently we ship the data in its native format or convert it to Analyze format for shipment. In the future, we will ship the data to users of the center in the format they request. All data conversion, for data from submitting authors and for data requests, will be done by the Center.

What type of data does the fMRIDC handle?

The Center will accept data from any peer-reviewed fMRI study. In addition to the fMRI data (i.e. structural and functional MR data), the Center will also accept additional data that is part of the study (e.g., physiological recordings, behavioral data, etc).

Does the fMRIDC archive raw data, analyzed data, or both?

The Center asks submitting authors to submit the following data:

• Raw, reconstructed image volumes from the scanner.
• Pre-processed images used for statistical analyses (for example, the images that have resulted from slice timing correction, spatial normalization, smoothing, etc. just prior to their entry into your statistical analysis engine) and detailed descriptions of the image processing steps that were applied.
• High-resolution anatomical images from all subjects (eg. SPGR, MPRAGE, Etc.)
• Image volumes of your final statistical results for each subject as well as statistical group maps.

Please note that sending all images resulting from intermediate steps of data processing is not necessary. The pre-processed image volumes should include only the completely processed data that would then be entered into a statistical analysis engine. However, please include descriptions of all image processing steps taken and which analysis engine was used.

The Center's goal is to archive the raw, pre-processed and analyzed data.

Is analysis of raw data supported within the fMRIDC database?

Currently, users can only request raw data from the Center. However, in the future the Center will offer users the ability to do some form of analyses on the data stored within the fMRIDC database. This will allow users to develop and apply novel data mining techniques to the data housed at the Center.

What descriptors can be stored to describe a study?

The most important aspect for the storage and organization of any scientific experiment is the storage and description of the experimental protocols involved. Neuroimaging data is associated with a variety of descriptors that define an experiment (e.g. information regarding the equipment, imaging protocol, experimental paradigm, statistical methods). The Center requests certain information deemed necessary in describing various aspects of the experiment. In addition to this information, the Center requests that each submitting author supply all the information they deem necessary to fully describe their experiment. In order to accommodate this, the fMRIDC database is designed so that it can be extended to store a researchers unique descriptors. This allows the Center to store all information concerning an experiment that the author and imaging community feel are necessary to fully describe the experiment. This framework also allows the Center to store information for descriptors as methods evolve and new methods emerge.

What quality control measures are implemented at the fMRIDC?

Their are a few issues regarding quality control that are of vital importance to the data center. First, the Center accepts data which are part of peer-reviewed articles. The quality control for the studies which the Center accepts lies with the neuroimaging community and its peer-review process. Second, all information concerning a study is provided by the authors themselves and the final study description is approved by the authors. Third, the Center does examine aspects of the submitted data to check the integrity of the stored and disseminated data.

How are data presently archived and how will the archiving, search and retrieval mechanisms scale with the growth of the Center?

Currently, data housed at the Center are primarily stored on RAID5 disk arrays. In the future it is anticipated that data from imaging studies will be stored in a hierarchical fashion (i.e. stored on disk ,CD and tape). However, the descriptive information for each study (i.e. summary information concerning the data that comprise the study as well as detailed information regarding the equipment, imaging protocol, experimental paradigm) will always be stored on disk and will always be immediately available for queries and browsing.

Does the fMRIDC database use a centralized, distributed or hybrid model?

The fMRIDC is a centralized resource for peer-reviewed neuroimaging studies. However, it is also important to realize that the fMRIDC database is part of a larger distributed federation of neuroscience databases being constructed. Therefore, it is of vital importance that fMRIDC database be able to interoperate with this federation.

Are the interface and query languages standard or customized?

Features for the query interface have been derived from the PubMed search engine - a search engine that many in fMRIDC's target user community are already familiar with. Users can begin constructing searches at the fMRIDC site without needing to learn a new search language. However, the interface has been customized with some new features. An extensible grammar allows for both Boolean operators as well as field tags to limit search results. Basic searches available now can be enhanced for complex queries on the underlying data and metadata. Flexible field tag settings allow the user to indicate how closely their search phrases should match and weights may be assigned to field tags allowing the user to emphasize those search criteria that are important to them.

Are the interfaces web-based, or do they require a dedicated application?

The interface to the Center's database is web-based and can be accessed by anyone with a computer and a modern internet browser that supports JavaScript.

How can data be retrieved from the fMRIDC Center?

Users can query the studies housed at the Center through a web browser. If a user requests a data set, the data will be provided electronically. Special arrangements must be made if you require the data burned to CDs.

What rules govern the use of data derived from the database, who has the right to publish findings based on these data, and how is credit assigned?

A model for data sharing has already been established by other communities. Namely, the genetics community has established policies for data stored in the various sequence database (e.g. GenBank). For data housed at the Center, anyone has the right to publish findings based on these datasets. Papers whose results are based on datasets obtained from the Center should credit the authors of the original study and acknowledge the Center and accession number of the dataset.

What rules protect the confidentiality of experimental subjects, and how are these kept in alignment with local IRB regulations and informed consent procedures?

Subject confidentiality is a key issue for the success of the fMRIDC. For each data set being submitted: