What is a BrainVISA database ?

What is a BrainVISA database ?
Prev	Chapter 4. Databases and ontology	Next

A BrainVISA database is composed of two parts:

A directory containing data files that are organized in a hierarchy that follows BrainVISA ontology.
A relational database built from the ontology and allowing to make efficient selection requests on data files according to their attributes. For example: get all the Raw T1 MRI acquired in a given protocol.

So there are two ways of seeing the same data connected together by a common ontology.

Figure 4.1. Correspondence between database directory and BrainVISA relational database

BrainVISA database stores the following information about data:

Data type: identify the content of the data (image, mesh, functional image, anatomical MRI, etc). The data types are organized in a hierarchy making it possible to decline a generic type in several specialized types. For example, there is a 4D Image type which is specialized in 3D Image (indeed, a three-dimensional image is a particular case of a four-dimensional one); the type 3D Image is itself declined in several types, for example T1 MRI and Brain mask.
File format: Represents the format of files used to record a data (generally indicated by the file extension). BrainVISA is able to recognize several file formats (for example Nifti, GIS, Analyse, etc). It is possible to add new data formats and to provide converters to make it possible for existing processes to use these new formats.
Files: a data is generally composed of one or several files.
Attributes: an attribute is an association between a name and a value. A set of attributes is associated to each element of the BrainVISA database. This set represents all of the characteristics of a data (the name of the protocol, the subject, acquisition parameters...).

This information is filled in when the user imports new data in the database and determines the way data will be written in BrainVISA database directory. When new data and results are generated by a process, the database information is generally given by the process. For example, the process Brain Mask Segmentation generates a T1 Brain Mask image with the same protocol and subject attributes than the input data and there is a rule in the ontology that determines the way this type of data must be written in BrainVISA database.

The ontology rules that defines the type of data and their organization is useful in two ways. It enables to scan a database directory and extract all needed information from the files organization. It also enables BrainVISA to know how to write new data and results in the database according to the type of data declared in the process.