Morphometry statistics

Gathers a number of morphometric figures about some regions of a set of subjects folds or ROI graphs

Description

This process does not actually make statistics, it only gives some descriptors on regions, according to a model. This process has been primarily written for cortical folds graphs and it is still its main usage, but it can also be used to obtain descriptotrs for gyri or ROIs. Other models could be added afterwards.

Imput graphs are generated by automatic segmentation tools: see the Morphologist process.

The subjects graphs must be at least partially labelled in order to extract the correct ROIs. It is possible to automatically label sulcus graphs using the morphometry / automatic recognition process (but it is very strongly recommended to verify the identifications)

The model holds the anatomical knowledge (list of regions...) and knows the associated descriptors and how to calculate them (virtual "experts" of the related anatomy).

The choice of regions is guided by a "nomenclature" which allows to select regions in a hierarchical way: by lobe or by large sulcus for instance. The editor is not very convivial yet (it will be improved in the future): up to now you must type regions names by hand.

It is possible to obtain figures on sulci and/or on relations linking to a sulcus. The process writes a file for each sulcus or relation (more precisely a file for an "expert" of the model). In each file, there will be one line for each subject.

Descriptors and models

Each data type (sulci, ROIs, gyri...) can me analyzed by one or several models. Each model embeds the different morphometry features descriptors calculation, so a model may provide some measurements, and another model different measurements.

As each model holds its own descriptors and their meaning, it is not possible to provide a common documentation for them. So you may use the viewer process on the model parameter (the eye icon) to display documentation of the specific chosen model.

Parameters

data_graphs: ListOf ( input )

Subjects ROI graphs (cortical folds, etc)

model: Model graph ( input )

Model graph which will provide its own set of morphometric descriptors

nomenclature: Nomenclature ( optional, input )

region: LabelSelection ( input )

Selection of names to process. This "selection" parameter type is a bit special: the associated selection file is optional (if omitted, the selection will not be saved for later usage). Region names are chosen using an editor that is opened using the "..." button. This editor uses the selected model and nomenclature. If no selection is done (if the user just doesn't use the "..." button), a default selection will be done and will contain all valid names found both in the model and the nomenclature. Most of the time, this is what we want.

output_directory: Directory ( input )

output_filename_prefix: String ( optional, input )

region_type: Choice ( optional, input )

Region: only the selected region Relations with region: models of the relations linking this region with ohters All: region + relations

label_attribute: Choice ( input )

run_dataMind: Boolean ( input )

Starts the statistics module interface. Needs R installed. R can be found on http://www.r-project.org/

Technical information

Toolbox : Morphologist

User level : 0

Identifier : morphometry

File name : brainvisa/toolboxes/morphologist/processes/morphometry/morphometry.py

Supported file formats :

model :

Graph and data

nomenclature :

Hierarchy

output_directory :

Directory