Spherical Brain Mapping

Spherical Brain Mapping

The Spherical Brain Mapping (SBM) is a feature extraction and visualization framework intended to map the internal structures and features of the brain onto a 2D image that summarizes all this information.

[kc-dual-button text1="Documentation" icon1="ion-document-text" icon_position1="left" text2="CODE" icon2="ion-code-download" icon_position2="left" middle_text="icon" icon3="ion-android-globe" _id="152628" link1="http://mapbrain.readthedocs.io/en/latest/quickstart.html||" show_icon1="yes" left_css="{`kc-css`:{}}" link2="https://github.com/SiPBA/mapBrain||" show_icon2="yes" right_css="{`kc-css`:{}}" separator_css="{`kc-css`:{}}" animate="||"]

Insight

Visualize 3D medical image volumes in 2D maps that highlight the statistical and textural properties of any tissue.

Analysis

Use the statistical radial imaging biomarkers to analyze population differences.

Labelling

Compute statistical tests on SBM maps and superimpose a template to locate significant brain regions.

Prediction

Train and test a classifier with SBM maps to avoid the small sample size problem.

SBM allows three-to-two dimension mapping by means of some statistical measures. The system is based on a conversion from 3D spherical to 2D rectangular coordinates. For each spherical coordinate pair (θ,φ), a vector v(θ,φ) oriented in that direction and starting at the defined origin (usually the Anterior Commisure), is defined. This vector creates a set of voxels V(θ,φ) that contains the intensities of all voxels crossed by v(θ,φ).

From V(θ,φ), a number of SBM measures can be computed, including including statistical values (average, entropy, kurtosis) and morphological values (tissue thickness, distance to the central point, number of non-zero blocks). These values conform a two-dimensional image that can be computationally and visually analysed.

A number of publications have already explored the ability of SBM in the diagnosis of Alzheimer's Disease:

  • F.J. Martinez-Murcia et al. Assessing Mild Cognitive Impairment Progression using a Spherical Brain Mapping of Magnetic Resonance ImagingJournal of Alzheimer's Disease (Pre-print). 2018. DOI: 10.3233/JAD-170403
  • F.J. Martinez-Murcia et al. A Spherical Brain Mapping of MR images for the detection of Alzheimer's DiseaseCurrent Alzheimer Research 13(5):575-88. 2016.
  • F.J. Martinez-Murcia et al. A Structural Parametrization of the Brain Using Hidden Markov Models-Based Paths in Alzheimer's DiseaseInternational Journal of Neural Systems 26(6) 1650024. 2016.

Installation

mapBrain is now available via pypi and can be installed directly from:

pip install mapBrain

Otherwise, copy the *.py files directly to the working directory, and import the library with import mapBrain.

Quickstart

The Statistical Brain Mapping is structured as a class that can be invoked from every script. The simplest approach would be using:

import mapBrain
import nibabel as nib

img = nib.load('MRIimage.nii')
sbm = mapBrain.SphericalBrainMapping()
map = sbm.doSBM(img.get_data(), measure='average', show=True)

And it's free!

Spherical Brain Mapping follows a GNU General Public License v3.0, so the code and all its derivatives are free (as in beer and in speech).