Welcome to the MUQUBIA Tool in Neugrid!


This algorithm was developed for the Italian Network of Neuroscience Neurorehabilitation (RIN) (https://www.reteneuroscienze.it/en/)

MUQUBIA allows to make a differential diagnosis classifying subjects in: alzheimer’s disease, lewy body dementia, frontotemporal dementia, healthy subject.

How to run the analysis:

You have to upload a T1-weighted 3D MRI scan as a zip file containing DICOMS or nii.gz archive and a Diffusion Tensor Imaging as a zip file containing DICOMS or Nifti files (bval and bvec bvec files must be provided and must containt “bval” and “bvec” text in the filename), then provide age, gender and CDR values. Optionally a FLAIR scan can be uploaded to improve pial surfaces.

The MRI image will be processed with FreeSurfer 6.0 and relevant brain volumes and thickness will be extracted. Subcortical volumes are normalized on eTIV. In conjunction with information provided the pipeline assigns the subject a diagnosis.

Here you can download the classifier: DOWNLOAD


 

Select subject’s age: 

Select subject’s gender:

Select CDR score: 

Select T13D to upload:

Select DTI to upload:

(OPTIONAL) Select FLAIR to upload:


Description of MUQUBIA (MUltimodal QUantification of Brain whIite matter biomArkers in dementia)

MUQUBIA is a machine learning tool that uses age, gender, CDR, and MRI data to classify a person as having probable Alzheimer’s dementia, Frontotemporal dementia (FTD), dementia with Lewy bodies (DLB), or as a cognitively normal (CN) subject.

Images of 354 subjects from 5 different datasets (ADNI, FTLDNI, NACC, Newcastle, PDBP) were processed using Freesurfer v6.0 and TRACULA to obtain the full set of features.

MUQUBIA uses 22 features that maximize the classification accuracy to distinguish to which class the person belongs.

Representation of brain regions corresponding to the imaging features selected by MUQUBIA to distinguish the different diagnostic classes (AD, DLB, FTD, CN). The color of each brain region reflects the ability of the corresponding feature to discriminate among the different classes.