FreeSurfer v5.3 (Recon all)

Clinical/Biological Phenomenon of study:
Study of hippocampal anatomy. Hippocampal volumes in template space (MNI) are contrasted against a normative population of 200 healthy control subjects.
Description:
Freesurfer is a set of automated tools for reconstruction of the hippocampus (together with other brain cortical areas and subcortical brain structures from MRI data).
Resources required:
2GB, 1 day, single core

FreeSurfer v6.0 (Recon all)

Clinical/Biological Phenomenon of study:
Clinical/Biological Phenomenon of study: Study of the main cortical and subcortical regions: hippocampal, amygdala, ventricles, thalamus, basal ganglia, whole brain volume and temporal cortex atrophy. These variables are contrasted against a normative population of 532 healthy control subjects. All values are normalized on total intracranial volume. Additionally, the segmentation of hippocampal subfields is available.
Description:
Freesurfer is a set of automated tools for reconstruction of brain cortical areas and subcortical brain structures from MRI data
Resources required:
2GB, 1 day, single core

FreeSurfer v7.1 (Recon all)

Clinical/Biological Phenomenon of study:
Study of the main cortical and subcortical regions: hippocampal, amygdala, ventricles, thalamus, basal ganglia, whole brain volume and temporal cortex atrophy. These variables are contrasted against a normative population of 385 healthy control subjects. All values are normalized on total intracranial volume. Additionally, the segmentation of hippocampal subfields and amygdalar nuclei is available.
Description:
Freesurfer is a set of automated tools for reconstruction of brain cortical areas and subcortical brain structures from MRI data.
Resources required:
2GB, 1 day, single core

Adaboost (ICBM152)

Clinical/Biological Phenomenon of study:
AdaBoost is a machine learning method that is used to segment hippocampus regions from 3D T1-weighted structural brain Magnetic Resonance (MR) scans. This has been made possible thanks to the support of the following contributors: Paul Thompson (USC) and DECIDE initiative. Hippocampal volumes are contrasted against a normative population of 200 healthy control subjects. Volumes obtained are in template (ICBM152) space.
Description:
Hippocampal volume segmentation.
Resources required:
2GB, 1 hour, single core.

Adaboost (native space)

Clinical/Biological Phenomenon of study:
AdaBoost is a machine learning method that is used to segment hippocampus regions from 3D T1-weighted structural brain Magnetic Resonance (MR) scans. This has been made possible thanks to the support of the following contributors: Paul Thompson (USC) and DECIDE initiative. Hippocampal volumes are contrasted against a normative population of 421 healthy control subjects. Volumes obtained are in native space. All values are normalized on total intracranial volume computed using SPM12.
Description:
Hippocampal volume segmentation.
Resources required:
2GB, 1 hour, single core.

Hypometabolic Convergence Index (HCI)

Clinical/Biological Phenomenon of study:
HCI is an AD-related hypometabolic convergence index. HCI has been shown able to discriminate patients with clinical AD from healthy older persons. This has been made possible thanks to the support of the following contributors: Kewei Chen and Eric Reiman (Banner Alzheimer’s Institute, Phoenix).
Description:
Brain hypometabolism.
Resources required:
1GB, 1 hour, single core.

METAROI

Clinical/Biological Phenomenon of study:
MetaROI is an average metabolism index computed on a set of analytically derived regions of interest (i.e.: left angular, right angular, left temporal, right temporal, and bilateral posterior cingulate binary masks in Montreal Neurological Institute space) reflecting AD hypometabolism pattern. MetaROI index is normalized using the pons and cerebellar vermis ROIs in MNI space. MetaROI has been shown sensitive in the detection of longitudinal cognitive and functional changes in AD and MCI patients. This has been made possible thanks to the support of the following contributor: William Jagust (UC Berkeley & Lawrence Berkeley National Laboratory, California).
Description:
Brain hypometabolism.
Resources required:
0.5GB, 1 hour, single core.

SPMgrid

Clinical/Biological Phenomenon of study:
SPMgrid is an automated voxel-based analysis algorithm used to detect cortical hypometabolism or hypermetabolism on FDG-PET brain scans. These scans are contrasted against a normative population of 225 healthy control subjects.
Description:
Brain hypometabolism and hypermetabolism.
Resources required:
1GB, 1 hour, single core.

Categorization, Clustering and Classification (CCC)

Clinical/Biological Phenomenon of study:
CCC allows a deep characterization and phenotyping of the Alzheimer’s Disease (AD) defining the subject clustering based on the multimodal combination of clinical, neuropsychological, biological and imaging variables. To interpret CCC results each cluster is cross-classified with hundreds of classical diagnosis of a reference dataset (Redolfi et al. 2020). The CCC algorithm is a machine learning (ML) tool originally developed by Tel Aviv University (Mitelpunkt et al., 2015).
Description:
Refinement of the AD spectrum diagnoses.
Resources required:
0.5 GB, 1 hour, single core.

Subtype and Stage Inference (SuStaIn)

Clinical/Biological Phenomenon of study:
SuStain is a data driven model that combines subtype model and temporal progression for AD-related brain atrophy. The model is based on volumes of relevant brain regions derived from T1-weighted 3D MRI scans processed with Freesurfer 5.3. SuStaIn algorithm was originally developed at University College London (Young et al. 2017), and the atrophy model was developed and validated at IRCCS Istituto Centro San Giovanni di Dio Fatebenefratelli in Brescia (Archetti at al. 2021).
Description:
Subtype and stage subjects on the basis of brain atrophy.
Resources required:
2 GB, 15 hours, single core.

LEArning Spatiotemporal Patterns in Python (Leaspy)

Clinical/Biological Phenomenon of study:
The Leaspy tool predicts the evolution of cognitive scores for an Alzheimer’s Disease patient given baseline and optional follow-up information. The Leaspy model allows tha modelling of longitudinal data into a logistic shape. The model employed on NeuGRID2  was trained on MMSE and ADAS-Cog-13 cognitive scores of ADNI amyloid-positive (PET or CSF Ab42) subjects. Leaspy was originally developed at Paris Brain Institute (Schiratti et al., 2017) and has been used for a variety of applications in medical data (Maheux et al., 2020; Koval et al., 2021).
Description:
4-year prediction of cognitive decline for AD patients.
Resources required:
0.5 GB, 15 minutes, single core.

MUQUBIA

Clinical/Biological Phenomenon of study:
Muqubia is a classifier based on sociodemographic, clinical and MRI data. It classifies an individual as healthy control (CN) or affected by Alzheimer’s Dementia (AD), Frontotemporal Dementia (FTD) or Dementia with Lewey Bodies (DLB). The user is retunred with hte classification of the individual wioth the probability of the classification for each class, alongisde the results of the MRI processing pipelines. More info can be found in De Francesco et al, 2023.
Description:
Individual classification between CN/AD/FTD/DLB.
Resources required:
2 GB, 2 days, single core.

Harmonization to RIN

Clinical/Biological Phenomenon of study:
This harmonization tool performs a MRI segmentation with Freesurfer  V7.3, then outputs of freesurfer are harmonized with either ComBat (Fortin et al., 2018) or Neuroharmony (Garcia-Dias et al., 2020) models trained on data from 297 individuals enrolled for the RIN project whose MRIs were acquired according to the RIN protocol (Nigri et al, 2022). The user is returned with the acquisiton parameters or each of the uploaded scan, raw Freesurfer outputs and harmonized Freesurferfe ouputs.
Description:
Harmonization of freesurfer outputs according to RIN protocol.
Resources required:
2GB, 1 day/10 scans, 10-cores