Premium
P3‐418: INTEGRATIVE PREDICTION OF COGNITIVE DECLINE USING ATROPHY NETWORKS, CSF BIOMARKERS, AND GENOTYPE DATA
Author(s) -
Richiardi Jonas,
Bigoni Claudia,
Roche Alexis,
Maréchal Bénédicte,
Peyratout Gwendoline,
Aouri Manel,
Braissant Olivier,
Henry Hugues,
Meuli Reto,
Kober Tobias,
Popp Julius
Publication year - 2018
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2018.06.1781
Subject(s) - atrophy , default mode network , cognitive decline , cognitive impairment , medicine , temporal lobe , cognition , neuroimaging , neuroscience , psychology , alzheimer's disease neuroimaging initiative , brain size , cardiology , pathology , disease , magnetic resonance imaging , radiology , dementia , epilepsy
classification of MCI/SMC (accuracy1⁄46161%) to the CSF benchmark model (accuracy1⁄46161%). Finally, models predicted MCI conversion to AD (mean yrs of conversion 1⁄4 2.161.8) with accuracy of 6864%, similar to the CSF benchmark (7062%). Comparison analyses revealed structural connectome estimates contributed to these models. Conclusions:Our validated results show the feasibility of the multi-modal MRI, particularly structural connectomes, combined with multi-variate machine learning algorithms as an accurate non-invasive brain marker of AD predicting diagnosis and disease progression. Optimizing the high-throughput phenotyping techniques and machine learning algorithms may improve the multi-modal MRI-based predictive modeling.