Multilevel Modeling with Structured Penalties for Classification from Imaging Genetics data - Université Pierre et Marie Curie Accéder directement au contenu
Communication Dans Un Congrès Année : 2017

Multilevel Modeling with Structured Penalties for Classification from Imaging Genetics data

Résumé

In this paper, we propose a framework for automatic classification of patients from multimodal genetic and brain imaging data by optimally combining them. Additive models with unadapted penalties (such as the classical group lasso penalty or L_1-multiple kernel learning) treat all modalities in the same manner and can result in undesirable elimination of specific modalities when their contributions are unbalanced. To overcome this limitation, we introduce a multilevel model that combines imaging and genetics and that considers joint effects between these two modalities for diagnosis prediction. Furthermore, we propose a framework allowing to combine several penalties taking into account the structure of the different types of data, such as a group lasso penalty over the genetic modality and a L_2-penalty on imaging modalities. Finally , we propose a fast optimization algorithm, based on a proximal gradient method. The model has been evaluated on genetic (single nucleotide polymorphisms-SNP) and imaging (anatomical MRI measures) data from the ADNI database, and compared to additive models. It exhibits good performances in AD diagnosis; and at the same time, reveals relationships between genes, brain regions and the disease status.
Fichier principal
Vignette du fichier
10.1007:978-3-319-67675-3 21.pdf (1.6 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01578441 , version 1 (29-08-2017)

Identifiants

  • HAL Id : hal-01578441 , version 1

Citer

Pascal Lu, Olivier Colliot. Multilevel Modeling with Structured Penalties for Classification from Imaging Genetics data. 3rd MICCAI Workshop on Imaging Genetics (MICGen 2017), Sep 2017, Québec City, Canada. pp.230-240. ⟨hal-01578441⟩
514 Consultations
158 Téléchargements

Partager

Gmail Facebook X LinkedIn More