Advanced data analysis in brain functional studies with magnetic resonance imaging (Contributo in atti di convegno)

Type
Label
  • Advanced data analysis in brain functional studies with magnetic resonance imaging (Contributo in atti di convegno) (literal)
Anno
  • 2006-01-01T00:00:00+01:00 (literal)
Alternative label
  • Vanello N.; Santarelli M. F.; Positano V.; Ricciardi E.; Pietrini P.; Landini L. (2006)
    Advanced data analysis in brain functional studies with magnetic resonance imaging
    in Sixth International Congress on Progress in Bioengineering, Pisa
    (literal)
Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#autori
  • Vanello N.; Santarelli M. F.; Positano V.; Ricciardi E.; Pietrini P.; Landini L. (literal)
Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#note
  • In: Sixth International Congress on Progress in Bioengineering (Pisa, 10-11 October 2006). Proceedings, pp. 47-. (Biomedicine & Pharmacotherapy). Elsevier Science, 2006. (literal)
Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#descrizioneSinteticaDelProdotto
  • ABSTRACT: Functional magnetic resonance imaging (fMRI) is currently a well established technique that can be used to study brain function. It takes advantage of good localization skills of magnetic resonance imaging (MRI) systems to detect local changes in blood flow and oxygenation, as in the case of blood oxygenation level dependent (BOLD) related changes. Its applications range from basic neuroscience research to clinical applications as for presurgical planning or to study recovery after brain lesions. fMRI studies are addressing more and more complex cognitive tasks. Despite of this fact, even in the simpler task, the experimenters rely upon an incomplete model of the signal generation. In fact most of the sources of data variances, as physiological changes during the task, attention or abituation effects, are not known and can be difficulty controlled or measured. Moreover the experimenters cannot model exactly a priori the BOLD changes elicited by the task. For these reasons the results of classical confirmatory approaches, as those derived on the General Linear Model that are based on experimenters' a priori hypotheses, may lead to biased results. The interest of the research community is addressing exploratory approaches, that can be used to evidentiate unexpected or unmodelled phenomena from the data. Independent Component Analysis (ICA) has been proven to be useful in detecting physiological and task related signal changes as well as separate artefacts from data, exploiting the hypothesis of statistical independence of the extracted components. The general assumptions used in the ICA model, as a drawback, imply that the extracted components are difficult to be classified. Furthermore the best number of components to be extracted, i.e. the model order, is not known a priori. In this work we will introduce a method to classify the ICA exploiting the residual dependencies among the extracted components. A distance measure derived from mutual information is proposed. A hierarchical clustering stage is then applied to the distances among the components in order to classify spatially independent maps of fMRI and identify interesting groups of components. This method could be used both to reveal informative relationships between the ICs and to give information about the model order. The method is tested on simulated datasets, showing its capabilities in grouping the components resulting from a splitting process due to model order overestimation. Results on real datasets are reported, and advantages in the application of the method are discussed. (literal)
Titolo
  • Advanced data analysis in brain functional studies with magnetic resonance imaging (literal)
Prodotto di
Autore CNR
Insieme di parole chiave

Incoming links:


Prodotto
Autore CNR di
Insieme di parole chiave di
data.CNR.it