Skip to Main content Skip to Navigation
Theses

Modélisation en bosses pour l'analyse de motifs oscillatoires reproductibles dans l'activité de populations neuronales: applications à l'apprentissage olfactif chez l'animal et à la détection précoce de la maladie d'Alzheimer

Abstract : The method presented here, namely « bump modeling », provides a simple representation of time-frequency maps obtained by wavelet transformation of signals; the representation is parsimonious in terms of number of parameters. Time-frequency features can be extracted from the resulting models, which allows (i) a statistical analysis of large sets of signal recordings, and (ii) the detection of reproducible time-frequency patterns. We apply that method to the analysis of electrophysiological signals: LFP signals recorded from freely behaving rats responding to odorants, and EEG recordings of short duration, obtained from patients who are conjectured to be developing Alzheimer's disease. Our approach shows that the extraction of correlates of sensory information processing, and the early detection of pathological states, is possible from the analysis of complex oscillatory activity patterns generated by large neuronal populations.
Document type :
Theses
Complete list of metadatas

Cited literature [135 references]  Display  Hide  Download

https://pastel.archives-ouvertes.fr/pastel-00001508
Contributor : Ecole Espci Paristech <>
Submitted on : Friday, January 20, 2006 - 8:00:00 AM
Last modification on : Friday, October 23, 2020 - 4:37:25 PM
Long-term archiving on: : Thursday, September 30, 2010 - 6:53:04 PM

Identifiers

  • HAL Id : pastel-00001508, version 1

Citation

François Benoît Vialatte. Modélisation en bosses pour l'analyse de motifs oscillatoires reproductibles dans l'activité de populations neuronales: applications à l'apprentissage olfactif chez l'animal et à la détection précoce de la maladie d'Alzheimer. Sciences du Vivant [q-bio]. Université Pierre et Marie Curie - Paris VI, 2005. Français. ⟨pastel-00001508⟩

Share

Metrics

Record views

1111

Files downloads

1712