Title :
Estimation of the neuronal activation using fMRI data: An observer-based approach
Author :
Laleg Kirati, T.M. ; Arabi, H. ; Tadjine, M. ; Zayane, Chadia
Author_Institution :
Electr. & Math. Sci. & Eng. Div. (CEMSE), King Abdullah Univ. of Sci. & Technol. (KAUST), Thuwal, Saudi Arabia
Abstract :
This paper deals with the estimation of the neuronal activation and some unmeasured physiological information using the Blood Oxygenation Level Dependent (BOLD) signal measured using functional Magnetic Resonance Imaging (fMRI). We propose to use an observer-based approach applied to the balloon hemodynamic model. The latter describes the relation between the neural activity and the BOLD signal. The balloon model can be expressed in a nonlinear state-space representation where the states, the parameters and the input (neuronal activation), are unknown. This study focuses only on the estimation of the hidden states and the neuronal activation. The model is first linearized around the equilibrium and an observer is applied to this linearized version. Numerical results performed on synthetic data are presented.
Keywords :
biomedical MRI; blood; haemodynamics; medical image processing; neurophysiology; observers; physiological models; state-space methods; BOLD signal; balloon hemodynamic model; blood oxygenation level dependent signal; fMRI data; functional magnetic resonance imaging; hidden state estimation; neural activity; neuronal activation estimation; nonlinear state-space representation; observer-based approach; unmeasured physiological information; Adaptation models; Blood; Hemodynamics; Mathematical model; Numerical models; Observers; adaptive observers; fMRI; neuronal activation; nonlinear hemodynamic model;
Conference_Titel :
American Control Conference (ACC), 2013
Conference_Location :
Washington, DC
Print_ISBN :
978-1-4799-0177-7
DOI :
10.1109/ACC.2013.6580691