Eventos Anais de eventos
COBEM 2021
26th International Congress of Mechanical Engineering
ESTIMATION OF THE BRAIN TEMPERATURE IN DEEP BRAIN STIMULATION APPLICATION WITH THE PARTICLE FILTER METHOD
Submission Author:
Caroline R. Pereira , RJ
Co-Authors:
Caroline R. Pereira, Luiz A. S. Abreu, Diego Knupp, Lucas Jardim
Presenter: Lucas Jardim
doi://10.26678/ABCM.COBEM2021.COB2021-1582
Abstract
Deep Brain Stimulation (DBS) is a well-established medical therapy that consists of sending electrical pulses to specific brain areas using electrodes implanted inside the brain. The thermal effect involved is patient-specific and difficult to predict. Although the use of DBS is widely spread around the world and has a well-tolerated surgical procedure, some brain injuries could be related to internal burns and non-expected hot spots around the electrodes. This work deals with the sequential estimation of the internal temperature of the brain containing a DBS lead by solving a state estimation problem with the Particle Filter methods. The classical bidimensional bioheat equation was considered to modeled this problem. The associated direct problem was solved with the Generalized Integral Transform Technique (GITT) and this solution was verified with a finite element approach implemented through the NDSolve function, intrinsic to the Mathematica software. In addition, some brain parameters will be estimated in order to take in account the thermal properties differences that occurs between the patients. The state estimation problem was solved with two particle filters: the sampling importance resampling (SIR) algorithm and the Liu and west Particle Filter. Experimental temperatures ware obtained supposedly with a sensor located inside the brain electrode. The uncertainties related to the physical and geometric parameters of the mathematical model, aiming at a better prediction of the temperature field of the tissues inside the brain ware taken into account. Uncertainties in the evolution and observation models were assumed as additive, Gaussian, uncorrelated and with zero means. For the brain parameters estimation, Gaussian prior information ware considered. The results were obtained considering simulated temperature measurements, with different noise levels and accurate estimations were obtained with both algorithms especially for small noise levels. The results shown a promising approach to reducing the risks of lesions related with the Deep Brain Stimulation technique.
Keywords
deep brain stimulation, bioheat transfer, Inverse problem, particle filter methods
