The New Jersey Institute of Technology's
Electronic Theses & Dissertations Project

Title:	Motor data scaling by respiration frequencies in rest
Author:	Kamble, Amit Madhukar
View Online:	njit-etd2010-117 (xi, 42 pages ~ 1.3 MB pdf)
Department:	Department of Biomedical Engineering
Degree:	Master of Science
Program:	Biomedical Engineering
Document Type:	Thesis
Advisory Committee:	Biswal, Bharat (Committee co-chair) Alvarez, Tara L. (Committee co-chair) Adamovich, Sergei (Committee member)
Date:	2010-08
Keywords:	Functional magnetic resonance imaging Motor data Respiration frequencies Scaling data
Availability:	Unrestricted
Abstract:	Functional Magnetic Resonance Imaging (fMRI) is widely used as a tool to see activations in the different brain regions. Motor data acquired from fMRI scan is accompanied with signal due to hemodynamic changes taking place during the scan. This hemodynamic signal is dominated by parameter alteration in the large vessels of brain. Scaling of task induced BOLD signal with hypercapnic or breathold data is one of the effective methods to minimize the signal due to large vessels. Patient discomfort and compliance has been a major issue with these methods hence in this study we used respiration frequencies in rest data to scale task induced data and compared results with breathold scaling. The correlation between respiration frequencies and breathold signal was very good indicating presence of hemodynamic component in respiration. Scaling was done in both time and frequency domain. Standard deviation was used in time domain and frequency fluctuation amplitude was used in frequency. Results of scaling with respiration frequencies and with breathold signal; were comparable and active areas during performance of motor task reduced after scaling, minimizing the effect of large vessels. These outcomes showed that respiration frequencies can be efficiently used for scaling data.