Search

The study was based on simulated magnetic resonance images (MRI) created by MATLAB software. The aim of the study was to estimate the effects of signal to noise ratio (SNR) and flip-angle variations, in the determination of longitudinal relaxation time T1. Simulated MRI experimentswere used to determine the optimal image pixel sampling method to be used in the calculation of T1. Five sampling methods were used to determine mean values of MR image: (1) Dividing the image into four equal blocks, (2) Dividing the image in the direction of flip angle variation, (3) Dividing the image in the direction perpendicular to the flip angle variation, (4) Taking mean pixel value of whole image, and (5) considering the values pixels by pixel. Calculation of T1 using the whole image was identified as the most effective approach, with a minimum deviation of 2.3 ms from the reference value 1000 ms. Minimum and maximum error percentages for flip-angle range 1600 to 1800 and SNR = 20 were 0.283% (whole image) and 3.288% (in the direction of flip angle variation) respectively. As expected, results show that the minimum deviations occur in narrowed flip angle regions towards 180 degrees and with high SNR values. T1 based extracellular volume (ECV) fraction error was found to be 21.78% for the whole image averaging method and 36.39% for four equal slice method at the lowest SNR (2.5) value., Annual Research Symposium, University of Colombo 2016