Study and implementation of the SENSE method for MR image reconstruction from multiple coils acquisition.

The performance of scanners for magnetic resonance imaging (MRI) had a leap in the last decade thanks to the advent of parallel acquisition, or acquisition with multiple coils. This type of acquisition allows either reducing the time of scanning, which is always important in acquisition of medical images, particularly in studies of functional magnetic resonance imaging (fMRI); or improving the signal-to-noise ratio (SNR) of the acquired images. The aim of this work was to study the reconstruction of MR images acquired with multiple coils, particularly using the SENSE method [1]. The choice of reconstruction method was due to the fact that our institution has a 3T MR scanner (Philips), which works with acquisition through multiple channels, and performs the reconstruction through a "black box " based on the SENSE method. In order to achieve our goal, we first undertook a study concerning the physical principles of MR signal generation and the reconstruction of MR images acquired in standard fashion (single coil acquisition). Next we studied the SENSE method and the file format of the raw data output by our scanner. The SENSE (Sensitivity Encoding) method was presented by Pruessmann et al. in 1999 [1], and it is based on the fact that the MR signal measured by a detector near the sample varies significantly depending on its position. That is, knowledge about the spatial sensitivity of a detector carries information on the origin of the MR signal detected, which can be used t