Functional Characterization of Brain Areas Using Functional Magnetic Resonance Imaging

Introduction

Functional magnetic resonance imaging (fMRI) is a non-invasive neuroimaging modality that is continuously growing, both in the clinical and scientific fields. The analysis of these images requires a very complex and varied post-processing of the obtained images. This causes the results of different studies to be non-comparable or difficult to characterize. In order to simplify the processing and obtain objective results with analyzable metrics, this work proposes the development of an analysis methodology to obtain statistical values on brain activation areas segmented by region.

Methods

Several specialized tools were used. First, the FreeSurfer scientific package was used for brain segmentation. Then, widely used software for the analysis of fMRI data, FSL, was used for activation areas. The last processing and analysis steps were performed with the 3D Slicer medical image visualization tool. Based on these tools, a method was defined to obtain fMRI activation metrics for each of the 148 brain structures obtained by FreeSurfer.

This method was applied to the database proposed by [Masterson TD et al., 2016], where studies of neuronal response to visual food stimuli were carried out at two different times of the day. From this, it is possible to perform a comparative analysis of different metrics for each functional area and to define the levels of significance.

Results

A method was obtained for generating metrics that characterize functional studies based on the use of open-source scientific tools. This method was then applied to a database of fMRI images. Finally, a non-parametric statistical study was used based on the characterization of each area for all subjects, obtaining the most significant features.

Conclusion

This work applied a methodology for automatic fMRI image processing to obtain metrics and perform the most convenient statistical analyses. This method can be extended to different intra- and inter-patient comparisons.