| Resumo: |
Resting State functional Magnetic Resonance Imaging (rs-fMRI) provides information about the functional connectivity of brain areas. However, prior to calculating the functional connectivity of the brain, there is a choice of several preprocessing steps that need to be selected. A critical source of variation between studies arises from distinct preprocessing approaches prior to the functional connectivity analysis. Therefore, a study to examine the reliability of different methods for pre-processing data from rs-fMRI is necessary. In this study, seven preprocessing strategies were tested and the reliability was evaluated between them in Graph Theoretical (GT). The sample used in this study is from a public database and consists of control subjects. Measures of GT were calculated using different strategies, after applying a method of subdividing the brain into 190 regions of interest. The following measures were calculated: global efficiency, characteristic path length, clustering coefficient and local efficiency. The results indicate that there is a significant difference in measurements of GT depending on the preprocessing strategy selected. It was also found that noise estimation parameters are correlated with GT measures. Moreover, it is observed that the level of thresholding chosen in the connectivity matrix can affect the measurements of GT, therefore further studies regarding this topic are needed. It was concluded, based on the sample used in this work that the method of scrubbing by outliers could increase the reliability of measurements of GT and reduce dependence on the movement of the patient's head. |
|---|
