FUNCTIONAL NEUROIMAGING DATA ANALYSIS: FROM PREPROCESSING FUNCTIONAL CONNECTIVITY AND MULTIVOXEL PATTERN ANALYSIS

The course aims at introducing the fundamentals of functional neuroimaging data analysis. I will cover the principles of blood oxygenation level-dependent (BOLD) imaging, with insights into image artifacts, the temporal properties of the hemodynamic response function, and the design of task-based and resting state neuroimaging experiments. Using publicly available functional magnetic resonance imaging (fMRI) data, I will outline preprocessing steps that are commonly employed in modern analysis pipelines, such as slice timing correction, head motion correction, extraction of nuisance signals, spatial smoothing, signal intensity normalization and coregistration of anatomical and functional images. I will then explain how to estimate single-participant task-related brain responses using the mass-univariate general linear model approach and how to derive univariate maps of statistical significance at the group level. I will cover issues related to the multiple comparisons and how to adjust the level of significance to control for false positive results. Using resting state fMRI data, I will introduce the basics of seed-based connectivity and functional connectomics. Lastly, I will present two multivariate techniques that offer the possibility to study information content: representational similarity analysis and brain decoding. Practical sessions will be based on bash and MATLAB scripting.