|Yvette Bohraus||Semester project|
|Life Science Section, EPFL||January 2008|
In this project, we used functional Magnetic Resonance Imaging (fMRI) data for a single-subject experiment to investigate brain activation patterns following auditory stimulation. Stimulation consisted of tones with four different frequencies, presented through a special headphone. Thanks to the multivariate machine learning algorithm called "Support Vector Machines" (SVM) as well as the application of statistical methods, such as t-testing and non-parametric permutation testing, important voxels for the classification task have been determined for each stimulation frequency. This allowed us to finally visualize several decision maps for tonotopic organization of the human auditory cortex. An illustration is shown in the figure shown below:
Activation pattern of the four frequencies where the stimulus frequency increases from frequency 1 (blue) to frequency 4 (red).