Probabilistic and Statistical Methods

Learning objectives: 

Students are able to critically reflect on mathematical formulations of data analytical methods within the cognitive neurosciences. They have a functional understanding and formal knowledge of common statistical and model-based paradigms used to analyze imaging data. Students are also able to use this knowledge to evaluate and plan empirical investigations, particularly in the research areas of cognitive neuroscience, and are aware of the significance and limitations thereof.

Content:

Building upon the knowledge gained in previous studies, students deepen their understanding of the following topics: correlation and regression, multiple and logistic regression, application of the general linear model and multilevel models, frequentist and Bayesian reasoning with approaches to control error rates (especially type 1 errors). Students gain experience in practically applying their knowledge of multivariate analysis methods using data set examples from cognitive neuroscience while under supervision,  and are also able to gain experience with approaches based on machine learning. Advanced methods of neuroimaging data analysis such as biophysical modeling approaches (e.g., psychophysiological interactions, dynamic causal modeling, etc.) are implemented in programming languages such as Matlab, RStudio, or Python using toolbox implementations.