Time: 15+16 October 2018
Place: UVA Amsterdam, REC H1.01, Roetersstraat 13, Amsterdam.
Teacher: Johannes Fahrenfort
Capacity: 20 participants
Credit: 1 EC
Renewed Summary: (5 Sept.2018)
In this two-day workshop you will learn about decoding and related multivariate approaches to analyzing MEG/EEG data. Approximately 50% of the workshop is devoted to interactive lectures, the other 50% of the workshop consists of hands-on tutorials/analyses. The lectures are generic, the practicals make use of the Amsterdam Decoding and Modeling toolbox (ADAM), which is written using MATLAB (see here: https://doi.org/10.3389/fnins.2018.00368)
The workshop will cover:
– A brief introduction of the historical background and physiological substrate of MEG/EEG
– Designing EEG/MEG experiments
– Backward decoding models in MVPA: how do they work, concepts and analytical approach
– The temporal generalization method
– Using classifier scores to map brain to behavior
– Forward encoding models in MVPA: how do they work, concepts and analytical approach
The examples in the course will be mostly from the field of consciousness, attention and working memory, but the concepts are easily applicable to other fields as well. On the first day you will learn the basic principles of decoding and run through a tutorial that makes use of an existing EEG dataset. In this tutorial you will compare ERPs to decoding results and become acquainted with the temporal generalization method. On the second day you will learn about more advanced multivariate methods, and you will be given the opportunity to analyze your own dataset (one you have already acquired), or – if you do not have data of your own – set up the analyses from scratch of a dataset that I will provide. In principle any ERP dataset that you have acquired lends itself to decoding/multivariate analysis.
At the end of the course, you will be able to execute decoding analyses using ADAM, and interpret the results on your own data.
– knowledge of MATLAB (opening and closing of files, executing code) is advantageous but not required
– Experience with collecting/analyzing EEG/MEG data is beneficial but not required