Course information

Please register for the course in MOSES or via email to secure a spot. Registration in ISIS will not be considered binding.

The seminar will take place Fridays 10-12 in-person (potentially hybrid/remote) in room MAR 4.062.

Learning Outcomes:

Knowledge: Students will understand the basic concepts of statistical and physical forward and inverse problems. They will be familiar with the physics of magneto- and encephalography (M/EEG) measurements and the steps to calculate M/EEG forward models. They will know different mathematical frameworks for inverse modeling, such as dipole fitting, beamforming, distributed imaging, blind source separation, and Bayesian modeling. They will be familiar with the concept of regularization and ways to encode prior knowledge into inverse solvers. They will learn to address technical challenges such as estimating the noise level/choosing the regularization parameter.

Skills: Students will be capable of modeling inverse problems as unsupervised or supervised machine learning problems. Students will also acquire or refine skills to independently review and systematically structure the literature of a well circumscribed field in order to address a given set of questions, and will gain experience in presenting the outcome to a critical audience as well as in participating in scientific discussions.

Competencies: Students will be able to discuss the advantages and disadvantages of different modeling approaches depending on the problem setting and to make informed modeling decisions.

Description of Teaching and Learning Methods:

The module consists of a single seminar. Students will prepare a presentation to a specific topic based on a provided collection of published material. Student presentations will be framed by short lecture segments introducing, contextualizing and connecting the presented topics. Each course slot will contain discussion periods, in which active participation is fostered. To this end, students will also work out a set of preparatory questions for each topic, as well as a brief summary.

Prerequisites :

A BSc degree in Computer Science, Biomedical Engineering or a comparable field is recommended. 

Basis knowledge of linear algebra, stochastics, and numerical optimization is advantageous.

Content:

- common inverse problems in biomedicine, in particular neuroimaging
- physical foundations of magneto-/electroencephalography (M/EEG)
- forward modeling and physics simulation for M/EEG
- dipole fits, beamforming, scanning techniques
- penalized likelihood approach: smoothness, structured sparsity, elastic net, total variation denoising
- Bayesian inference: maximum a-posteriori estimation, hierarchical and empirical Bayes, sparse Bayesian learning
- noise learning approaches
- blind source separation as a statistical inverse problem
- simulations and validation
- research software
- applications in brain-computer interfacing and neurology

The module grade is calculated based on
1. The quality of a scientific presentation (50%).
2. The quality of short summaries for each topic (50%).