Motivierte Studierende haben die Möglichkeit in der Arbeitsgruppe ihre Masterarbeit zu schreiben. Eine Auswahl an möglichen Projekten findet sich nachfolgend.
EEG is recorded from electrodes on the scalp. These electrodes are quite far away from the neural tissue and it is difficult to determine where the recorded signal is originating from. Source modeling approaches allow to estimate the origin of neural activity. These models can be improved by taking the shape of the cortex of each individual into account. However, these models are still far from optimal. In this project we want to use ECoG data (recorded in epilepsy patients) to predict EEG activity. ECoG is recorded with electrodes that are directly located on the brain and therefore allow to estimate the origin of the recorded signal quite well. We are going to use this data to compute a forward model to predict the signal that can be recorded on the scalp using EEG. This work may help to improve source models. The project will work on existing data, no extra data acquisition will be required.
In daily life, we constantly interact with each other. In all our interactions we have clear expectations about the actions of others and make constant predictions about the other person’s action. As long as our predictions are met we can interact effortlessly with others. As soon as a person deviates from what we expect we have to take an extra effort to react adequately. To study this aspect of expectation building we use the so-called word-by-word game (taken from improvisational theater). In this game, two people build sentences together, one word at a time. As long as both have the same idea what the sentence is about this joint sentence building is easy and the words follow each other quickly. As soon as both have a different idea of the sentence we observe a significant slowing down. In this project, you will re-analyze an existing dataset of a behavioral study. We collected data from 30 pairs playing this game together and recorded everything they said. This project will work on existing behavioral data, no extra data acquisition will be required. This project requires advanced statistics.
EEG is generally recorded for experiments in the lab that normally do not last longer than two hours. We have acquired EEG data during a workday while participants did an auditory task using mobile ear-EEG. For this data, we find the expected event-related potentials (ERP). In this project, you will re-analyses the existing dataset focusing on the on-going EEG instead of the event-related potentials. This project will work on existing EEG data, no extra data acquisition will be required. This project requires a lot of signal processing.