We do not have any open postdoc or PhD positions at the moment. Please contact Athina Tzovara for inquiries.
The human brain is characterized by rich neural dynamics, which are the result of coordinated electrical neural activity. Stimuli from our environment, like sounds, trigger strong reactions in distributed brain regions. In our recent work, we are studying how the human brain processes sounds from the environment via rich recordings of neural activity in patients with epilepsy.
Computational techniques are increasingly employed to analyze the large amounts of data that are generated from invasive recordings of neural activity of the human brain. In this project, we will employ machine learning and simulation techniques to analyse neural signals, with the goal of better understanding how our brains react to sounds. These signals come with incredible temporal and spatial resolution and are thus ideal to study brain processes at the micro- and mesoscopic levels.
The project is suitable for a Master thesis project. The student working on it should be motivated to program in Python for analyzing rich datasets. This project will give experience with machine learning, signal processing, neuroscience and working with human data. No prior experience with neural signals analysis is required.
For more information please contact:
Riccardo Cusinato: riccardo.cusinato@unibe.ch
Athina Tzovara: athina.tzovara@unibe.ch
Artificial intelligence is increasingly used for studying brain functions in health and disease. AI techniques can assist with automating sleep scoring or detecting epileptic activity, via electrophysiological (EEG) recordings.
EEG is a non-invasive technique that measures time-series of electric activity of the brain, though electrodes placed on the scalp. EEG is used as a diagnostic tool in neurological disorders, like epilepsy, and also for monitoring sleep, via polysomnography recordings. Today, there are several techniques for automating sleep scoring, and for identifying epileptic spikes. However, these techniques are not yet used in the clinical practice, and need further validation.
This master thesis will use signal processing and machine learning techniques to analyse rich EEG datasets of patients with epilepsy. The goals are to (a) validate existing techniques for detecting epileptic spikes with the use of AI, and (b) to validate existing AI algorithms for sleep scoring, for the case of sleep scoring in epilepsy patients, who have abnormal patterns of EEG signals.
The student working on this project will gain experience in signal processing of time series data, machine learning for biomedical data, and in the development of clinical biomarkers.
For more infomration please contact:
Athina Tzovara: athina.tzovara@inf.unibe.ch
Isolated REM-sleep behavior disorder (iRBD) is an early stage of alpha-synucleinopathy and may occur in isolation more than 10 years before the diagnosis of PD or as part of its symptomatology, then termed secondary. Clinically, it is characterized by vivid dreams and their acting out during sleep. Currently, there is no available biomarker indicating disease progression of iRBD or its conversion to Parkinson’s disease. In analogy to Alzheimer’s disease, we know that a disturbance of the sleep-wake rhythm and circadian regulation occurs before the onset of full-blown Parkinson’s disease.
A method that can detect subtle early signs is actigraphy. Actigraphy uses continuous measurement of movement (acceleration moments) per minute over several days. Advantages of this method are numerous: on one hand its clinical reliability as well as good acceptance by patients, and on the other hand, the objective reliability and subsequent statistical evaluation of the measured values.
The goal of this master thesis is to analyze a rich actigraphy dataset. The student will then use this method retrospectively in a mixed cohort of patients with isolated and secondary RBD in order to examine the analysis results of the clinical subgroups in comparison to the control group and, in the case of several measurements of individual patients, to analyze the data over the course of the disease follow up. For this project programming skills in Python, Matlab or R are needed and interest in analyzing actigraphy data, recorded from wearable devices. The student that will work in this project will gain experience analyzing rich actigraphy datasets, and will have the opportunity to apply advanced data analysis methods on a clinical application. For more information, please contact:
Dr. Carolin Schäfer, Oberärztin, SWEZ, Neurologie, Inselspital : Carolin.schaefer@insel.ch
Prof. Athina Tzovara: athina.tzovara@inf.unibe.ch