Dr M.W. Tangermann (Michael)

My expertise is in developing fundamental machine learning algorithms which meet challenges arising when applied to time series data obtained from neural recordings and behavioural measurements. These signals typically are high-dimensional, very noisy and non-stationary, i.e. they display changing signal distributions over time, which poses specific problems to decoding algorithms.

I work with recordings of the electroencephalogram (EEG), stereoelectroencephalography (sEEG), electrocorticogram (ECoG), local field potentials (LFP), magnetoencephalogram (MEG) and electromyogram (EMG) and signals recorded by external devices. In addition, behavioural data like responses to motor and language tasks are included.

I also apply the developed machine learning methods by implementing closed-loop neurotechnological systems that interact with the nervous system. Examples of such systems are brain-computer interfaces for communication, control and rehabilitation and adaptive deep brain stimulation systems for patients with neurological or psychiatric disorders. The closed-loop interaction realized by these systems requires machine learning solutions also for choosing optimal stimulation parameters given different brain states and task contexts.

I translate methods developed into neurotechnological applications for clinical- and non-clinical purposes, e.g., brain-computer interfaces for communication, control and monitoring, adaptive deep brain stimulation, language rehabilitation after stroke, and hand motor rehabilitation.

• Natural Computing & Neurotechnology
• Language and Communication

• Brain Computer Interfacing