Neuronal pattern recognition

The neuronal network commonly referred to as the human brain or more general the central nervous system still poses many riddles that are yet to be solved. The usage of in vitro based neuronal cultures has facilitated the analysis of these networks, regarding the effects of electric, electromagnetic or biochemical stimuli. Yet a detailed assessment of the culture’s reaction is still quite complicated due to the high complexity of the network that nonetheless remains. In order to improve the results of such studies it is crucial to investigate the networks on a single cell level, so that changes in network activity or communication patterns can be identified correctly.

Figure 1: Microelektrode-array with cultured cells & recorded cell signals.

With this task at hand we established new neuronal pattern recognition methods that are able to discriminate the signals generated by different neurons. The resulting spike sorting algorithm combines elaborated feature extraction techniques, e.g. wavelet packet analysis, with powerful classification or clustering methods.

Figure 2: The integration of this spike sorting algorithm into our analysis software Dr. Cell enables detailed studies on neuronal network behaviour and offers the opportunity to conduct more complex measurements and analyses

The integration of this spike sorting algorithm into our analysis software Dr. Cell enables detailed studies on neuronal network behaviour and offers the opportunity to conduct more complex measurements and analyses.

Intention recognition via brain-computer-interfaces

Brain-Computer-Interfaces (BCI) are a very elegant way to access the human brain and to detect human thoughtprocesses. This non-invasive approach allows EEG measurements in everyday situations, such as driving a vehicle or using an entertainement device, and hence information on the effect to the test subject can be derived. - Consequently this kind of data offers great opportunities in various areas e.g. the field of safety improvements and entertainment sector anstatt There is a great variety of possible uses for this kind of data e.g. in the field of safety improvements and entertainment sector.

A particularly interesting area is identification of the intention of the subject. By analysing the EEG data derived by the BCI system it is possible to recognise the reaction of the subject through classification algorithms or pattern recognition techniques respectively. The gained information can then be used to facilitate any of the user’s upcoming actions.

Figure 3: Origin of EEG data. Changes in the activity of specific brain regions can be detected in the electrical brain signals derived by the BCI electrodes.

Recent publications:

R. Bestel, A. W. Daus, C. Thielemann, A novel automated spike sorting algorithm with adaptable feature extraction, accepted in 2012, Journal of Neuroscience Methods. Open access