Data Analysis and Statistical Modelling of Dynamical Processes in Neurology
Final Report Abstract
Numerous clinical phenomena in neurology are of dynamic oscillatory nature and are caused by inte-acting subsystems. Analyzing the individual time series and the detection of the network structure of such subsystems allows for an understanding of the underlying processes in the healthy and the diseased state. This project has developed and applied mathematical methods in order to infer from measured time series the dynamic properties of the subsystems and their interactions. The development of the mathematical methods was driven by the challenges in time series analysis posed by clinical applications. The application of the developed methods enhanced clinical understanding, provided diagnostic and prognostic factors and will eventually guide treatment strategies. The analysis of time resolved interaction plays an important role in many applications. In the scope of this DFG funded project we have developed a time-resolved spectral analysis technique, which allows also the investigation of time-resolved Granger-causality, to be precise time-resolved renormalized partial directed coherence. With this technique we were able to estimate the direct directed interactions structure together with the strength of interaction in an application to Parkinson tremor as well as epilepsy. Whenever modeling using ordinary differential equations is possible, these ordinary differential equations enable through their parameters a detailed description and understanding of the behavior of the system. In the cerebral autoregulation application we, thereby, suggested a new quantity to measure CO2 reactivity in stroke diagnosis.
Publications
- Application of a multivariate seizure detection and prediction method to non-invasive and intracranial long-term EEG recordings. Clin. Neurophysiol. 119, 2008, 197-211
Schad A., Schindler K., Winterhalder M., Schelter B., Maiwald T., Brandt A., Timmer J., Schulze-Bonhage A.
- Multivariate analysis of dynamical processes: Point processes and time series. Europ. Phys. J. 165, 2008, 25-34
Henschel K., Hellwig B., Amtage F., Jachan M., Lücking C.H., Timmer J., Schelter B.
- Assessing the strength of directed influences among neural signals using renormalized partial directed coherence. J. Neurosci. Meth. 179, 2009, 121-130
Schelter B., Timmer J., Eichler M.
- High functional connectivity of tremor related subthalamic neurons in Parkinson's disease. Clinical Neurophysiol. 120, 2009, 1755-1761
Amtage F., Henschel K., Schelter B., Vesper J., Timmer J., Lücking C.H., Hellwig B.
- Parametric versus nonparametric transfer function estimation of cerebral autoregulation from spontaneous blood-pressure oscillations. Cardiovasc. Eng. 9, 2009, 72-82
Jachan M., Reinhard M., Spindeler L., Hetzel A., Schelter B., Timmer J.
- Time-variant estimation of directed influence during Parkinsonian tremor. J. Physiol. 103, 2009, 348-352
Sommerlade L., Henschel K., Wohlmuth J., Jachan M., Amtage F., Hellwig B., Lücking C.H., Timmer J., Schelter B.
- Distinguishing direct and indirect interactions in oscillatory networks with multiple time scales. Phys. Rev. Lett. 104, 2010, 038701
Nawrath J., Romano M.C., Thiel M., Kiss I.Z., Wickramasinghe M., Timmer J., Kurths J., Schelter B.
- Identification of preseizure states in epilepsy: A data-driven approach for multichannel EEG recordings. Frontiers in Computational Neuroscience 5, 2011, 32
Feldwisch-Drentrup H., Staniek M., Schulze-Bonhage A., Timmer J., Dickten H., Elger C.E., Schelter B., Lehnertz K.
- On the estimation of the direction of information flow in networks of dynamical systems. J. Neurosci. Meth. 196, 2011, 182-189
Sommerlade L., Amtage F., Lapp O., Hellwig B., Lücking C.H., Timmer J., Schelter B.
