Prediction Error Profiles allowing a Seizure Forecasting in Epilepsy ?

Niederhoefer, Christian; Tetzlaff, Ronald

doi:10.1109/cnna.2006.341607

Cited by 11 publications

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Prediction error profile as described in [3]. Electrode points 40-60 show reasonable prediction performance.…”

Section: Patientmentioning

confidence: 89%

“…A seizure precursor may be detected by changes of the so-called signal features derived from EEG signals. Many publications [3][4][5][6][7][8][9][10] have been addressed to this field of research, but the problem remains unsolved. It has been shown that algorithms based on CNN [11][12][13] provided new results in the field of EEG-signal analysis.…”

Section: F Gollas C Niederhöfer and R Tetzlaffmentioning

confidence: 99%

See 1 more Smart Citation

Toward an autonomous platform for spatio‐temporal EEG‐signal analysis based on cellular nonlinear networks

Gollas

Niederhofer

Tetzlaff

2008

Circuit Theory & Apps

Self Cite

View full text Add to dashboard Cite

SUMMARYAlthough in the field of epileptic seizure prediction many spatio-temporal approaches have been carried out, the precursor detection problem remains unsolved up to now. It can be observed that an increasing number of algorithms are developed based on cellular nonlinear networks (CNNs). They are dealing with the extraction of signal features using intracranial EEG recordings in order to detect possible preseizure states. In general, reliable precursor detections cannot be obtained for all treated cases. The performance of these algorithms can be enhanced by adapting them to specific patients. Combining different features in a feature vector in a future seizure anticipation platform may lead to a reliably working seizure prediction system.In this contribution we focus on two different CNN-based algorithms-a nonlinear identification approach and a prediction algorithm. They will be discussed in detail and recently obtained results will be given.

show abstract

“…Prediction error profile as described in [3]. Electrode points 40-60 show reasonable prediction performance.…”

Section: Patientmentioning

confidence: 89%

Section: F Gollas C Niederhöfer and R Tetzlaffmentioning

confidence: 99%

Toward an autonomous platform for spatio‐temporal EEG‐signal analysis based on cellular nonlinear networks

Gollas

Niederhofer

Tetzlaff

2008

Circuit Theory & Apps

Self Cite

View full text Add to dashboard Cite

show abstract

“…They were also able to detect a pre-seizure phase for several minutes prior to the ictal state. Niederhoefer and Tetzlaff [49] used prediction error profiles to forecast seizures. They were able to show distinct changes minutes before seizure onset.…”

Section: Prediction Of Ictal States Using Eeg Analysis 251 the Preimentioning

confidence: 99%

Automated epileptic seizure onset detection

Ponnambalam

2010

2010 International Conference on Autonomous and Intelligent Systems, AIS 2010

View full text Add to dashboard Cite

“…Our previous results 10,11 showed that two CNN based approaches -the identification and prediction of EEG signals using Reaction-Diffusion CNN and delay-type discrete-time CNN (DT-CNN) -could be used to discriminate between a supposed pre-seizure state and a seizure free state in a statistical sense for the different long term recordings. However, these results take into account direct local couplings of EEG channels only and disregard possible couplings of spatially seperated regions of the brain.…”

Section: Introductionmentioning

confidence: 99%

Spatio-temporal coupling of EEG signals in epilepsy

2011

Self Cite

View full text Add to dashboard Cite

Approximately 1% of the world's population suffer from epileptic seizures throughout their lives that mostly come without sign or warning. Thus, epilepsy is the most common chronical disorder of the neurological system. In the past decades, the problem of detecting a pre-seizure state in epilepsy using EEG signals has been addressed in many contributions by various authors over the past two decades. Up to now, the goal of identifying an impending epileptic seizure with sufficient specificity and reliability has not yet been achieved. Cellular Nonlinear Networks (CNN) are characterized by local couplings of dynamical systems of comparably low complexity. Thus, they are well suited for an implementation as highly parallel analogue processors. Programmable sensor-processor realizations of CNN combine high computational power comparable to tera ops of digital processors with low power consumption. An algorithm allowing an automated and reliable detection of epileptic seizure precursors would be a"huge step" towards the vision of an implantable seizure warning device that could provide information to patients and for a time/event specific treatment directly in the brain. Recent contributions have shown that modeling of brain electrical activity by solutions of Reaction-Diffusion-CNN as well as the application of a CNN predictor taking into account values of neighboring electrodes may contribute to the realization of a seizure warning device.In this paper, a CNN based predictor corresponding to a spatio-temporal filter is applied to multi channel EEG data in order to identify mutual couplings for different channels which lead to a enhanced prediction quality. Long term EEG recordings of different patients are considered. Results calculated for these recordings with inter-ictal phases as well as phases with seizures will be discussed in detail.

show abstract

Prediction Error Profiles allowing a Seizure Forecasting in Epilepsy ?

Cited by 11 publications

References 12 publications

Toward an autonomous platform for spatio‐temporal EEG‐signal analysis based on cellular nonlinear networks

Toward an autonomous platform for spatio‐temporal EEG‐signal analysis based on cellular nonlinear networks

Automated epileptic seizure onset detection

Spatio-temporal coupling of EEG signals in epilepsy

Contact Info

Product

Resources

About