Correlation dimension and integral do not predict epileptic seizures

Abstract

Reports in the literature have indicated potential value of the correlation integral and dimension for prediction of epileptic seizures up to several minutes before electrographic onset. We apply these measures to over 2000 total hours of continuous electrocortiogram, taken from 20 patients with epilepsy, examine their sensitivity to quantifiable properties such as the signal amplitude and autocorrelation, and investigate the influence of embedding and filtering strategies on their performance. The results are compared against those obtained from surrogate time series. Our conclusion is that neither the correlation dimension nor the correlation integral has predictive power for seizures.

The Impact of EEG/MEG Signal Processing and Modeling in the Diagnostic and Management of Epilepsy

“…A general shortcoming was a lack of specificity regarding the identification of a preictal state. In addition, Harrison et al [41] examined the sensitivity of the correlation integral and dimension for prediction of epileptic seizures comparing its performance against surrogate time series and concluded that neither the correlation dimension nor the correlation integral has predictive power for epileptic seizures. Further, they [42] reported that the accumulated energy, as a measure to predict seizures, did not appear to have predictive abilities for these data sets.…”

State-of-the-Art of Seizure Prediction

“…Some of the results were encouraging, whereas other results illustrated that certain approaches are unlikely to be worthwhile. In line with this inconclusiveness are recent controversies about the relevance of nonlinear approaches for the prediction of epileptic seizures (McSharry et al, 2003a,b;Maiwald et al, 2004;Mormann et al, 2005) and studies raising doubts about the reproducibility of previously reported claims (AschenbrennerScheibe et al, 2003;De Clerq et al, 2003;Harrison et al, 2005aHarrison et al, , 2005bLai et al, 2003Lai et al, , 2004Lehnertz et al, 2003;Maiwald et al, 2004;Winterhalder et al, 2003). Although there is evidence from several methods for identifiable precursors preceding partial onset seizures, one should keep in mind that this evidence is based on retrospective analyses of mostly intracranial EEG data recorded during evaluation for resective surgery.…”

Adaptive prediction of epileptic seizures from intracranial recordings

“…Using nonlinear features such as the largest Lyapunov exponent [15,16], the correlation dimension [17,18], and the dynamical similarity index [19,20], it was suggested that precursors of seizures may arise several minutes before the seizure, but later evaluations showed that similarity index [21] and correlation dimension [22,23] were ineffective for their prediction. The combination of multiple features in order to improve seizure prediction was suggested by a neural network study [24], which predicted seizures 3.45 min before occurrence on average.…”