Key result
An SVM-based algorithm classified HRV signals into epileptic and non-epileptic classes with 86.74% accuracy for right-sided and 79.41% accuracy for left-sided focus seizures.
Why the study?
Does HRV signal analysis using SVMs improve classification accuracy of epileptic vs non-epileptic signals in epileptic patients?
Observational (n=16)
Does HRV signal analysis using SVMs improve classification accuracy of epileptic vs non-epileptic signals in epileptic patients?
Absolute Event Rate: 86.74% vs 79.41%
HRV analysis using SVMs can classify epileptic seizures with high accuracy, demonstrating that right-sided seizures are associated with ictal tachycardia and left-sided seizures with ictal bradycardia.
May support HRV-SVM for seizure focus lateralization; leaves open prospective validation before clinical adoption.
OBJECTIVE: Epileptic onsets often affect the autonomic function of the body during a seizure, whether it is in ictal, interictal or post-ictal periods. The different effects of localization and lateralization of seizures on heart rate variability (HRV) emphasize the importance of autonomic function changes in epileptic patients. On the other hand, the detection of seizures is of primary interests in evaluating the epileptic patients. In the current paper, we analyzed the HRV signal to develop a reliable offline seizure-detection algorithm to focus on the effects of lateralization on HRV. MATERIALS AND METHODS: We assessed the HRV during 5-min segments of continuous electrocardiogram (ECG) recording with a total number of 170 seizures occurred in 16 patients, composed of 86 left-sided and 84 right-sided focus seizures. Relatively high and low-frequency components of the HRV were computed using spectral analysis. Poincaré parameters of each heart rate time series considered as non-linear features. We fed these features to the Support Vector Machines (SVMs) to find a robust classification method to classify epileptic and non-epileptic signals. Leave One Out Cross-Validation (LOOCV) approach was used to demonstrate the consistency of the classification results. RESULTS: Our obtained classification accuracy confirms that the proposed scheme has a potential in classifying HRV signals to epileptic and non-epileptic classes. The accuracy rates for right-sided and left-sided focus seizures were obtained as 86.74% and 79.41%, respectively. CONCLUSIONS: The main finding of our study is that the patients with right-sided focus epilepsy showed more reduction in parasympathetic activity and more increase in sympathetic activity. It can be a marker of impaired vagal activity associated with increased cardiovascular risk and arrhythmias. Our results suggest that lateralization of the seizure onset zone could exert different influences on heart rate changes. A right-sided seizure would cause an ictal tachycardia whereas a left-sided seizure would result in an ictal bradycardia.
No takes yet. Share an insight, caveat, or question.
Behbahani et al. (2016) conducted an observational in Epilepsy (n=16). Right-sided focus seizures vs. Left-sided focus seizures was evaluated on Classification accuracy of HRV signals to epileptic and non-epileptic classes. An SVM-based algorithm classified HRV signals into epileptic and non-epileptic classes with 86.74% accuracy for right-sided and 79.41% accuracy for left-sided focus seizures.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: