Marta Amengual-Gual1, Iván Sánchez Fernández2, Tobias Loddenkemper3. 1. Pediatric Neurology Unit, Department of Pediatrics, Hospital Universitari Son Espases, Universitat de les Illes Balears, Palma, Spain; Division of Epilepsy and Clinical Neurophysiology, Department of Neurology, Boston Children's Hospital, Harvard Medical School, Boston, MA, USA. Electronic address: marta.amengual.gual@gmail.com. 2. Division of Epilepsy and Clinical Neurophysiology, Department of Neurology, Boston Children's Hospital, Harvard Medical School, Boston, MA, USA; Department of Child Neurology, Hospital Sant Joan de Déu, Universidad de Barcelona, Spain. Electronic address: ivan.fernandez@childrens.harvard.edu. 3. Division of Epilepsy and Clinical Neurophysiology, Department of Neurology, Boston Children's Hospital, Harvard Medical School, Boston, MA, USA. Electronic address: tobias.loddenkemper@childrens.harvard.edu.
Abstract
BACKGROUND: The occurrence of epileptic seizures in seemingly random patterns takes a great toll on persons with epilepsy and their families. Seizure prediction may markedly improve epilepsy management and, therefore, the quality of life of persons with epilepsy. METHODS: Literature review. RESULTS: Seizures tend to occur following complex non-random patterns. Circadian oscillators may contribute to the rhythmic patterns of seizure occurrence. Complex mathematical models based on chaos theory try to explain and even predict seizure occurrence. There are several patterns of epileptic seizure occurrence based on seizure location, seizure semiology, and hormonal factors, among others. These patterns are most frequently described for large populations. Inter-individual variability and complex interactions between the rhythmic generators continue to make it more difficult to predict seizures in any individual person. The increasing use of large databases and machine learning techniques may help better define patterns of seizure occurrence in individual patients. Improvements in seizure detection -such as wearable seizure detectors- and in seizure prediction -such as machine learning techniques and artificial as well as neuronal networks- promise to provide further progress in the field of epilepsy and are being applied to closed-loop systems for the treatment of epilepsy. CONCLUSIONS: Seizures tend to occur following complex and patient-specific patterns despite their apparently random occurrence. A better understanding of these patterns and current technological advances may allow the implementation of closed-loop detection, prediction, and treatment systems in routine clinical practice.
BACKGROUND: The occurrence of epileptic seizures in seemingly random patterns takes a great toll on persons with epilepsy and their families. Seizure prediction may markedly improve epilepsy management and, therefore, the quality of life of persons with epilepsy. METHODS: Literature review. RESULTS:Seizures tend to occur following complex non-random patterns. Circadian oscillators may contribute to the rhythmic patterns of seizure occurrence. Complex mathematical models based on chaos theory try to explain and even predict seizure occurrence. There are several patterns of epilepticseizure occurrence based on seizure location, seizure semiology, and hormonal factors, among others. These patterns are most frequently described for large populations. Inter-individual variability and complex interactions between the rhythmic generators continue to make it more difficult to predict seizures in any individual person. The increasing use of large databases and machine learning techniques may help better define patterns of seizure occurrence in individual patients. Improvements in seizure detection -such as wearable seizure detectors- and in seizure prediction -such as machine learning techniques and artificial as well as neuronal networks- promise to provide further progress in the field of epilepsy and are being applied to closed-loop systems for the treatment of epilepsy. CONCLUSIONS:Seizures tend to occur following complex and patient-specific patterns despite their apparently random occurrence. A better understanding of these patterns and current technological advances may allow the implementation of closed-loop detection, prediction, and treatment systems in routine clinical practice.
Authors: Christopher J Re; Alexander I Batterman; Jason R Gerstner; Russell J Buono; Thomas N Ferraro Journal: Front Neurol Date: 2020-06-23 Impact factor: 4.003
Authors: Kathryn A Salvati; George M P R Souza; Adam C Lu; Matthew L Ritger; Patrice Guyenet; Stephen B Abbott; Mark P Beenhakker Journal: Elife Date: 2022-01-04 Impact factor: 8.140