DynamoSort: Using machine learning approaches for the automatic classification of seizure dynamotypes

Abstract Objective Epilepsy is characterised by unprovoked and recurring seizures, which can be electrically measured using electroencephalograms (EEG). To better understand the underlying mechanisms of seizures, researchers are exploring their temporal dynamics through the lens of dynamical systems modelling. Seizure initiation and termination patterns of spiking amplitude and frequency can be sorted into “dynamotypes”, which may be able to serve as biomarkers for intervention. However, manual classification of these dynamotypes requires trained raters and is prone to variability. To address this, we developed DynamoSort, a machine-learning algorithm for automatic seizure onset and offset classification. Methods We used approximately 2100 seizures from an intra-amygdala kainic acid (IAK) mouse model of mesial temporal lobe epilepsy, categorized by five trained raters. MATLAB’s classification learner application was used to create an ensemble model to score and label dynamotypes of individual seizures based on spiking and frequency features. Results Dynamotype classification of real EEG data lacks a definitive ground truth, with mean inter-rater agreement at 73.4% for onset and 64.2% for offset types. Despite this, DynamoSort achieved a mean area under the curve (AUC) of 0.81 for onset and a mean AUC of 0.75 for offset types. Machine-human agreement was not significantly different from human-to-human agreement. To address the lack of ground truth in ratings, DynamoSort assigns probabilistic scores (-20 to 20), to indicate similarity to each seizure dynamotype based on spiking features, allowing for a characterization of seizure dynamics on a spectrum rather than the traditional qualitative taxonomy. Significance Automating the classification of dynamotypes is a critical step for their inclusion as a biomarker in clinical and research applications. DynamoSort is a straightforward, open-access tool that uses automatic labelling and probabilistic scoring to quantify subtle changes in seizure onset and offset dynamics. Key Points Dynamotypes are a promising seizure categorization system, but is prone to interrater variability and lacks a ground truth. Machine learning can be used to automatically classify seizure onsets and offsets into appropriate dynamotypes based on spike features. Agreement between DynamoSort and human raters rivals typical agreement rates in trained human raters. DynamoSort uses probabilistic scoring to quantify subtle changes in seizure onset and offset, allowing for a quantitative characterisation.