Loading…

Advances in Epilepsy Monitoring by Detection and Analysis of Brain Epileptiform Discharges

Objective: Brain interictal and preictal epileptiform discharges (EDs) are transient events occurring between two or before seizure onsets visible in intracranial electroencephalogram. In the diagnosis of epilepsy and localization of seizure sources, both interictal and ictal recordings are extremel...

Full description

Saved in:
Bibliographic Details
Published in:Psychology & Neuroscience 2022-12, Vol.15 (4), p.375-394
Main Authors: Abdi-Sargezeh, Bahman, Sanei, Saeid
Format: Article
Language:English
Subjects:
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Objective: Brain interictal and preictal epileptiform discharges (EDs) are transient events occurring between two or before seizure onsets visible in intracranial electroencephalogram. In the diagnosis of epilepsy and localization of seizure sources, both interictal and ictal recordings are extremely informative. For this purpose, computerized intelligent spike and seizure detection techniques have been researched and are constantly improving. This is not only to detect more EDs from over the scalp but also to classify epileptic and non-epileptic discharges. Method: Tensor factorization and deep learning are two advanced and powerful techniques which have been recently suggested for ED detection. Here, our main contribution is to review recent ED detection methods with emphasis on multiway analysis and deep learning approaches. Results: The performance measures (accuracy, sensitivity, specificity, etc.) of the applied methods are reported. Different researchers use different data, making the performance comparison difficult and sometimes impossible given that, unlike many other approaches, in our work we identify both scalp-visible and scalp-invisible EDs. Conclusions: It is shown that both multi-way analysis techniques (namely, Tucker and CANDECOMP/PARAFAC decomposition methods) and deep learning (particularly convolutional neural networks and long short-term memory) are powerful in identifying the EDs. These techniques have opened a new window to the epilepsy diagnosis and management spheres. Public Significance Statement The outcome of this research has significant impact on seizure diagnosis as it enables detection of invisible epileptiform spikes from scalp EEGs without any invasive operation. This potentially benefits over 65 million people worldwide suffering from epilepsy.
ISSN:1984-3054
1983-3288
DOI:10.1037/pne0000275