Detecting fast-ripples on both micro- and macro-electrodes in epilepsy: A wavelet-based CNN detector.

Background: Fast-ripples (FR) are short (∼10 ms) high-frequency oscillations (HFO) between 200 and 600 Hz that are helpful in epilepsy to identify the epileptogenic zone. Our aim is to propose a new method to detect FR that had to be efficient for intracerebral EEG (iEEG) recorded from both usual clinical macro-contacts (millimeter scale) and microwires (micrometer scale).
New Method: Step 1 of the detection method is based on a convolutional neural network (CNN) trained using a large database of > 11,000 FR recorded from the iEEG of 38 patients with epilepsy from both macro-contacts and microwires. The FR and non-FR events were fed to the CNN as normalized time-frequency maps. Step 2 is based on feature-based control techniques in order to reject false positives. In step 3, the human is reinstated in the decision-making process for final validation using a graphical user interface.
Results: WALFRID achieved high performance on the realistically simulated data with sensitivity up to 99.95 % and precision up to 96.51 %. The detector was able to adapt to both macro and micro-EEG recordings. The real data was used without any pre-processing step such as artefact rejection. The precision of the automatic detection was of 57.5. Step 3 helped eliminating remaining false positives in a few minutes per subject.
Comparison With Existing Methods: WALFRID performed as well or better than 6 other existing methods.
Conclusion: Since WALFRID was created to mimic the work-up of the neurologist, clinicians can easily use, understand, interpret and, if necessary, correct the output.
Competing Interests: Declaration of Competing Interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Emmanuel J. Barbeau reports a relationship with Avrio MedTech that includes: equity or stocks. Christophe Hurter reports a relationship with Avrio MedTech that includes: equity or stocks. Ludovic Gardy reports a relationship with Avrio MedTech that includes: equity or stocks. Emmanuel J. Barbeau has patent #WO2023067114A1 pending to Avrio MedTech. Christophe Hurter has patent #WO2023067114A1 pending to Avrio MedTech. Ludovic Gardy has patent #WO2023067114A1 pending to Avrio MedTech. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
(Copyright © 2024 The Authors. Published by Elsevier B.V. All rights reserved.)