Neuromorphic deep spiking neural networks for seizure detection
The vast majority of studies that process and analyze neural signals are conducted on cloud computing resources, which is often necessary for the demanding requirements of deep neural network workloads. However, applications such as epileptic seizure detection stand to benefit from edge devices that...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IOP Publishing
2023-01-01
|
| Series: | Neuromorphic Computing and Engineering |
| Subjects: | |
| Online Access: | https://doi.org/10.1088/2634-4386/acbab8 |
| _version_ | 1827964987445870592 |
|---|---|
| author | Yikai Yang Jason K Eshraghian Nhan Duy Truong Armin Nikpour Omid Kavehei |
| author_facet | Yikai Yang Jason K Eshraghian Nhan Duy Truong Armin Nikpour Omid Kavehei |
| author_sort | Yikai Yang |
| collection | DOAJ |
| description | The vast majority of studies that process and analyze neural signals are conducted on cloud computing resources, which is often necessary for the demanding requirements of deep neural network workloads. However, applications such as epileptic seizure detection stand to benefit from edge devices that can securely analyze sensitive medical data in a real-time and personalised manner. In this work, we propose a novel neuromorphic computing approach to seizure detection using a surrogate gradient-based deep spiking neural network (SNN), which consists of a novel spiking ConvLSTM unit. We have trained, validated, and rigorously tested the proposed SNN model across three publicly accessible datasets, including Boston Children’s Hospital–MIT (CHB-MIT) dataset from the U.S., and the Freiburg (FB) and EPILEPSIAE intracranial electroencephalogram datasets from Germany. The average leave-one-out cross-validation area under the curve score for FB, CHB-MIT and EPILEPSIAE datasets can reach 92.7 $\%$ , 89.0 $\%$ , and 81.1 $\%$ , respectively, while the computational overhead and energy consumption are significantly reduced when compared to alternative state-of-the-art models, showing the potential for building an accurate hardware-friendly, low-power neuromorphic system. This is the first feasibility study using a deep SNN for seizure detection on several reliable public datasets. |
| first_indexed | 2024-04-09T17:24:48Z |
| format | Article |
| id | doaj.art-4a04f8e6e6f04deabbe59c85bf0c9798 |
| institution | Directory Open Access Journal |
| issn | 2634-4386 |
| language | English |
| last_indexed | 2024-04-09T17:24:48Z |
| publishDate | 2023-01-01 |
| publisher | IOP Publishing |
| record_format | Article |
| series | Neuromorphic Computing and Engineering |
| spelling | doaj.art-4a04f8e6e6f04deabbe59c85bf0c97982023-04-18T13:54:07ZengIOP PublishingNeuromorphic Computing and Engineering2634-43862023-01-013101401010.1088/2634-4386/acbab8Neuromorphic deep spiking neural networks for seizure detectionYikai Yang0Jason K Eshraghian1Nhan Duy Truong2Armin Nikpour3Omid Kavehei4https://orcid.org/0000-0002-2753-5553School of Biomedical Engineering, Faculty of Engineering, The University of Sydney , Sydney, NSW 2006, AustraliaThe Department of Electrical and Computer Engineering, University of California Santa Cruz , Santa Cruz, CA 95064, United States of AmericaSchool of Biomedical Engineering, Faculty of Engineering, The University of Sydney , Sydney, NSW 2006, Australia; The University of Sydney Nano Institute , Sydney, NSW 2006, AustraliaComprehensive Epilepsy Service and Department of Neurology, Royal Prince Alfred Hospital , Sydney, NSW 2050, Australia; Faculty of Medicine and Health, Central Clinical School, The University of Sydney , Sydney, NSW 2006, AustraliaSchool of Biomedical Engineering, Faculty of Engineering, The University of Sydney , Sydney, NSW 2006, Australia; The University of Sydney Nano Institute , Sydney, NSW 2006, AustraliaThe vast majority of studies that process and analyze neural signals are conducted on cloud computing resources, which is often necessary for the demanding requirements of deep neural network workloads. However, applications such as epileptic seizure detection stand to benefit from edge devices that can securely analyze sensitive medical data in a real-time and personalised manner. In this work, we propose a novel neuromorphic computing approach to seizure detection using a surrogate gradient-based deep spiking neural network (SNN), which consists of a novel spiking ConvLSTM unit. We have trained, validated, and rigorously tested the proposed SNN model across three publicly accessible datasets, including Boston Children’s Hospital–MIT (CHB-MIT) dataset from the U.S., and the Freiburg (FB) and EPILEPSIAE intracranial electroencephalogram datasets from Germany. The average leave-one-out cross-validation area under the curve score for FB, CHB-MIT and EPILEPSIAE datasets can reach 92.7 $\%$ , 89.0 $\%$ , and 81.1 $\%$ , respectively, while the computational overhead and energy consumption are significantly reduced when compared to alternative state-of-the-art models, showing the potential for building an accurate hardware-friendly, low-power neuromorphic system. This is the first feasibility study using a deep SNN for seizure detection on several reliable public datasets.https://doi.org/10.1088/2634-4386/acbab8deep spiking neural networkseizure detectionneuromorphic computing |
| spellingShingle | Yikai Yang Jason K Eshraghian Nhan Duy Truong Armin Nikpour Omid Kavehei Neuromorphic deep spiking neural networks for seizure detection Neuromorphic Computing and Engineering deep spiking neural network seizure detection neuromorphic computing |
| title | Neuromorphic deep spiking neural networks for seizure detection |
| title_full | Neuromorphic deep spiking neural networks for seizure detection |
| title_fullStr | Neuromorphic deep spiking neural networks for seizure detection |
| title_full_unstemmed | Neuromorphic deep spiking neural networks for seizure detection |
| title_short | Neuromorphic deep spiking neural networks for seizure detection |
| title_sort | neuromorphic deep spiking neural networks for seizure detection |
| topic | deep spiking neural network seizure detection neuromorphic computing |
| url | https://doi.org/10.1088/2634-4386/acbab8 |
| work_keys_str_mv | AT yikaiyang neuromorphicdeepspikingneuralnetworksforseizuredetection AT jasonkeshraghian neuromorphicdeepspikingneuralnetworksforseizuredetection AT nhanduytruong neuromorphicdeepspikingneuralnetworksforseizuredetection AT arminnikpour neuromorphicdeepspikingneuralnetworksforseizuredetection AT omidkavehei neuromorphicdeepspikingneuralnetworksforseizuredetection |