# arXiv q-bio.NC: Evolved Spiking Neural Networks for EEG Seizure Detection [Canonical reading](https://escapemeridian.com/readings/arxiv-q-bio-nc-evolved-spiking-neural-networks-for-eeg-seizure-057d9579) In her paper’s abstract, Tuba Aksoy reports that EONS, a framework that evolves spiking-network designs without gradients, found improvements in EEG seizure detection with per-neuron calcium parameters or evolved synaptic delays. On 10 held-out CHB-MIT patients, per-neuron calcium raised record-level Matthews correlation coefficient (MCC) from 0.592 to 0.726 and lowered the false-alarm rate from 0.17 to 0.10 across 20 independent runs. Fixed or shared calcium parameters showed no significant benefit, and axonal delays did not improve results. On Siena, the reported MCC gains were +0.096 in one patient-disjoint split and +0.059 in another; the second result had p = 0.052. Aksoy says these results suggest calcium and delays may make the evolutionary search more effective, and motivate testing on more data and in neuromorphic hardware. These findings concern the reported datasets and tests; the abstract reports no clinical-use results. Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims. ## Original source arXiv q-bio.NC · Article Original source date: Nov 2, 2026 (2026-11-02). Date basis: Date basis not provided. Added to Escape Meridian: 2026-11-10T05:01:14.895Z. First observed: 2026-11-04T05:12:32.344Z. Retrieved: 2026-11-04T05:31:00.664Z. - [Evolved Spiking Neural Networks for EEG Seizure Detection: Calcium Dynamics and Delay Mechanics](https://arxiv.org/abs/2611.03841) ## Limits The abstract reports results on 10 held-out CHB-MIT patients and two patient-disjoint Siena splits; it does not report tests on more data or in neuromorphic hardware. The abstract reports seizure-detection metrics, not performance in clinical use. ## Open question What results would testing on more data and in neuromorphic hardware produce? ## Source-stated Aksoy reports that EONS evolves neuromorphic spiking-network topology, weights, thresholds, and delays without gradients. > The EONS framework evolves the topology, weights, thresholds and delays of neuromorphic spiking networks without gradients. - [Supporting source](https://arxiv.org/abs/2611.03841) ## Source-stated On 10 held-out CHB-MIT patients, per-neuron calcium raised record-level MCC from 0.592 to 0.726 and lowered the false-alarm rate from 0.17 to 0.10 over 20 independent runs. > On 10 held-out CHB-MIT patients never seen in training, per-neuron calcium raised the record-level Matthews correlation coefficient (MCC) from 0.592 to 0.726 over 20 independent runs (p = 0.008) and lowered the false-alarm rate from 0.17 to 0.10. - [Supporting source](https://arxiv.org/abs/2611.03841) ## Source-stated Aksoy reports no significant benefit from fixed or per-network calcium, and a similar gain from IF networks with evolved synaptic delays of 1–3 ticks; axonal delays did not help. > Fixed or per-network calcium gave no significant benefit. IF networks with evolved synaptic delays of 1-3 ticks gave a similar gain (p = 0.005), whereas axonal delays did not. - [Supporting source](https://arxiv.org/abs/2611.03841) ## Source-stated On Siena, Aksoy reports MCC gains over IF of +0.096 in one patient-disjoint split and +0.059 in the other, with p = 0.052 for the latter. > On Siena, with unchanged settings, per-neuron calcium gained +0.096 MCC over IF in one patient-disjoint split (p = 0.011) and +0.059 in the other (p = 0.052). - [Supporting source](https://arxiv.org/abs/2611.03841) ## Source-stated Aksoy says the results suggest calcium and delays may make evolutionary search more effective, and motivate testing on more data and in neuromorphic hardware. > These results suggest that calcium and delays may make the evolutionary search more effective, and motivate testing on more data and in neuromorphic hardware. - [Supporting source](https://arxiv.org/abs/2611.03841) ## Source and contributor context Aksoy, Tuba · Article author These are credits stated by the source; they do not establish a profile match or authorship of every passage. ## Source media No suitable media was found by the supported page extraction. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Neural network A system of connected simple calculations, loosely inspired by the brain, that learns patterns from data. - [Definition reference](https://developers.google.com/machine-learning/glossary#neural-network) ## arXiv arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed. - [Explore this context](/entities/arxiv) - [Reference](https://arxiv.org/) - [Reference](https://info.arxiv.org/help/api/index.html) ## Editorial provenance Version 1. Publication ID: f576d519-7a46-4f07-8a29-1344ac5a181e. Prepared by gpt-6-luna; grounding review by gpt-6-luna. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-10T05:01:14.895Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: material\_development. Observation: eba7c5244930f7adccb62ffb8de20e7ad390d05cb0b84a8e31f8cb340124273c. Source text hash: 13a7ada95ce148a22cc5d092d069b4f1413c320b03262b26699278393b128b1c. ## Revision and source-check history ### Version 1 — 2026-11-10T05:01:14.895Z In her paper’s abstract, Tuba Aksoy reports that EONS, a framework that evolves spiking-network designs without gradients, found improvements in EEG seizure detection with per-neuron calcium parameters or evolved synaptic delays. On 10 held-out CHB-MIT patients, per-neuron calcium raised record-level Matthews correlation coefficient (MCC) from 0.592 to 0.726 and lowered the false-alarm rate from 0.17 to 0.10 across 20 independent runs. Fixed or shared calcium parameters showed no significant benefit, and axonal delays did not improve results. On Siena, the reported MCC gains were +0.096 in one patient-disjoint split and +0.059 in another; the second result had p = 0.052. Aksoy says these results suggest calcium and delays may make the evolutionary search more effective, and motivate testing on more data and in neuromorphic hardware. These findings concern the reported datasets and tests; the abstract reports no clinical-use results. ## Original source arXiv q-bio.NC · Article Original source date: Nov 2, 2026 (2026-11-02). Date basis: Date basis not provided. Added to Escape Meridian: 2026-11-10T05:01:14.895Z. First observed: 2026-11-04T05:12:32.344Z. Retrieved: 2026-11-04T05:31:00.664Z. - [Evolved Spiking Neural Networks for EEG Seizure Detection: Calcium Dynamics and Delay Mechanics](https://arxiv.org/abs/2611.03841) ## Limits The abstract reports results on 10 held-out CHB-MIT patients and two patient-disjoint Siena splits; it does not report tests on more data or in neuromorphic hardware. The abstract reports seizure-detection metrics, not performance in clinical use. ## Open question What results would testing on more data and in neuromorphic hardware produce? ## Source-stated Aksoy reports that EONS evolves neuromorphic spiking-network topology, weights, thresholds, and delays without gradients. > The EONS framework evolves the topology, weights, thresholds and delays of neuromorphic spiking networks without gradients. - [Supporting source](https://arxiv.org/abs/2611.03841) ## Source-stated On 10 held-out CHB-MIT patients, per-neuron calcium raised record-level MCC from 0.592 to 0.726 and lowered the false-alarm rate from 0.17 to 0.10 over 20 independent runs. > On 10 held-out CHB-MIT patients never seen in training, per-neuron calcium raised the record-level Matthews correlation coefficient (MCC) from 0.592 to 0.726 over 20 independent runs (p = 0.008) and lowered the false-alarm rate from 0.17 to 0.10. - [Supporting source](https://arxiv.org/abs/2611.03841) ## Source-stated Aksoy reports no significant benefit from fixed or per-network calcium, and a similar gain from IF networks with evolved synaptic delays of 1–3 ticks; axonal delays did not help. > Fixed or per-network calcium gave no significant benefit. IF networks with evolved synaptic delays of 1-3 ticks gave a similar gain (p = 0.005), whereas axonal delays did not. - [Supporting source](https://arxiv.org/abs/2611.03841) ## Source-stated On Siena, Aksoy reports MCC gains over IF of +0.096 in one patient-disjoint split and +0.059 in the other, with p = 0.052 for the latter. > On Siena, with unchanged settings, per-neuron calcium gained +0.096 MCC over IF in one patient-disjoint split (p = 0.011) and +0.059 in the other (p = 0.052). - [Supporting source](https://arxiv.org/abs/2611.03841) ## Source-stated Aksoy says the results suggest calcium and delays may make evolutionary search more effective, and motivate testing on more data and in neuromorphic hardware. > These results suggest that calcium and delays may make the evolutionary search more effective, and motivate testing on more data and in neuromorphic hardware. - [Supporting source](https://arxiv.org/abs/2611.03841) ## Source and contributor context Aksoy, Tuba · Article author These are credits stated by the source; they do not establish a profile match or authorship of every passage. ## Source media No suitable media was found by the supported page extraction. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Neural network A system of connected simple calculations, loosely inspired by the brain, that learns patterns from data. - [Definition reference](https://developers.google.com/machine-learning/glossary#neural-network) ## arXiv arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed. - [Explore this context](/entities/arxiv) - [Reference](https://arxiv.org/) - [Reference](https://info.arxiv.org/help/api/index.html) ## Editorial provenance Version 1. Publication ID: f576d519-7a46-4f07-8a29-1344ac5a181e. Prepared by gpt-6-luna; grounding review by gpt-6-luna. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-10T05:01:14.895Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: material\_development. Observation: eba7c5244930f7adccb62ffb8de20e7ad390d05cb0b84a8e31f8cb340124273c. Source text hash: 13a7ada95ce148a22cc5d092d069b4f1413c320b03262b26699278393b128b1c. ### uncertain — 2026-11-10T05:01:14.895Z — anthropic/claude-opus-5.5 Publication: f576d519-7a46-4f07-8a29-1344ac5a181e attribution — pass: The work is correctly attributed to sole author Tuba Aksoy on arXiv q-bio.NC, and results are framed as her reported findings. One minor note: the draft says 'her paper', but the source does not state the author's pronouns. A small repair is 'In the paper's abstract, Tuba Aksoy reports...'. > Authors:Tuba Aksoy summary — uncertain: The numbers are faithful, but a non-specialist cannot follow the central idea. The draft never explains spiking neural networks, what an activity-dependent calcium current adds to integrate-and-fire (IF) neurons, what MCC measures, what synaptic versus axonal delays are, or what neuromorphic hardware is. The baseline 'IF' is used in the source but never defined in the summary. The phrase 'EONS ... found improvements' blurs who did what: the author tested Ca-IF neurons within EONS. Repair: add plain-language glosses or tooltips, and say the author tested adding calcium to neurons evolved by EONS. > We tested whether giving integrate-and-fire (IF) neurons an activity-dependent calcium current (Ca-IF neurons) improves evolved networks for EEG seizure detection date — pass: The originalPublished value of 2026-11-02 matches the v1 submission date. The later feed date of 2026-11-04 was correctly not used. > \[Submitted on 2 Nov 2026\] support — pass: Each claim matches the abstract verbatim, and the limitations faithfully note that the study has no clinical-use results and no hardware testing. The p = 0.052 caveat is preserved, as is the hedged 'may' in the conclusion. No implied benefits, costs or stakeholders were invented. The question follows directly from the author's stated next steps. > These results suggest that calcium and delays may make the evolutionary search more effective, and motivate testing on more data and in neuromorphic hardware. media — pass: The selected media is the verified direct arXiv abstract page for this paper, which is the most direct access to the source. ## Recorded references Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter. No public references recorded here yet. [Reference data (JSON)](/readings/057d9579c8713ac7e110c9d047c6fa2837b4098aa559a818066585041ff70abb/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/arxiv-q-bio-nc-evolved-spiking-neural-networks-for-eeg-seizure-057d9579.json)