mlsanityData to 14 Oct 2026

Neuron-Level Architecture Growth: A Controlled Evaluation for EEG Time-Series Decoding

arXiv 2609.33880 ↗First posted 27 Sep 2026eess.SP cs.HC cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Convolutional EEG decoders are trained at a fixed width, usually set by their authors on other data. Growing methods add neurons during training where the loss could decrease the most, but whether they improve compared to a reference width is untested on EEG. Here, we grow three convolutional backbones on 12 motor-imagery datasets under three protocols and compare each with its reference model per subject. The growing ShallowFBCSPNet scores 2.9 points above its reference model with only half the parameters (0.57x), SCCNet changes by at most 1.2 points. Deep4Net growing models show decreased accuracy, but they require adaptation that prevent to compare faithfully the results. These differences follow the selection step, which keeps a candidate neuron relying on a dynamic threshold from singular values decomposition. Overall, these results suggest that growth helps when its criterion can rank the candidate neurons, and that the rate of skipped neuron addition tells where a decoder can be grown small from scratch.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 20 of its 33 references that are in the map. First posted on arXiv 27 Sep 2026.

Builds on

20 of 33 references.

References · 20

  1. 2016EEGNet: A Compact Convolutional Network for EEG-based Brain-Computer Interfacescited by 628
  2. 2015Net2Net: Accelerating Learning via Knowledge Transfercited by 250
  3. 2018MOABB: Trustworthy algorithm benchmarking for BCIscited by 82
  4. 2022GradMax: Growing Neural Networks using Gradient Informationcited by 66
  5. 2019Splitting Steepest Descent for Growing Neural Architecturescited by 55
  6. 2021Firefly Neural Architecture Descent: a General Approach for Growing Neural Networkscited by 49
  7. 2025EEG Foundation Models: A Critical Review of Current Progress and Future Directionscited by 32
  8. 2024The largest EEG-based BCI reproducibility study for open science: the MOABB benchmarkcited by 29
  9. 2025Are Large Brainwave Foundation Models Capable Yet? Insights from Fine-tuningcited by 15
  10. 2024A Systematic Evaluation of Euclidean Alignment with Deep Learning for EEG Decodingcited by 12
  11. 2024Growing Tiny Networks: Spotting Expressivity Bottlenecks and Fixing Them Optimallycited by 9
  12. 2023Evaluating the structure of cognitive tasks with transfer learningcited by 8
8 more
  1. 2022MI-BMInet: An Efficient Convolutional Neural Network for Motor Imagery Brain--Machine Interfaces with EEG Channel Selectioncited by 7
  2. 2021Minimizing subject-dependent calibration for BCI with Riemannian transfer learningcited by 6
  3. 2025Adaptive Width Neural Networkscited by 4
  4. 2023Deep comparisons of Neural Networks from the EEGNet familycited by 3
  5. 2024When To Grow? A Fitting Risk-Aware Policy for Layer Growing in Deep Neural Networkscited by 3
  6. 2022Cross Task Neural Architecture Search for EEG Signal Classificationscited by 2
  7. 2025Growth strategies for arbitrary DAG neural architecturescited by 2
  8. 2024Towards a Network Expansion Approach for Reliable Brain-Computer Interfacecited by 1

Built on by

Not cited yet.