EUVIMEDEUVIMED entdecken
Crossref

Neurofuse: Feature Fusion with Soft-Reset Spiking Networks for Efficient BCI

SURUTHEKA S

International Journal of Science, Strategic Management and Technology · 29.03.2026 · 02 (03) · 1-9

DOI: 10.55041/ijsmt.v2i3.366 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Science, Strategic Management and Technology, online erschienen am 29.03.2026. Verfasst von SURUTHEKA S.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate decoding of brain signals is vital for intra-cortical brain–computer interfaces (iBCIs). Traditional methods relying on hand-crafted neural activity features often lack accuracy, while deep learning approaches require high computational resources. In this paper (2025), we propose a spiking neural network (SNN) framework combined with a feature fusion strategy to achieve both efficiency and high decoding performance. Our model integrates manually extracted neural activity vector features with deep representations, enabling improved classification of motor-related signals. Experiments on rhesus macaque datasets show that the proposed method outperforms artificial neural network baselines in accuracy while being tens to hundreds of times more energy efficient, making it well-suited for real-world iBCIs applications.

Zitieren

APA

S, S. (2026). Neurofuse: Feature Fusion with Soft-Reset Spiking Networks for Efficient BCI. International Journal of Science, Strategic Management and Technology, 02(03), 1-9. https://doi.org/10.55041/ijsmt.v2i3.366

Vancouver

S S. Neurofuse: Feature Fusion with Soft-Reset Spiking Networks for Efficient BCI. International Journal of Science, Strategic Management and Technology. 2026;02(03):1-9. doi:10.55041/ijsmt.v2i3.366

BibTeX
@article{S2026,
  author = {S, SURUTHEKA},
  title = {Neurofuse: Feature Fusion with Soft-Reset Spiking Networks for Efficient BCI},
  journal = {International Journal of Science, Strategic Management and Technology},
  year = {2026},
  volume = {02},
  number = {03},
  pages = {1--9},
  doi = {10.55041/ijsmt.v2i3.366}
}
Kostenlos starten →