# Phoneme- vs. Character-Level Targets and Selective State-Space Models for Intracortical Brain-to-Text

Preprint · Lucas Zamora Vera, Jose A. Gonzalez-Lopez

- Archive id: arxiv:2607.26751
- Published: 2026-07-29
- arXiv: https://arxiv.org/abs/2607.26751
- Source URL: https://arxiv.org/abs/2607.26751
- PDF: https://arxiv.org/pdf/2607.26751
- Provenance: arxiv

## Abstract

State-of-the-art intracortical brain-to-text systems pair a neural-sequence phone decoder with an external language model. Two design axes remain underexplored: whether selective state-space models (Mamba) improve on recurrent decoders, and how the output target (phonetic vs.\\ character) interacts with that choice. On the public Brain-to-Text '25 benchmark, we study a controlled 2x2 grid (GRU vs.\\ hybrid Mamba decoder; phonetic vs.\\ character targets) trained with a CTC objective under one reproducible protocol. The recurrent baseline remains strongest: the best phonetic GRU reaches 12.62\\% PER and 21.19\\% WER, while the best textual GRU after LM rescoring reaches 13.39\\% CER and 26.28\\% WER. The Mamba hybrid is competitive but does not surpass it. Ablations isolate architectural contributions, and error analysis shows representation-dependent failures: articulatory-like phoneme confusions vs.\\ lexical and word-boundary errors.

_Abstract reproduced under CC0-1.0 (arXiv descriptive metadata)._

---
Space Kitty Archive · https://space-kitty.com · API: https://space-kitty.com/api/v1 · MCP: https://space-kitty.com/mcp · Premium: $0.20/request via x402 (Base Sepolia (testnet)). Metadata: arXiv (CC0), OpenAlex (CC0), Europe PMC (CC BY/CC0 works), ClinicalTrials.gov.