# CIPHER: Conformer-based Inference of Phonemes from High-density EEG

arXiv: https://arxiv.org/abs/2604.02362 · HTML: https://mlsanity.com/paper/2604.02362/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 33 references that are in the map.
- First posted on arXiv 22 Mar 2026.

## Builds on (most cited first, 5 of 5)

- [mixup: Beyond Empirical Risk Minimization](https://mlsanity.com/paper/1710.09412/) (2017)
- [Squeeze-and-Excitation Networks](https://mlsanity.com/paper/1709.01507/) (2017)
- [Conformer: Convolution-augmented Transformer for Speech Recognition](https://mlsanity.com/paper/2005.08100/) (2020)
- [EEGNet: A Compact Convolutional Network for EEG-based Brain-Computer Interfaces](https://mlsanity.com/paper/1611.08024/) (2016)
- [Decoding speech perception from non-invasive brain recordings](https://mlsanity.com/paper/2208.12266/) (2022)

## Built on by (most cited first, 1 of 1)

- [Leakage-Audited Benchmarking Reveals Limited Evidence for Cross-Subject Auditory-Evoked EEG Vowel Perception Decoding](https://mlsanity.com/paper/2605.00865/) (2026)

## Abstract (arXiv)

Decoding speech information from scalp EEG remains difficult due to low SNR and spatial blurring. We present CIPHER (Conformer-based Inference of Phonemes from High-density EEG Representations), a dual-pathway model using (i) ERP features and (ii) broadband DDA coefficients. On OpenNeuro ds006104 (24 participants, two studies with concurrent TMS), binary articulatory tasks reach near-ceiling performance but are highly confound-vulnerable (acoustic onset separability and TMS-target blocking). On the primary 11-class CVC phoneme task under full Study 2 LOSO (16 held-out subjects), performance is substantially lower (real-word WER: ERP 0.671 +/- 0.080, DDA 0.688 +/- 0.096, indicating limited fine-grained discriminability. We therefore position this work as a benchmark and feature-comparison study rather than an EEG-to-text system, and we constrain neural-representation claims to confound-controlled evidence.