Decoding continuous speech from intracortical recordings is a central challenge for brain-computer interfaces (BCIs), with transformative potential for individuals with conditions that impair their ability to speak. While recent microelectrode array (MEA) decoders achieve impressive accuracy, their performance is fundamentally limited by the small size of existing datasets, they remain brittle to session-to-session variability, and their ability to generalize across participants remains unexplored. We introduce BrainWhisperer, a neural speech decoder that integrates high-resolution MEA recordings with a large pretrained automatic speech recognition (ASR) model. Building on interpretability findings showing that Whisper's encoder learns phoneme-selective representations with localized attention, we train a customized version of Whisper, modified to process neural features, using a hybrid objective that combines CTC loss on phonemes--predicted from the third encoder layer--and cross-entropy loss on word tokens. We introduce domain-informed modifications including windowed self-attention to capture articulatory continuity, hierarchical month/day-specific low-rank projections to address non-stationarity, and subject-specific embedders enabling cross-subject training. Evaluated on a publicly available MEA dataset (Card et al.), BrainWhisperer matches or outperforms prior state-of-the-art decoders. Critically, cross-dataset training improves performance even on individual datasets without fine-tuning, demonstrating unprecedented generalization. The model supports dual decoding paths: a high-accuracy phoneme-based path with external language model rescoring, and a fast direct text generation path enabling sub-100ms inference with minimal hardware requirements.
Figures & tables
Figure 1: A. Our decoding pipelines: the low-compute end-to-end branch uses an encoder–decoder architecture with beam search, while the high-compute branch feeds phoneme logits into a weighted finite-state transducer to generate candidates, which are then rescored. B. BrainWhisperer’s architecture. Neural features are first mapped into a shared embedding space using a convolutional front-end (embedder) and then processed by a stack of transformer encoder layers. A phoneme head predicts phoneme sequences using a CTC loss, while the remaining encoder layers feed a transformer decoder trained with cross-entropy on text tokens. C. The proposed month-specific and day-specific low-rank projections. D. Our windowed attention mechanism. Windowing is applied to the phoneme-selective encoder layers.
Table 1: Accuracy in terms of phoneme error rate (PER) and word error rate (WER) for the evaluated models. Results are shown for both the test and competition splits of the Card dataset. The 5-gram subscript indicates that extensive post-processing with a 5-gram model and rescoring was applied.
Table 2: Examples of decoded transcriptions. Each example reports the ground truth transcription, the one predicted by the model and the corresponding WER. We selected examples associated with the 10 th , 20 th , …, 100 th percentiles. Errors are highlighted in red.
Speech brain--computer interfaces require decoders that translate intracortical activity into linguistic output while remaining robust to limited data and day-to-day variability. While prior high-performing systems have largely relied on framewise phoneme decoding combined with downstream language models, it remains unclear what contextual sequence-to-sequence decoding contributes to sublexical neural readout, robustness, and interpretability. We evaluated a multitask Transformer-based sequence-to-sequence model for attempted speech decoding from area 6v intracortical recordings. The model jointly predicts phoneme sequences, word sequences, and auxiliary acoustic features. To address day-to-day nonstationarity, we introduced the Neural Hammer Scalpel (NHS) calibration module, which combines global alignment with feature-wise modulation. We further analyzed held-out-day generalization and attention patterns in the encoder and decoders. On the Willett et al. dataset, the proposed model achieved a state-of-the-art phoneme error rate of 14.3%. Word decoding reached 25.6% WER with direct decoding and 19.4% WER with candidate generation and rescoring. NHS substantially improved both phoneme and word decoding relative to linear or no day-specific transform, while held-out-day experiments showed increasing degradation on unseen days with temporal distance. Attention visualizations revealed recurring temporal chunking in encoder representations and distinct use of these segments by phoneme and word decoders. These results indicate that contextual sequence-to-sequence modeling can improve the fidelity of neural-to-phoneme readout from intracortical speech signals and suggest that attention-based analyses can generate useful hypotheses about how neural speech evidence is segmented and accumulated over time.
Recent speech brain-computer interfaces (BCIs) increasingly rely on large models and complex multi-stage pipelines for neural decoding. However, in speech BCI, where training data are scarce, noisy, and expensive to collect, the benefits of model scaling remain unclear. This work revisits a fundamental question: are large models truly necessary for accurate neural decoding? We propose BEST (Brain-Encoding-Speech-Text), a two-stage framework for efficient neural-to-text decoding. Instead of relying on large generative models or phoneme-centric pipelines, BEST introduces hierarchical intermediate representations and a compact speech-oriented decoder that directly maps neural activity to text. This design preserves essential structure while significantly reducing computational cost, enabling robust generalization in low-resource settings and practical deployment under constrained resources. On the Brain-to-Text '24 and '25 benchmarks, BEST achieves word error rates (WERs) of 6.19% and 1.62%, respectively, using only 3.85% of the parameters required by comparable methods. Beyond these results, we provide three key insights: (1) decoder scaling yields non-monotonic gains; (2) cross-modal alignment is more critical than model size; and (3) overfitting degards sequence-level performance. Representation analysis further reveals a consistent neural-audio-text alignment pathway.
Zehui Feng, Weichuan Wang, Xiaohan Chen +2
Shanghai Jiao Tong University · Nanyang Technological University · Zhejiang University
The loss of speech limits communication for individuals with paralysis. Direct neural-to-speech synthesis is challenging due to the limited availability of neural data for training speech brain-computer interfaces. Most existing systems rely on cascaded neural-to-text-to-speech pipelines, which increase inference latency and propagate errors across stages. We present Brain2Speech-Net, a single-stage neural-to-speech generation framework without intermediate text decoding. We use a differentiable phoneme bottleneck and a deep-HMM alignment mechanism to map long neural recordings into the latent space of a text-to-speech (TTS) model, enabling high-quality speech synthesis. Brain2Speech-Net is the only system in our comparison that produces intelligible speech while generating faster than real time.
Shreeram Suresh Chandra, Zexin Cai, Yu Tsao +2
Johns Hopkins University, USA · Academia Sinica, Taiwan · University of Edinburgh, UK