Brain to text decoding runs fast and accurately on mobile devices
FETTUCCINE: Fast and efficient brain-to-text decoding on mobile devices
Machine Learning
Summary
People who can no longer speak due to illness might use brain signals to communicate, but current systems are slow and need powerful computers. The authors developed FETTUCCINE, a new method that quickly turns brain signals into written words using speech recognition technology, and can run on a phone. Their tests show it is accurate and can keep working well over several days. This method also helps locate which brain signals are most important for decoding speech.
What this means in practice
- •For speech neuroprosthesis developers: Create portable brain-to-text communication devices for users with speech impairments that run efficiently on common mobile hardware.
- •For mobile device engineers: Integrate brain-signal decoding models into smartphones to enable real-time neural communication interfaces without cloud reliance.
Authors
Jonathan D. McCart, Pranav I. Deevi, Mehdi Azabou, Nanda H. Krishna, Caleb A. McKinney, Nicholas S. Card, Sergey D. Stavisky, Chethan Pandarinath
Abstract
Recent advances in speech neuroprostheses have enabled new communication avenues for people who have lost the ability to speak due to neurological illness or injury. These systems can decode neural activity during attempted speech into text by relying on language priors to achieve high decoding accuracy. This performance currently comes at the cost of using a compute-intensive language modeling stack, presenting a major bottleneck for real-world use. Further, current language models incur highly variable and disruptive inference latencies, limiting communication throughput and preventing users from reliably participating in natural conversation. We introduce FETTUCCINE, an end-to-end brain-to-text framework that leverages the speed and portability of automatic speech recognition (ASR) models to address the aforementioned practical challenges, while achieving usable (<10%) word error rates (WER), efficient adaptation to neural non-stationarities, and reliable performance on mobile devices. When evaluated using publicly available data from an intracortical speech neuroprosthesis user, FETTUCCINE outperforms previous end-to-end methods, achieving as low as 4.39% WER. Most notably, our models successfully run on a commercially-available mobile device with throughput reaching up to 370x real-time, providing the first demonstration of brain-to-text decoding on a mobile device. We also show that our models can be successfully finetuned to extend usable performance to future days. Beyond performance, our approach enables localization of the neural features most relevant for decoding via gradient-based saliency maps, which we show align with well-established physiological priors across different ASR model families. Taken together, our results show that FETTUCCINE overcomes core infrastructural and computational barriers, yielding a new class of portable brain-to-text communication.