Removing Timing Shortcuts Improves Non-Invasive Brain-to-Text
Abstract
We find that major reported improvements in decoding words from non-invasive brain recordings are largely reproducible without any brain data. In the influential work of d'Ascoli et al. (2025), time series of brain activity from subjects perceiving continuous speech are segmented into fixed-length windows starting at each word. A neural network then generates predictions for all of the words in a sentence together. Neighbouring windows partially overlap, implicitly revealing the interval between words. Since these intervals indicate the duration of the words spoken, and different words tend to have different durations - for example, "the" is much shorter than "supercalifragilisticexpialidocious" - the neural network can improve its predictions of words without relying on the underlying brain activity. Consistent with this, the method reaches 22.0% balanced accuracy on synthetic signals containing no brain information, compared with 22.3% on real brain recordings. To prevent the network from learning this shortcut, we make a single, simple change. Instead of jointly encoding all windows in a sentence, we process each independently. As a result, the neural network achieves better performance by learning underlying word-specific information from brain recordings. This makes two existing strategies become much more effective than before. Both aggregating predictions from distinct neural responses to the same word and using a pretrained LLM as a linguistic prior now substantially improve results. On our perceived speech benchmark, this simple recipe (SimpleB2T) achieves a word error rate of 36.6% with five observations per word, approaching past invasive speech decoding performance, albeit under different conditions. The results in this work expose an important shortcut in brain-to-text decoding and show that removing it leads to a simple and considerably more effective strategy.
Related
- Mapping the neural basis of the perception of “flow of time”: interval timing and mental time travel in the human brain
- Pubertal timing is a critical dimension of adolescent brain vulnerability
- Forniceal deep brain stimulation improves memory in Rett syndrome mice by promoting the maturation of newborn hippocampal neurons
- Epigenetic mechanisms of Non-Invasive Brain Stimulation: A Systematic Review of Human and Preclinical Evidence
- Targeting the RNA regulator HuR in acute traumatic brain injury mitigates secondary injury and improves functional outcome by suppressing neuroinflammation
BibTeX
@misc{jayalath2026removing,
title = {{Removing Timing Shortcuts Improves Non-Invasive Brain-to-Text}},
author = {Dulhan Jayalath and Oiwi Parker Jones},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.40359},
eprint = {2609.40359},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.40359},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Removing Timing Shortcuts Improves Non-Invasive Brain-to-Text
AU - Dulhan Jayalath
AU - Oiwi Parker Jones
PY - 2026
DA - 2026/09/30
PB - arXiv
DO - 10.48550/arxiv.2609.40359
UR - https://arxiv.org/abs/2609.40359
AB - We find that major reported improvements in decoding words from non-invasive brain recordings are largely reproducible without any brain data. In the influential work of d'Ascoli et al. (2025), time series of brain activity from subjects perceiving continuous speech are segmented into fixed-length windows starting at each word. A neural network then generates predictions for all of the words in a sentence together. Neighbouring windows partially overlap, implicitly revealing the interval between words. Since these intervals indicate the duration of the words spoken, and different words tend to have different durations - for example, "the" is much shorter than "supercalifragilisticexpialidocious" - the neural network can improve its predictions of words without relying on the underlying brain activity. Consistent with this, the method reaches 22.0% balanced accuracy on synthetic signals containing no brain information, compared with 22.3% on real brain recordings. To prevent the network from learning this shortcut, we make a single, simple change. Instead of jointly encoding all windows in a sentence, we process each independently. As a result, the neural network achieves better performance by learning underlying word-specific information from brain recordings. This makes two existing strategies become much more effective than before. Both aggregating predictions from distinct neural responses to the same word and using a pretrained LLM as a linguistic prior now substantially improve results. On our perceived speech benchmark, this simple recipe (SimpleB2T) achieves a word error rate of 36.6% with five observations per word, approaching past invasive speech decoding performance, albeit under different conditions. The results in this work expose an important shortcut in brain-to-text decoding and show that removing it leads to a simple and considerably more effective strategy.
ER -