Neuro2Semantic: A Transfer Learning Framework for Semantic Reconstruction of Continuous Language from Human Intracranial EEG

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shams, Siavash, Antonello, Richard, Mischler, Gavin, Bickel, Stephan, Mehta, Ashesh, Mesgarani, Nima
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909630385356800
author Shams, Siavash
Antonello, Richard
Mischler, Gavin
Bickel, Stephan
Mehta, Ashesh
Mesgarani, Nima
author_facet Shams, Siavash
Antonello, Richard
Mischler, Gavin
Bickel, Stephan
Mehta, Ashesh
Mesgarani, Nima
contents Decoding continuous language from neural signals remains a significant challenge in the intersection of neuroscience and artificial intelligence. We introduce Neuro2Semantic, a novel framework that reconstructs the semantic content of perceived speech from intracranial EEG (iEEG) recordings. Our approach consists of two phases: first, an LSTM-based adapter aligns neural signals with pre-trained text embeddings; second, a corrector module generates continuous, natural text directly from these aligned embeddings. This flexible method overcomes the limitations of previous decoding approaches and enables unconstrained text generation. Neuro2Semantic achieves strong performance with as little as 30 minutes of neural data, outperforming a recent state-of-the-art method in low-data settings. These results highlight the potential for practical applications in brain-computer interfaces and neural decoding technologies.
format Preprint
id arxiv_https___arxiv_org_abs_2506_00381
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Neuro2Semantic: A Transfer Learning Framework for Semantic Reconstruction of Continuous Language from Human Intracranial EEG
Shams, Siavash
Antonello, Richard
Mischler, Gavin
Bickel, Stephan
Mehta, Ashesh
Mesgarani, Nima
Computation and Language
Audio and Speech Processing
Signal Processing
Decoding continuous language from neural signals remains a significant challenge in the intersection of neuroscience and artificial intelligence. We introduce Neuro2Semantic, a novel framework that reconstructs the semantic content of perceived speech from intracranial EEG (iEEG) recordings. Our approach consists of two phases: first, an LSTM-based adapter aligns neural signals with pre-trained text embeddings; second, a corrector module generates continuous, natural text directly from these aligned embeddings. This flexible method overcomes the limitations of previous decoding approaches and enables unconstrained text generation. Neuro2Semantic achieves strong performance with as little as 30 minutes of neural data, outperforming a recent state-of-the-art method in low-data settings. These results highlight the potential for practical applications in brain-computer interfaces and neural decoding technologies.
title Neuro2Semantic: A Transfer Learning Framework for Semantic Reconstruction of Continuous Language from Human Intracranial EEG
topic Computation and Language
Audio and Speech Processing
Signal Processing
url https://arxiv.org/abs/2506.00381