MapGuide: A Simple yet Effective Method to Reconstruct Continuous Language from Brain Activities

TL;DR

This paper introduces a straightforward method for reconstructing continuous language from brain signals by directly comparing predicted text embeddings, significantly outperforming previous indirect approaches and emphasizing the importance of accurate brain-to-embedding mapping.

Contribution

The paper presents a novel direct comparison approach for brain-to-text reconstruction, achieving substantial performance improvements over existing methods.

Findings

Our method improves BLEU scores by 77% and METEOR scores by 54%.

Better brain-to-embedding mapping leads to more accurate language reconstruction.

The approach simplifies the process of decoding language from brain activity.

Abstract

Decoding continuous language from brain activity is a formidable yet promising field of research. It is particularly significant for aiding people with speech disabilities to communicate through brain signals. This field addresses the complex task of mapping brain signals to text. The previous best attempt reverse-engineered this process in an indirect way: it began by learning to encode brain activity from text and then guided text generation by aligning with predicted brain responses. In contrast, we propose a simple yet effective method that guides text reconstruction by directly comparing them with the predicted text embeddings mapped from brain activities. Comprehensive experiments reveal that our method significantly outperforms the current state-of-the-art model, showing average improvements of 77% and 54% on BLEU and METEOR scores. We further validate the proposed modules through detailed ablation studies and case analyses and highlight a critical correlation: the more precisely we map brain activities to text embeddings, the better the text reconstruction results. Such insight can simplify the task of reconstructing language from brain activities for future work, emphasizing the importance of improving brain-to-text-embedding mapping techniques.