Saved in:
Bibliographic Details
Main Authors: Zheng, Ruizhe, Sun, Lichao
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2409.17987
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910621379854336
author Zheng, Ruizhe
Sun, Lichao
author_facet Zheng, Ruizhe
Sun, Lichao
contents Decoding visual-semantic information from brain signals, such as functional MRI (fMRI), across different subjects poses significant challenges, including low signal-to-noise ratio, limited data availability, and cross-subject variability. Recent advancements in large language models (LLMs) show remarkable effectiveness in processing multimodal information. In this study, we introduce an LLM-based approach for reconstructing visual-semantic information from fMRI signals elicited by video stimuli. Specifically, we employ fine-tuning techniques on an fMRI encoder equipped with adaptors to transform brain responses into latent representations aligned with the video stimuli. Subsequently, these representations are mapped to textual modality by LLM. In particular, we integrate self-supervised domain adaptation methods to enhance the alignment between visual-semantic information and brain responses. Our proposed method achieves good results using various quantitative semantic metrics, while yielding similarity with ground-truth information.
format Preprint
id arxiv_https___arxiv_org_abs_2409_17987
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle LLM4Brain: Training a Large Language Model for Brain Video Understanding
Zheng, Ruizhe
Sun, Lichao
Computer Vision and Pattern Recognition
Human-Computer Interaction
Decoding visual-semantic information from brain signals, such as functional MRI (fMRI), across different subjects poses significant challenges, including low signal-to-noise ratio, limited data availability, and cross-subject variability. Recent advancements in large language models (LLMs) show remarkable effectiveness in processing multimodal information. In this study, we introduce an LLM-based approach for reconstructing visual-semantic information from fMRI signals elicited by video stimuli. Specifically, we employ fine-tuning techniques on an fMRI encoder equipped with adaptors to transform brain responses into latent representations aligned with the video stimuli. Subsequently, these representations are mapped to textual modality by LLM. In particular, we integrate self-supervised domain adaptation methods to enhance the alignment between visual-semantic information and brain responses. Our proposed method achieves good results using various quantitative semantic metrics, while yielding similarity with ground-truth information.
title LLM4Brain: Training a Large Language Model for Brain Video Understanding
topic Computer Vision and Pattern Recognition
Human-Computer Interaction
url https://arxiv.org/abs/2409.17987