Meta-learning In-Context Enables Training-Free Cross Subject Brain Decoding

TL;DR

This paper introduces a meta-learning approach for cross-subject brain decoding from fMRI data that generalizes without fine-tuning by conditioning on a few examples, enabling robust visual decoding across individuals.

Contribution

A novel meta-optimized, training-free method for cross-subject brain decoding that leverages in-context learning and hierarchical inference without requiring retraining.

Findings

Achieves strong cross-subject and cross-scanner generalization.

Does not require anatomical alignment or stimulus overlap.

Operates without fine-tuning or retraining.

Abstract

Visual decoding from brain signals is a key challenge at the intersection of computer vision and neuroscience, requiring methods that bridge neural representations and computational models of vision. A field-wide goal is to achieve generalizable, cross-subject models. A major obstacle towards this goal is the substantial variability in neural representations across individuals, which has so far required training bespoke models or fine-tuning separately for each subject. To address this challenge, we introduce a meta-optimized approach for semantic visual decoding from fMRI that generalizes to novel subjects without any fine-tuning. By simply conditioning on a small set of image-brain activation examples from the new individual, our model rapidly infers their unique neural encoding patterns to facilitate robust and efficient visual decoding. Our approach is explicitly optimized for in-context learning of the new subject's encoding model and performs decoding by hierarchical inference, inverting the encoder. First, for multiple brain regions, we estimate the per-voxel visual response encoder parameters by constructing a context over multiple stimuli and responses. Second, we construct a context consisting of encoder parameters and response values over multiple voxels to perform aggregated functional inversion. We demonstrate strong cross-subject and cross-scanner generalization across diverse visual backbones without retraining or fine-tuning. Moreover, our approach requires neither anatomical alignment nor stimulus overlap. This work is a critical step towards a generalizable foundation model for non-invasive brain decoding.