Meta Learning In Context Enables Training Free Cross Subject

---
name: meta-learning-in-context-enables-training-free-cross-subject
description: "Visual decoding from brain signals is a key challenge at the intersection of computer vision and neuroscience, requiring methods that bridge neural re... Activation: cs.LG, q-bio.NC, neuroscience, research"
---

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

## Overview

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.

## Source Paper

- **Title:** Meta-learning In-Context Enables Training-Free Cross Subject Brain Decoding
- **Authors:** Mu Nan, Muquan Yu, Weijian Mai, Jacob S. Prince, Hossein Adeli
- **arXiv:** 2604.08537v1
- **Published:** 2026-04-09
- **Categories:** cs.LG, q-bio.NC
- **PDF:** https://arxiv.org/abs/2604.08537
- **Value Score:** 29

## Core Concepts

### Key Contributions

Based on the abstract analysis, this paper contributes to understanding cs.LG, q-bio.NC through:

1. **Novel Methodology**: 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

2. **Technical Approach**: The paper employs cs.LG and q-bio.NC methodologies to address the research problem.

3. **Applications**: Potential applications include decoding.

## Research Context

This work sits at the intersection of cs.LG and q-bio.NC, building on prior work in:
- Computational neuroscience methods
- Neural dynamics analysis
- Brain network modeling

## Implementation Notes

For implementation of the methods described in this paper:

```python
# Reference implementation structure
# Based on: Meta-learning In-Context Enables Training-Free Cross Subject Brain Decoding

import numpy as np

# Key methodology components:
# 1. Data preprocessing
# 2. Model training / analysis
# 3. Evaluation and validation

# TODO: Implement specific methodology from paper
# See full paper at https://arxiv.org/abs/2604.08537
```

## Related Work

- Papers in similar categories: cs.LG, q-bio.NC
- See also skills in ai_collection for related neuroscience methods

## Activation Keywords

- cs.LG, q-bio.NC
- neuroscience
- brain network
- neural dynamics
- computational neuroscience

## Latest Research Updates

### arXiv:2604.08537v1 (2026-04-09)
**Title:** Meta-learning In-Context Enables Training-Free Cross Subject Brain Decoding
**Authors:** Mu Nan, Muquan Yu, Weijian Mai et al.
**Link:** https://arxiv.org/abs/2604.08537v1


## References

- Mu Nan et al. (2026). "Meta-learning In-Context Enables Training-Free Cross Subject Brain Decoding." arXiv:2604.08537v1.