---
title: "Beyond Trial Averaging: Anchoring Neural and Visual Representations for Few-Repetition Brain-to-Image Retrieval"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.19128v1"
permalink: "https://feed.astralor.com/items/itm_4095fe063a6f95aeddd2621c"
published: "2026-08-19T17:23:30.000Z"
score: 78
channel: "ai"
tags: ["NEAR","THINGS-EEG2","Few-Shot Learning","Neural Decoding","Brain-to-Image Retrieval","Brain-Computer Interface"]
---

# Beyond Trial Averaging: Anchoring Neural and Visual Representations for Few-Repetition Brain-to-Image Retrieval

## AI commentary

> What makes this paper truly valuable is that it re-examines the 'few-repetition retrieval collapse' phenomenon, usually attributed to noise, and uncovers a subtler representation alignment structure. For researchers in brain-computer interfaces and neural decoding, NEAR could mean future experiments no longer need subjects to view images dozens of times—images could be retrieved from a single trial. Though it's a preprint without formal peer review, the method's validation across three neuroimaging modalities is convincing.

## At a glance

Brain-to-image retrieval typically requires averaging up to 80 neural trials per image to suppress noise; when only one or a few repetitions are available, retrieval accuracy drops sharply. This paper discovers a non-transitive alignment pattern: low-repetition query signals and image representations each align with the high-repetition center, but not directly with each other, indicating that query noise is only part of the problem and gallery placement also matters. The authors propose NEAR, a neural-anchor-based retrieval framework that treats the high-repetition center as an anchor and approaches it from both sides: a denoiser pulls the noisy query toward the true anchor, and a small network predicts each candidate's pseudo-anchor from its image and pulls the image toward it. Across four datasets spanning EEG, MEG, and fMRI, NEAR consistently improves retrieval in the few-repetition regime. On THINGS-EEG2, it improves 200-way Top-1 accuracy by 5.7 and 9.3 percentage points when averaging one and four repetitions, respectively. By anchoring neural and visual representations, NEAR reduces reliance on repeated acquisition and brings neural retrieval closer to real-world deployment.

## Key points

- Identifies a non-transitive alignment pattern: low-repetition queries and image representations each align with the high-repetition center but not directly with each other, suggesting gallery placement also affects retrieval.
- Proposes NEAR, which anchors on the high-repetition center and approaches it from both sides: denoising the query and predicting pseudo-anchors from images.
- Consistently improves retrieval in the few-repetition regime across four datasets spanning EEG, MEG, and fMRI.
- On THINGS-EEG2, improves 200-way Top-1 accuracy by 5.7 and 9.3 percentage points with one and four averaged repetitions, respectively.

## Full text

Decoding visual information from brain signals probes neural representations and enables neuro-rehabilitation and dream decoding. Recent brain-to-image retrieval approaches have achieved promising performance, typically by averaging many (up to 80) neural trials per image, requiring repeated stimulus presentation that increases latency, cost, and user burden. When only one or a few repetitions are available, the retrieval accuracy drops sharply. This drop is commonly attributed to query noise because averaging suppresses noise and increases signal stability. However, we find a non-transitive alignment pattern: the low-repetition query signal and the image representation each align with the high-repetition center, but not directly with each other. This pattern shows that query noise is only part of the problem and that gallery placement also affects retrieval. We therefore propose a neural-anchor-based retrieval (NEAR) framework that treats the high-repetition center as an anchor and approaches it from both sides: a denoiser pulls the noisy query toward the true anchor, and a small network predicts each candidate's pseudo anchor from its image and pulls the image toward it. Across four datasets spanning EEG, MEG and fMRI, NEAR consistently improved retrieval in the few-repetition regime. On THINGS-EEG2, it improved 200-way Top-1 accuracy by 5.7 and 9.3 percentage points respectively, when averaging one and four repetitions. By anchoring neural and visual representations, NEAR reduces reliance on repeated acquisition and brings neural retrieval closer to real-world deployment.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_4095fe063a6f95aeddd2621c