A Systematic Evaluation of Sample-Level Tokenization Strategies for MEG Foundation Models

SungJun Cho, Chetan Gohil, Rukuang Huang, Oiwi Parker Jones, Mark W. Woolrich · 2026

Read A Systematic Evaluation of Sample-Level Tokenization Strategies for MEG Foundation Models by SungJun Cho, Chetan Gohil, Rukuang Huang, Oiwi Parker Jones, Mark W. Woolrich (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-16626.txt.

Description

SungJun Cho and co-authors test sample-level token schemes for MEG brain models in a head-to-head way. Brain models must chop smooth neural time series into tokens, but the effect of that chop is poorly known. The study pits learned choppers against fixed ones on magnetoencephalography data, checking rebuild faith, token forecasts, lifelike generated data, subject info kept, and later tasks.

A new autoencoder-based learned chopper is brought in. Tests span three public MEG sets from varied sites, scanners, and tasks. Both learned and fixed schemes rebuild well and score much the same on most checks. The upshot is that plain fixed sample-level token plans can serve brain models, sparing builders needless fuss.

Catalog record

CREATORSungJun Cho, Chetan Gohil, Rukuang Huang, Oiwi Parker Jones, Mark W. Woolrich
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE64 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-16626

Browse the full library · About the Tower of Babel