Learning Time-Scale Invariant Population-Level Neural Representations

TL;DR

This paper introduces a novel pretraining method called Time-scale Augmented Pretraining (TSAP) that enhances the invariance of neural population representations to different time-scales, improving robustness and generalization for neural decoding tasks.

Contribution

The paper proposes TSAP, a new pretraining approach that increases time-scale invariance in population-level neural representations, addressing a key challenge in neural model generalization.

Findings

TSAP improves robustness to time-scale mismatches across tasks.

Existing models lack invariance to time-scale variations.

Handling preprocessing diversity is crucial for neural foundation models.

Abstract

General-purpose foundation models for neural time series can help accelerate neuroscientific discoveries and enable applications such as brain computer interfaces (BCIs). A key component in scaling these models is population-level representation learning, which leverages information across channels to capture spatial as well as temporal structure. Population-level approaches have recently shown that such representations can be both efficient to learn on top of pretrained temporal encoders and produce useful representations for decoding a variety of downstream tasks. However, these models remain sensitive to mismatches in preprocessing, particularly on time-scales, between pretraining and downstream settings. We systematically examine how time-scale mismatches affects generalization and find that existing representations lack invariance. To address this, we introduce Time-scale Augmented Pretraining (TSAP), which consistently improves robustness to different time-scales across decoding tasks and builds invariance in the representation space. These results highlight handling preprocessing diversity as a key step toward building generalizable neural foundation models.