NeurIPT: Foundation Model for Neural Interfaces

2025-10-18

AI Summary

To address the poor cross-subject, cross-task, and cross-device generalization of EEG signals, this paper introduces the first foundational model specifically designed for electroencephalography (EEG)-based neural interfaces. Methodologically, we propose an amplitude-aware masked pretraining strategy and develop a spatiotemporal unified Transformer architecture that integrates 3D electrode coordinate positional encoding, intra- and inter-regional brain pooling, and a progressive Mixture-of-Experts (MoE) design—enabling robust representation learning across heterogeneous EEG data. After fine-tuning on eight mainstream BCI downstream tasks, our model consistently outperforms state-of-the-art approaches, demonstrating superior generalization capability and plug-and-play usability. This work establishes a scalable framework and delineates key technical pathways for foundational model development in EEG analysis.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Humans and AI: Brain-Sensing and Analysis

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Electroencephalography (EEG) has wide-ranging applications, from clinical diagnosis to brain-computer interfaces (BCIs). With the increasing volume and variety of EEG data, there has been growing interest in establishing foundation models (FMs) to scale up and generalize neural decoding. Despite showing early potential, applying FMs to EEG remains challenging due to substantial inter-subject, inter-task, and inter-condition variability, as well as diverse electrode configurations across recording setups. To tackle these open challenges, we propose NeurIPT, a foundation model developed for diverse EEG-based Neural Interfaces with a Pre-trained Transformer by capturing both homogeneous and heterogeneous spatio-temporal characteristics inherent in EEG signals. Temporally, we introduce Amplitude-Aware Masked Pretraining (AAMP), masking based on signal amplitude rather than random intervals, to learn robust representations across varying signal intensities beyond local interpolation. Moreover, this temporal representation is enhanced by a Progressive Mixture-of-Experts (PMoE) architecture, where specialized expert subnetworks are progressively introduced at deeper layers, adapting effectively to the diverse temporal characteristics of EEG signals. Spatially, NeurIPT leverages the 3D physical coordinates of electrodes, enabling effective transfer of embedding across varying EEG settings, and develops Intra-Inter Lobe Pooling (IILP) during fine-tuning to efficiently exploit regional brain features. Empirical evaluations across eight downstream BCI datasets, via fine-tuning, demonstrated NeurIPT consistently achieved state-of-the-art performance, highlighting its broad applicability and robust generalization. Our work pushes forward the state of FMs in EEG and offers insights into scalable and generalizable neural information processing systems.

Problem

• Addresses EEG variability across subjects, tasks, and recording conditions
• Develops foundation model for scalable neural decoding in brain-computer interfaces
• Handles diverse electrode configurations and spatio-temporal EEG characteristics

Innovation

• Transformer model captures EEG spatio-temporal characteristics
• Amplitude-aware masking learns robust temporal representations
• Progressive MoE architecture adapts to diverse EEG signals