In recent years, foundation models—general-purpose AI models centered around Transformer architectures and trained on massive-scale data—have profoundly reshaped the paradigm of AI research. These models demonstrate exceptional capabilities in generalization, task transfer, and knowledge extraction, and have led to revolutionary breakthroughs in fields such as natural language processing, computer vision, and speech recognition.

At the same time, neuroscience is entering a new data-driven era. The rapid accumulation of large-scale multimodal neural data (e.g., MRI, fMRI, MEG, EEG, SEEG, ECoG, fNIRS) and behavioral data (e.g., eye movements, motor actions, decision-making) demands new, efficient modeling tools to uncover the structure, function, and cognitive mechanisms of the brain. Foundation models—especially those with open architectures and multimodal modeling capabilities—are becoming key technological engines for advancing brain science.

Moreover, some open-source, white-box foundation models offer observable, controllable, and analyzable computational structures, providing neuroscience with an unprecedented paradigm of a "dissectible" complex system. Studying the internal representational structures and learning mechanisms of these models may offer new theoretical tools and experimental support for addressing fundamental scientific questions such as neural information encoding, generalization mechanisms, and cognitive modeling.

To promote the deep integration of neuroscience and artificial intelligence, this special issue focuses on the interdisciplinary research at the intersection of foundation models for brain science. We cordially invite researchers from related fields to submit their work and jointly explore the frontiers of AI-neuroscience convergence.

Scope of Topics:

1. Theory of Foundation Models and Neuroscience

1. Theoretical investigations of foundation models in neuroscience
2. Methodology for modeling brain structure and function using neuroimaging modalities
3. Representation learning and pretraining models for neural signals

1. Foundation Models for Cognitive Science

1. Simulation and modeling of memory, attention and decision-making using foundation models
2. Multimodal integration of neural and behavioral data based on foundation models
3. Representation and analysis of cognitive processes (e.g., perception, reasoning, learning)

1. Foundation Models for Brain-Computer Interface (BCI)

1. Foundation models for neural decoding
2. Modeling and optimization of human-machine interaction strategies
3. Interpretation and feedback modeling of neural signals

1. Brain-Inspired Intelligence and Neuromorphic Foundation Models

1. Design and optimization of brain-inspired foundation model architectures
2. Analysis of transfer learning and generalization in brain-like neural networks
3. Modeling intelligent systems with cognitive capabilities

1. Foundation Models for Brain Disease and Clinical Applications

1. Foundation models for early detection of brain disorders
2. Risk prediction models for neurological and psychiatric diseases
3. Intelligent modeling for rehabilitation support and intervention planning

1. Alignment and Mechanism Analysis between Foundation Models and Brain

1. Alignment mechanisms and biological interpretability between neural representations and foundation models
2. Modeling the geometric structure and evolution of representational manifolds
3. Theoretical links and empirical studies on representational structures and generalization ability

Submission Guidelines

Authors should prepare papers in accordance with the format requirements of Tsinghua Science and Technology, with reference to the Instruction given at https://www.sciopen.com/journal/1007-0214, and submit the complete manuscript through the online manuscript submission system at https://mc03.manuscriptcentral.com/tst with manuscript type as “Special Issue on Foundation Models for Brain Science”.

Important Dates

Deadline for submissions: December 31, 2025

Guest Editors

• Quanying Liu, Southern University of Science and Technology, China
• Helen Juan Zhou, National University of Singapore, Singapore
• Jibin Wu, Hong Kong Polytechnic University, China
• Guoqi Li, Institute of Automation, Chinese Academy of Sciences, China
• Xiaomu Wang, Nanjing University, China
• Lina Yao, The University of New South Wales, Australia
• Chen Wei, University of Birmingham, UK