IBrain reads brain signals jointly from surface to spikes
iBrain: A Unified Foundation Model Reading the Brain from Surface to Spikes
Artificial Intelligence
Summary
Measuring brain activity is done using different tools that capture signals at different scales, like brain waves and neuron spikes. The authors present iBrain, a new model that learns from both types of signals together instead of separately. They trained iBrain on thousands of hours of brain recordings to help it understand the complex patterns in these signals. This model performs better than earlier methods and can work well in different brain recording setups.
What this means in practice
- •For neural engineers: Use joint-trained iBrain models to interpret brain signals more accurately across different recording methods during brain-machine interface development.
- •For medical device developers: Incorporate iBrain’s unified model to improve neural signal decoding for invasive diagnostic and therapeutic brain devices.$Commercial implications: Enables advanced brain signal analysis software for medical devices that improve diagnostics and treatments.
Authors
Ying Chen, Tiou Wang, Zhifeng Yue
Abstract
Invasive neural recordings provide high-fidelity measurements of brain activity, with signals such as intracranial EEG (iEEG) and intracortical spiking activity capturing neural dynamics at different spatial and temporal scales. Yet existing neural foundation models have largely been developed independently for different invasive recording paradigms, leaving joint pretraining across heterogeneous invasive signals underexplored. In this work, we introduce iBrain, a unified foundation model that jointly learns from iEEG and spiking activity. iBrain employs signal-specific encoders to accommodate their distinct signal characteristics and a shared spatiotemporal Transformer backbone to model dependencies across recording channels and time. We pretrain iBrain on over 7,000 hours of heterogeneous neural recordings using masked signal reconstruction and channel-view alignment, promoting contextual modeling of neural dynamics and robustness across different channels. iBrain consistently outperforms single-signal pretraining baselines and achieves state-of-the-art performance on multiple benchmarks. Further experiments demonstrate that iBrain exhibits transferability and data efficiency across diverse recording settings. These results highlight the potential of joint pretraining on heterogeneous invasive neural recordings to support scalable neural modeling and transferable representations across recording settings and downstream tasks.