Huanqiu Zhang

The paper defines neural code symbols as variable-duration packets in auditory spike trains, showing this approach enables real-time readout and maximizes information capacity.

The paper demonstrates that embedding connectome data into a geometric space improves prediction of neural activity compared to high-dimensional models.

The paper develops a predictive autoencoder model of the hippocampus to simulate sequence learning, showing CA3's predictive role and CA1's error signaling in mice.