Bei Hong

The paper reconstructs multiscale and multimodal communication networks in the suprachiasmatic nucleus using electron microscopy, revealing unique synaptic and electrotonic structures.

CellCUT improves spatial transcriptomics by accurately segmenting cells and assigning transcripts using nuclear images, outperforming existing methods in diverse tissues and challenging data conditions.

The paper uses deep learning to analyze electron microscopy data from mouse auditory cortex, revealing structural changes in synapses and mitochondria associated with fear memory.