Hanchuan Peng

This paper introduces MMB, a framework that encodes neuronal morphology into symbolic representations, revealing region-specific and scale-dependent principles of neuronal organization across the brain.

This study identifies spatial synaptic organization as a mechanism for stabilizing learning and introduces Spatial Synaptic Regularization (SSR) to improve AI learning stability.

The paper analyzes the mouse brain connectome to show that specific neuron connection patterns and modules contribute to global dynamical stability rather than optimal processing capacity.

This paper introduces a convolution framework for neuronal morphology reconstruction, showing improved performance on mesoscaled neuronal datasets.

The paper introduces LetsACT, a multimodal integration platform that combines morphological, spatial, and transcriptomic data to analyze morpho-genetic variations in human cortical neurons associated with tumor infiltration.

The paper identifies three organizational principles of post-synaptic sites in the mouse brain using a statistical approach and validates them across species and with electrophysiological data.

SpatialZ generates dense 3D cell atlases from planar spatial transcriptomics data by reconstructing virtual slices, validated with real 3D data and enabling new spatial analyses.

This paper constructs a high-resolution mouse brain connectome by integrating multi-scale, multi-modal data and validates it against existing connectomes and gene expression data.

The paper generates whole-brain connectivity by predicting detailed axonal morphologies from sparse data, enabling in silico experimentation of large brain regions.

The paper uses a whole-brain EM dataset of Drosophila to analyze synapses and microenvironmental features, revealing functional modules in the adult brain.

NeuroXiv provides an AI-powered database of neuron morphologies from mouse brains, enabling dynamic analysis and comparison across a standardized atlas.

This paper identifies intrinsic features that distinguish human and mouse cortical neurons, revealing significant differences in morphology, function, and separability across brain regions.

This paper constructs a mouse brain atlas based on dendritic microenvironments, revealing new subregions and improving projection mapping by integrating spatial and transcriptomic data.

The paper examines axonal bouton distribution in mouse neurons, showing non-homogeneous patterns and their impact on whole-brain network topology.

The paper introduces a method to map individual sensory nerve axons from digits to spinal cord at sub-micron resolution.

The paper maps single cortical neurons across the human brain to profile morphological diversity and stereotypy using a new tomography approach.

The paper introduces a collaborative platform augmented by AI for large-scale 3D neuron morphology reconstruction in mouse and human brains, showing its capability to generate tens of thousands of reconstructions.

The paper defines neuron connectivity types (c-types) using morphology and connectivity data, showing they better capture subtype distinctions than other features.

BigNeuron benchmarks automatic neuron reconstruction algorithms using gold standard annotations and image quality metrics, showing that consensus methods outperform individual algorithms.

The paper provides a guide to the BRAIN Initiative Cell Census Network (BICCN) data ecosystem, its tools, and resources for analyzing brain cell types across species.

The paper provides a detailed cellular-resolution map of the mouse primary motor cortex, defining cell types and their connectivity.

The paper introduces MorphoHub, a platform for generating petabyte-scale multi-morphometry data, enabling high-quality analysis of neuronal structures in whole brain datasets.

The paper characterizes transcriptomic and physiological properties of GABAergic cortical neurons in mice, identifying 20 interneuron met-types with consistent features across modalities.

Complete morphologies of 1,708 neurons, sparsely labelled and imaged whole-brain with fMOST, expose region-specific long-range projection rules and finer diversity than current transcriptomic subtypes capture.

The paper identifies transcriptomic diversity in human supragranular neurons and links it to functional specialization and disease vulnerability.

The paper introduces TeraVR, a virtual reality system for precise 3-D neuronal morphology reconstruction in whole mouse brains, enabling collaborative and efficient annotation of complex neuronal structures.

The paper classifies neurons in mouse visual cortex into 17 electrophysiological and 35 morphological types using unsupervised methods and links them to transcriptomic types.

The paper maps excitatory and inhibitory synapses across dendritic trees of cortical neurons and finds local balance in their ratio despite domain-specific density variations.

DeepNeuron is an open-source deep learning toolbox that improves neuron tracing in light microscopy images by learning features and rules from data to address challenges in detecting, connecting, and classifying neuronal structures.