Joergen M. R. Kornfeld

Constraints for spatially and temporally precise learning in a neural circuit model of reinforcement learning

The paper presents a neural circuit model of reinforcement learning in songbirds that demonstrates how temporal and spatial imprecision in dopaminergic signals can be overcome through fast reward prediction updates.

Songbird connectome reveals tunneling of migratory neurons in the adult striatum

The paper uses connectomics to show that migrating neurons in the adult zebra finch striatum physically reshape mature circuits by creating tunnels through synapse-rich regions.

The songbird basal ganglia connectome

This paper presents the first vertebrate basal ganglia connectome, revealing complex connectivity patterns that challenge the traditional view of discrete pathways.

Combinatorial protein barcodes enable self-correcting neuron tracing with nanoscale molecular context

The paper introduces PRISM, a platform that uses combinatorial protein barcodes to enable self-correcting neuron tracing with molecular context, improving tracing accuracy and handling spatial gaps.

Public volume electron microscopy data: An essential resource to study the brain microvasculature

The paper presents a meta-analysis of electron microscopy datasets to explore cerebrovascular structures and identifies features important for vascular function and segmentation improvements.

An anatomical substrate of credit assignment in reinforcement learning

The paper shows how the basal ganglia solve the credit assignment problem through synaptic architecture and biophysical modeling in songbirds.

Local axonal conduction delays underlie precise timing of a neural sequence

The paper shows that long axonal conduction delays in HVC neurons are necessary for precise millisecond timing of neural sequences during song production.

Learning cellular morphology with neural networks

The paper introduces CMNs to automate morphological analysis of neuronal reconstructions, identifying glia cells and resolving reconstruction errors.

High-Precision Automated Reconstruction of Neurons with Flood-filling Networks

This paper introduces flood-filling networks for automated neuron reconstruction, achieving high precision in tracing neurites from electron microscopy data.