GTree

GTree reconstructs neuronal morphologies from large-scale brain imaging datasets for population-level neuron tracing and single-axon resolution analysis.

Key Features:

• Automated Neuron Tracing: Integrates multiple automated tracing algorithms to trace neurons in single images and reconstruct neuronal populations from brain-wide image stacks.
• Error-Screening System: Implements an error-screening system to detect and correct submicron errors in densely packed neurites and align reconstructions with ground truth across long projections.
• Efficiency and Speed: Processes extensive datasets such as whole mouse-brain image stacks and has been reported to operate approximately five times faster than widely used commercial software on such datasets.
• Scalability: Handles very large datasets exceeding 10 terabytes and supports data from various light microscopy platforms.

Scientific Applications:

• Population-level neuronal reconstruction: Reconstructs neuronal populations at single-axon level across the mouse brain to support mapping of neural circuits.
• Anatomical mapping of long-range projections: Applied to reconstruct 35 long-projection neurons around an injection site in a mouse brain for detailed connectivity analysis.

Methodology:

Performs automated neuron tracing via integrated algorithmic approaches and applies an error-screening system to detect and correct submicron neurite errors on brain-wide image stacks, including datasets >10 TB.