Brain-Wide Shape Reconstruction of a Traced Neuron Using the Convex Image Segmentation Method

Area of Science:

• Neuroscience
• Computational Biology
• Image Analysis

Background:

• Neuronal shape reconstruction is vital for understanding neural circuits and neuron identity.
• Advances in optical imaging provide whole-brain neuron shape data, but challenges remain.
• Fuzzy neurites and closely attached neurites (crossover regions) complicate accurate reconstruction.

Purpose of the Study:

• To develop a robust convex image segmentation model for accurate neuronal shape reconstruction.
• To address challenges posed by fuzzy neurites and neurite crossover regions in large-scale datasets.

Main Methods:

• Proposed a convex image segmentation model for neuronal shape reconstruction.
• Segmented neurites into cross-sections along their traced skeletons.
• Utilized sparse gradient image properties and small radius rule for fuzzy neurite regions.
• Leveraged skeleton point relationships to identify neurites in crossover regions.

Main Results:

• Demonstrated model performance on datasets with fuzzy neurites and crossover regions.
• Verified robust neuron shape reconstruction on a brain-wide scale.
• The model successfully identified and reconstructed neurites in challenging areas.

Conclusions:

• The proposed convex image segmentation model offers a robust solution for neuronal shape reconstruction.
• The model effectively overcomes limitations of previous methods in handling complex neuronal structures.
• This technique advances large-scale neuronal anatomy and circuit mapping capabilities.