brainlit

brainlit analyzes neuronal morphology by fitting branching B-splines to SWC-formatted neuron traces and computing curvature and torsion to quantify the internal geometry of axonal arbors.

Key Features:

• Spline fitting with B-splines: Fits neuron trace points with branching B-splines, treating branches as differentiable curves.
• Frenet-Serret formulas: Computes curvature and torsion in closed form using the Frenet-Serret formulas.
• Continuous parameterization: Defines curvature and torsion continuously along the curve as a pointwise alternative to start-end metrics such as tortuosity.
• Application to cortical projection neurons: Applied to cortical projection neurons traced in two mouse brains to quantify geometric differences between primary, collateral, and terminal axon branches.
• Validation across samples: Produces consistent geometric characterizations across different brain samples.

Scientific Applications:

• Neuronal subtype identification: Quantifies geometric properties that can help distinguish neuronal subtypes by morphology.
• Understanding learning processes: Provides geometric metrics to investigate how structural variations in neurons relate to learning-related phenomena.
• Neurological disease research: Enables detailed neuromorphological analyses relevant to studying disease mechanisms and exploring morphological biomarkers.

Methodology:

Processes SWC-formatted neuron traces in a Python package by fitting branching B-splines and computing curvature and torsion via the Frenet-Serret formulas, then analyzes these parameters along axonal arbors.