Unit: Axons and Dendrites

Learning goals

• Distinguish axonal and dendritic structures in EM imagery.
• Use organelle and morphology cues for robust process classification.
• Handle ambiguous cases in proofreading with context-aware decisions.

Scope boundaries

• In scope:
  • morphology-first discrimination of axons vs dendrites
  • proofreading cues and common misclassification traps
• Out of scope:
  • full synapse-typing deep dive (handled in ultrastructure/synapse units)

Website draft blocks

Hero framing

Accurate axon/dendrite identification is one of the most important practical skills in connectomics. This unit focuses on visual cues that support reliable classification during proofreading.

Core cues

• Process caliber and branching patterns.
• Spine-associated vs bouton-associated context.
• Organelle signatures and continuity behavior across slices.

Practical workflow

1. Start with local morphology.
2. Check organelle/context clues.
3. Track continuity in neighboring sections.
4. Assign provisional class and confidence.
5. Resolve edge cases with broader neighborhood context.

Common pitfalls

• Overreliance on single-slice appearance.
• Misclassification in dense neuropil crossings.
• Ignoring nearby synaptic context and process continuity.

Slide draft sequence (v1)

1. Why axon/dendrite discrimination matters
2. Morphology cues overview
3. Organelle cues overview
4. Dendrite-focused examples
5. Axon-focused examples
6. Ambiguous examples and error taxonomy
7. Practical proofreading checklist
8. Summary + quick-reference cue table

Figure candidates

See: course/units/figures/06-axons-and-dendrites-selected-v1.md

Cross-links

• Related modules: module04, module09
• Related unit: 05-neuronal-ultrastructure
• Related tool: /tools/connectome-quality/

Open issues

• Add a learner-facing decision tree graphic for classification.
• Validate terminology consistency with Unit 05 glossary.

Scope check (expert pass)

• Emphasize probabilistic classification in ambiguous neuropil regions.
• Add explicit edge cases: en passant boutons, thin dendrites, branch points, truncated processes.
• Require multi-slice continuity checks before final class assignment.

Technical anchors to preserve

• Morphology + organelle + local connectivity cues should be combined.
• Confidence scoring improves auditability and inter-rater consistency.
• Misclassification at this stage propagates to network-level interpretation errors.