Landmarking the brain for geometric morphometric analysis: an error study.

Area of Science:

• Neuroimaging and Neuroscience
• Biomedical Engineering
• Human Anatomy

Background:

• Volumetric analysis is common for neuroanatomic phenotypes but cannot quantify shape changes or regional interrelationships.
• Statistical shape analysis using landmarks is superior for assessing brain shape but lacks standardized protocols and reliability data.
• Previous studies have used numerous landmarks but often fail to publish identification protocols or validation results.

Purpose of the Study:

• To collaboratively develop detailed data collection protocols for a set of 29 brain landmarks.
• To conduct an intra- and inter-observer validation study for the established set of landmarks.

Main Methods:

• Development of detailed protocols for 29 cortical and subcortical landmarks.
• Utilized a sample of 10 twelve-year-old boys for protocol development.
• Performed intra- and inter-observer reliability testing on the defined landmarks.

Main Results:

• Established detailed protocols for 29 specific brain landmarks.
• Achieved an average intra-observer error of 1.9 mm (range: 0.72 mm–5.6 mm).
• Achieved an average inter-observer error of 1.1 mm (range: 0.40 mm–3.4 mm).

Conclusions:

• Successfully established standardized landmark protocols for neuroanatomic phenotype assessment.
• The developed protocols demonstrate minimal error, suitable for reliable statistical shape analysis.
• These validated protocols can be adopted by researchers to enhance the consistency and accuracy of brain shape studies.