Generalizable MRI normative modelling to detect age-inappropriate neurodegeneration.

Area of Science:

• Neuroimaging
• Biomarkers
• Neurodegenerative Diseases

Background:

• Assessing age-appropriate brain atrophy from MRI scans is challenging.
• Automated structural MRI metrics show promise for neurodegeneration but lack standardized reference data.
• Normative modeling offers a solution by creating individualized, age- and sex-adjusted centile scores for MRI metrics.

Purpose of the Study:

• To develop and validate brain charts for individual-level assessment of brain structure.
• To investigate the relationship between brain charting centile scores, cognitive performance, and neuropathology.
• To apply brain charting to identify neurodegenerative atrophy in Alzheimer's disease and frontotemporal lobar degeneration.

Main Methods:

• Developed regional cortical thickness and amygdala/hippocampal volume brain charts using normative data from 56,173 participants.
• Applied brain charts to assess individual centile scores and group-level relationships with cognitive performance (mini-mental state examination).
• Validated brain charting in cohorts with Alzheimer's disease (ADNI-3, NACC) and frontotemporal lobar degeneration (NFLD).

Main Results:

• Brain charting successfully demonstrated individualized atrophy patterns in clinical cases.
• Group-level analysis showed regional centile scores predicted cognitive performance and tau deposition in Alzheimer's disease.
• Brain charting discriminated Alzheimer's disease groups from controls and identified distinct atrophy patterns in frontotemporal lobar degeneration phenotypes.

Conclusions:

• Regional centile scores from normative brain charting provide a generalizable method for objective atrophy identification.
• Brain charting enables the determination of neurodegenerative atrophy at the individual level.
• This approach enhances the interpretation of MRI scans for neurodegenerative disease diagnosis.