Adjusting for age-dependent bias in brain-age frameworks

Takeaway

  • A recently developed age-dependent bias-adjustment scheme for brain-age framework removes age dependency from predicted brain age results and results in statistically improved results for brain-age estimation frameworks.

Why this matters

  • Machine learning and brain imaging have been used to develop brain age-delta scores that estimate cognitively healthy aging and can be used clinically to identify individuals with abnormal cognitive aging.

  • Recently, an age-dependent bias in these scores has been detected; therefore, it is necessary to develop an accurate brain age prediction framework to mediate this bias.