Rxivist logo

The Cost of Untracked Diversity in Brain-Imaging Prediction

By Oualid Benkarim, Casey Paquola, Bo-yong Park, Valeria Kebets, Seok-Jun Hong, Reinder Vos de Wael, Shaoshi Zhang, B.T. Thomas Yeo, Michael Eickenberg, Tian Ge, Jean-Baptiste Poline, Boris C. Bernhardt, Danilo Bzdok

Posted 17 Jun 2021
bioRxiv DOI: 10.1101/2021.06.16.448764

Brain-imaging research enjoys increasing adoption of supervised machine learning for single-subject disease classification. Yet, the success of these algorithms likely depends on population diversity, including demographic differences and other factors that may be outside of primary scientific interest. Here, we capitalize on propensity scores as a composite confound index to quantify diversity due to major sources of population stratification. We delineate the impact of population heterogeneity on the predictive accuracy and pattern stability in two separate clinical cohorts: the Autism Brain Imaging Data Exchange (ABIDE, n=297) and the Healthy Brain Network (HBN, n=551). Across various analysis scenarios, our results uncover the extent to which cross-validated prediction performances are interlocked with diversity. The instability of extracted brain patterns attributable to diversity is located preferentially to the default mode network. Our collective findings highlight the limitations of prevailing deconfounding practices in mitigating the full consequences of population diversity.

Download data

  • Downloaded 528 times
  • Download rankings, all-time:
    • Site-wide: 82,001
    • In neuroscience: 14,050
  • Year to date:
    • Site-wide: 34,056
  • Since beginning of last month:
    • Site-wide: 52,897

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide