EUVIMEDOdkrijte EUVIMED
Open access

Interpretable ROI Identification in Brain Image Analysis: Overcoming CNN Black Box Challenges With Kriging‐Enhanced Adaptive Sampling

HyunAh Lee iD, Jihnhee Yu iD, Soyun Park iD, Zohi Sternberg

Statistics in Medicine · 2026-07 · 45 (15-17)

DOI: 10.1002/sim.70653 ↗

At a glance

Journal article in Statistics in Medicine, published online on 2026-06-23. Written by HyunAh Lee and 3 co-authors. Institution: Department of Biostatistics, The State University of New York, University at Buffalo, Buffalo, New York, USA. Major MEDLINE topics: Brain, Image Processing, Computer-Assisted, Neural Networks, Computer. 69 references. The full text is freely available in PubMed Central. Funded by: NINDS NIH HHS.

PovzetekSubjectsBibliografski podatkiCiteRelated studies

Abstract (original language)

Brain image analysis presents significant challenges due to limitations in precision, computational efficiency, and interpretability. Although neural networks have proven effective for modeling complex patterns, they often function as black‐box systems, making their predictions difficult to interpret and limiting their clinical utility. To address these challenges, we propose the adaptive spatial key‐region identification (ASKRI) framework—a novel method to identify region of interest, which combines adaptive sampling based on Shannon entropy, probability‐mean‐driven selection, and spatial uncertainty quantified via kriging method. ASKRI integrates block‐to‐block kriging with statistical inference to interpolate CNN‐derived classification performance, significantly reducing the computational burden of exhaustive model training without sacrificing predictive accuracy. Designed for seamless integration with convolutional neural networks (CNNs), ASKRI enhances both the accuracy and interpretability of ROI identification. Its effectiveness is demonstrated using the traumatic brain injury (TRACK‐TBI) dataset, where ASKRI reliably identifies spatially consistent and biologically meaningful regions associated with aging. These results underscore the framework's potential to advance brain image analysis, while offering transparent and resource‐efficient diagnostic support in clinical settings.

Subjects

MeSH terms (MEDLINE)

BrainImage Processing, Computer-AssistedNeural Networks, ComputerConvolutional Neural NetworksHumansBrain Injuries, TraumaticMagnetic Resonance Imaging

Author keywords

adaptive samplingblack‐box problemblock‐to‐block krigingcomputational efficiencyconvolutional neural networks (CNNs)regions of interest (ROI)

Authors and institutions

  • Department of Biostatistics, The State University of New York, University at Buffalo, Buffalo, New York, USA.
  • Global Biometrics and Data Science, Bristol Myers Squibb, Lawrence Township, New Jersey, USA.
  • Department of Neurology, The State University of New York, University at Buffalo, Buffalo, New York, USA.

ORCID profiles: HyunAh Lee, Jihnhee Yu, Soyun Park

Funding

  • NINDS NIH HHS

Cite

APA

Lee, H., Yu, J., Park, S., & Sternberg, Z. (2026). Interpretable ROI Identification in Brain Image Analysis: Overcoming CNN Black Box Challenges With Kriging‐Enhanced Adaptive Sampling. Statistics in Medicine, 45(15-17). https://doi.org/10.1002/sim.70653

Vancouver

Lee H, Yu J, Park S, Sternberg Z. Interpretable ROI Identification in Brain Image Analysis: Overcoming CNN Black Box Challenges With Kriging‐Enhanced Adaptive Sampling. Statistics in Medicine. 2026;45(15-17). doi:10.1002/sim.70653

BibTeX
@article{Lee2026,
  author = {Lee, HyunAh and Yu, Jihnhee and Park, Soyun and Sternberg, Zohi},
  title = {Interpretable ROI Identification in Brain Image Analysis: Overcoming CNN Black Box Challenges With Kriging‐Enhanced Adaptive Sampling},
  journal = {Statistics in Medicine},
  year = {2026},
  volume = {45},
  number = {15-17},
  doi = {10.1002/sim.70653}
}
Start for free →