B0176
Title: Mitigating inter-scanner biases in high-dimensional neuroimaging data via spatial Gaussian process
Authors: Jun Young Park - University of Toronto (Canada) [presenting]
Rongqian Zhang - University of Toronto (Canada)
Abstract: In neuroimaging studies, combining data collected from multiple study sites or scanners is becoming common to increase the reproducibility of scientific discoveries. At the same time, unwanted variations arise by using different scanners (inter-scanner biases), which need to be corrected before downstream analyses. While statistical harmonization methods such as ComBat have become popular in mitigating inter-scanner biases in neuroimaging, recent methodological advances have shown that harmonizing heterogeneous covariances results in higher data quality. A new statistical harmonization method is proposed called SAN-GP (spatial autocorrelation normalization via Gaussian process) that preserves homogeneous covariance vertex-level cortical thickness data across different scanners. SAN-GP uses an explicit Gaussian process to characterize scanner-invariant and scanner-specific variations to reconstruct spatially homogeneous data across scanners. SAN-GP is computationally efficient, and it easily allows the integration of existing harmonization methods. The utility of the proposed method is demonstrated using cortical thickness data from the social processes initiative in the neurobiology of the schizophrenia(s) (SPINS) study.
