As biomedical imaging has become a cornerstone of medical practice and research, the development and application of appropriate analytics for conducting clinical epidemiology is critical. The complex structure of imaging data can render classical statistical approaches inadequate. Indeed, even state-of-the-art specialized imaging statistics approaches have been criticized heavily in recent years for inflated type I error rates leading less reproducible results, especially in neuroscience. Furthermore, the immersive and transdisciplinary training of the next generation of methodologists, applied statisticians, and epidemiologists in these tools is critical for ensuring scientific rigor in imaging-based findings, as well as the development of generalizable imaging biomarkers.
The Penn Statistics in Imaging and Visualization Endeavor (PennSIVE), a Center of Excellence based in the Center for Clinical Epidemiology and Biostatistics (CCEB), includes laboratories developing and applying biostatistical methods for image analysis. Established in 2012, PennSIVE has been at the forefront of developing new frameworks for understanding and mitigating fundamental epidemiologic biases in multi-center and observational imaging studies. By collaborating with psychiatrists, neurologists, radiologists, and basic neuroscientists, investigators at PennSIVE have been building and applying the next generation of tools necessary for imaging epidemiology. Through immersive training programs at the pre- and post-doctoral levels, PennSIVE aims to produce a generation of highly sought-after imaging statisticians, as well as a network of co-trained bioengineers, clinician-scientists, and neuroscientists.