(Un)common space in infant neuroimaging studies: a systematic review of infant templates

• In Event: Poster Session 07
  In Poster Session: PS 07 Topic - Biological Processes: Neuroscience and Genetics

Abstract

Objective: In neuroimaging studies, spatial normalization is an important step of image processing that maps an individual’s brain onto a template brain permitting downstream statistical analyses. For adult studies, the MNI template is the established reference brain, allowing for easier comparisons across studies, data sharing, and meta-analyses. However, in infant neuroimaging, there remain several technical challenges that have prevented the establishment of a standardized template for spatial normalization. For example, the MNI template was constructed from adult data and is not suited for smaller infant brains. Also, brain morphometry rapidly changes through infancy, implying that multiple templates may be needed over development. As such, many different approaches are used throughout the literature, including analyzing neuroimaging data in native space (i.e. forgoing spatial normalization), creating a study specific template, or using one of the several available templates. To quantify the popularity and variability of these approaches in infant neuroimaging studies, we performed a systematic review of infant MRI studies from 2000 and later. Our goal is that, by quantifying what the most popular approaches are, standardized templates in infant neuroimaging can be established.

Methods: A PubMed search was performed to gather relevant articles, using the following search terms: “infant MRI” and “neonatal MRI”, “neonatal ‘fmri’”, “toddler ‘fmri’”, “’toddler fmri’”, “preterm fmri”, “neonate(s)”, “infant(s)”, “(((infant) OR (neonate)) Or (newborn)) AND ((fmri) OR (MRI) OR (DTI))”. Our initial search returned 37,782 articles. We excluded duplicate articles, articles from before the year 2000, case reports, articles written in languages other than English, articles for animal studies, methodological articles, reviews articles, articles using subjects older than 18 months, articles using fetal MRI, and articles using imaging modalities other than MRI. Here, we present results from the first 101 papers meeting inclusion criteria. Articles were classified into 1) processing data in native space (i.e. forgoing spatial normalization), 2) creating a study specific template, or 3) using a predefined template.

Results: Our preliminary analysis shows that native space processing was used in over 50% of the studies (native space: 53.5%, study specific: 18.8%; predefined: 27.72%, X2=8.9, p=.01). Of predefined templates, the JHU Neonate Atlas (n=9) was the most commonly used. The preferred approach varied over time. Studies published prior to 2010 overwhelmingly used native space compared to using a template (either predefined or study specific; native space: 91.4%, template: 8.6%, X2=11.8, p<.001). In contrast, studies published after 2010 primarily used a template; though native space processing remained popular (native space: 33.3%, template: 66.7%, X2=5.3, p=.02).
Conclusions: Using a systematic review of infant neuroimaging studies, we highlight a lack of an established template brain in these studies. Our work only contains 101 studies so far. Thus, it remains important future work to continue to code the other studies from our search. Given that infant neuroimaging is maturing in its use of image processing methods, establishing standardized templates is a needed next step in the field.

Authors

• C. Alice Hahn, Yale University
  Presenting Author

• Silvia Gini, University College London and Yale University
  Non-Presenting Author

• Alexis Alfano, Quinnipiac University
  Non-Presenting Author

• Hannah Peterson, Yale University
  Non-Presenting Author

• Saloni Mehta
  Non-Presenting Author

• Alexander Dufford, Northwestern University
  Non-Presenting Author

• Dustin Scheinost, Yale School of Medicine
  Non-Presenting Author