In this application, we propose a strategy to remove physiologic noise from the cardiac and respiratory processes as sources of bias and reduced specificity in functional magnetic resonance imaging (MRI) and functional connectivity MRI data. These noise sources are difficult to correct for without special equipment for parallel monitoring and it is difficult to ascertain the success of the correction process itself. As a result, many researchers necessarily ignore these sources or use amelioration methods that are not specific to these sources. Due to the size and variability of these sources across populations, this represents a critical barrier to progress in the field. However, recent progress in physiologic estimation and correction by the investigators has produced a tool to enable researchers to retrospectively correct their data, in a manner equivalent to correcting their data with parallel monitored noise sources. This is an important advance, but only a small set of researchers are currently using these tools. The reasons are twofold: 1) the tool is not easy to use and requires additional software and 2) past experience with physiologic correction is limited to those investigators who have access to monitoring equipment. To counter these problems, we propose to improve the integration of our tools with the Analysis of Functional NeuroImages (AFNI) library such that at the conclusion of our project physiologic correction is as easy to apply as volumetric motion correction currently is. In addition, we propose to produce validated physiologic estimators for the bulk of the publicly-available data maintained by the Functional Connectomes Project and its most recent initiative, the International Neuroimaging Data-sharing Initiative. With the recent increase in analyses of publicly-available MRI data, this project will dramatically increase community experience with physiologic correction and enable a shift from the analysis and reporting of physiologic- corrupted data to the analysis and reporting of physiologic-uncorrupted data.
Physiologic noise from cardiac and respiratory processes obscures useful signals in functional magnetic resonance imaging (MRI) data. The size of this noise varies randomly between individuals and systematically between disease populations and states. However it is difficult to correct for its presence without special equipment so many researchers necessarily ignore it, making it a critical barrier to progress in the field. This projct will remove that barrier by providing this correction for public neuroimaging data and a tool to correct researchers'and clinicians'data without special equipment.