An Examination of Timing Variables in Liquid Taste Administration fMRI Paradigms

Abstract

Use of liquid administration paradigms in functional magnetic resonance imaging (fMRI) research has rapidly increased and expanded. Despite rapid expansion, there are no studies examining the methods setup of this paradigm. The present study aims to address basic questions about the methodology of liquid administration paradigms using a series of three experiments. These questions include whether using timing files generated by the study computer accurately represent the time of taste receipt, whether variation in methods setup meaningfully impacts time of taste receipt, and whether inaccurate receipt time impacts first-level analysis of fMRI data. Experiment 1 used an equipment setup simulating the running of an fMRI study with three taste stimuli; tasteless solution, chocolate milkshake, and chocolate milk. For each taste stimulus, we captured and compared the times of administration recorded by the study computer and the time of receipt captured using two measurement systems. Experiment 2 used a similar setup with only one taste stimulus: chocolate milkshake. Using high-definition video, we captured the time elapsed as the taste stimulus passed through the tubing at three infusion speeds and varying temperatures. In Experiment 3, we used the fMRI data from a previous study, in which five participants were administered chocolate milkshake to demonstrate how the lag times captured in Experiment 1 affect first-level fMRI data analyses. There were significant differences in time of administration recorded by the computer and time of receipt, demonstrating the existence of a lag time. This lag time was greatest for chocolate milkshake. Semiautomatic pump speed also was significantly associated with lag time. Temperature and length of tubing were not significantly related to lag time. Additionally, our fMRI analyses show that the variation in the methods setup have a meaningful impact on first-level fMRI analyses. When using the computer-generated time that was not corrected for lag time, increased neural activity in response to the chocolate milkshake taste was shown in three out of the five participants. When correcting using the lag time, increased neural activity in response to chocolate milkshake taste was found in all five participants. Overall, this study demonstrates a “proof of concept” that the practice of estimating liquid administration times from computer-generated files may not be accurate, that variation in methods setup may affect the timing in this paradigm, and that these differences in timing introduce variation in the fMRI data of individual participants.