Quantifying the effects of head movement on magnetic resonance spectroscopy estimates of gamma-aminobutyric acid

There has been keen interest in measuring in vivo GABA. However, GABA signal is low and typically measured using techniques vulnerable to the confounding effects of in-scanner head movement. This issue is particularly problematic for clinical studies since it may lead to Type I or II error in testing for group differences. While solutions to mitigate the effects of movement have been proposed, fundamental and largely unexamined issues are the nature and scale of this effect. We developed a method to quantify and characterize head movement during GABA spectroscopy and found that two parameters of movement, displacement and instantaneous movement, were inversely correlated with and accounted for 12.1% and 20.2% variance of GABA estimates respectively. We conclude that head movement can significantly affect GABA measurements and the application of methods to account for movement may improve of GABA measurement accuracy and the detection of true group differences in clinical studies.