A brain-stimulation study reported a significant GABA-related reduction within the active-treatment group, but the active-versus-sham difference in change was not significant. The study illustrates why a significant result in one group and a nonsignificant result in another cannot substitute for a direct comparison.
Su and colleagues randomized 50 methamphetamine-dependent patients to four weeks of active or sham repetitive transcranial magnetic stimulation using an intermittent theta-burst protocol. Magnetic resonance spectroscopy measured GABA and the combined glutamate/glutamine signal, Glx, in the left dorsolateral prefrontal cortex. Results were expressed relative to N-acetylaspartate, or NAA. These were brain-metabolite ratios, not blood or urine drug concentrations.
GABA/NAA declined significantly within the active group (p = 0.040), while the sham group’s within-group change was nonsignificant. However, the time-by-group interaction for GABA/NAA was nonsignificant (p = 0.695), and the post-treatment group comparison was also nonsignificant. The data therefore did not establish a greater GABA/NAA reduction with active treatment.
Glx/NAA declined significantly within the sham group rather than the active group, but its time-by-group interaction was likewise nonsignificant. By contrast, craving and verbal-learning scores did show significant group-by-time interactions. A correlation between GABA-related change and a cognitive measure does not prove that a metabolite change caused or mediated the clinical result.
A ratio can change through its numerator, denominator or both, and a regional spectroscopy measurement does not directly quantify synaptic neurotransmission throughout the brain. The sample was small and only one target region was examined. The findings warrant further mechanistic research, but they do not validate GABA/NAA as an individual treatment-response test or an impairment marker. A careful account keeps the controlled clinical comparisons, within-group changes and exploratory correlations separate.
See also interpreting glutamate-related spectroscopy measurements.