A brain scan obtained during a thinking task also reflects how the person performs that task. If one group responds more slowly, an activation difference may partly reflect additional time spent processing each stimulus. A small 2011 methamphetamine study illustrates how accounting for this issue can change the reported imaging pattern.
Who was studied, and when?
Nestor and colleagues compared 10 methamphetamine-dependent adults with 18 comparison adults. The methamphetamine group had been abstinent for four to seven days. Participants performed a color-word Stroop task during functional MRI: they identified the ink color while ignoring the word, with matching and conflicting stimuli presented in separate blocks.
This was an observational human study during early abstinence, not an experiment administering a known methamphetamine dose. Screening and admission urine tests supported recent use in the dependent group. All participants tested negative for illicit drugs and breath alcohol on the testing day. The paper did not establish a blood methamphetamine concentration associated with any scan finding.
The behavioral difference came first
The methamphetamine group responded more slowly in both matching and conflicting conditions. Its increase in response time from matching to conflicting trials was also larger. Participants in that group made more errors overall, but the group-by-condition interaction for errors was not significant.
These results distinguish general slowing, conflict-related slowing and errors. They are related outcomes, but an error difference across both conditions does not establish a selectively larger conflict-related error effect.
What happened before response-time adjustment?
In the initial imaging comparison, group differences appeared for the contrast subtracting matching from conflicting trials. The dependent group showed lower activation in part of the left cerebellum and higher activation in several other regions, including posterior cingulate and medial frontal areas.
Reporting only this analysis could suggest that the main result was increased activation in the methamphetamine group. That would omit the investigators’ subsequent effort to address performance differences.
What changed after adjustment?
The investigators then controlled statistically for response time to address time spent on task. In this analysis, significant group differences appeared during the conflicting condition: the methamphetamine group showed lower activation in regions including the right inferior frontal gyrus, supplementary motor/anterior cingulate region and anterior insula. No region showed greater activation in that group in the adjusted comparison.
The adjusted and unadjusted analyses should therefore be named separately. Adjustment changes the question being estimated; it does not turn a cross-sectional comparison into proof of drug-caused tissue damage or reveal a person’s original brain function before drug use.
How far can these findings travel?
Only 10 methamphetamine-dependent participants were studied. Seven had lifetime histories of other substance abuse or dependence. Early withdrawal, unmeasured background differences and the particular task design remain relevant limitations.
The blocked design also prevented the investigators from separately examining trial-to-trial adjustments after conflict or distinguishing every successful from unsuccessful conflict-resolution event. Results from a task presenting stimuli in a different sequence may address a different aspect of control.
For an expert reviewing this historical evidence, the useful question is whether the reported brain difference was adjusted for the observed performance difference, and which contrast produced it. Neither analysis validated an individual test of intoxication, driving impairment or permanent injury.
Historical primary source
Nestor LJ, Ghahremani DG, Monterosso J, London ED. Prefrontal hypoactivation during cognitive control in early abstinent methamphetamine-dependent subjects. Psychiatry Research: Neuroimaging. 2011;194(3):287–295. doi:10.1016/j.pscychresns.2011.04.010. Original commentary on the supplied paper; no source figures or tables reproduced.
Leave a Reply