Can a Better Cognitive Retest Reconstruct an Earlier Alcohol Level?

No. Improvement on a later cognitive test cannot, by itself, reconstruct an earlier alcohol concentration or identify why the earlier performance was worse. A 2020 study found slower responses in hungover students who improved when retested about a week later. It also illustrates why hangover, sleep and repeated testing need attention before attributing a score change to a single cause.

In a DUI review, a later better performance may be relevant context. Converting that improvement into an earlier BAC, a precise recovery time or proof of alcohol as the sole cause goes beyond this evidence.

What the repeated-testing design added

Murgia and colleagues enrolled 36 university students, divided equally into a hangover group and a comparison group. At the first visit, the hangover group came after a drinking occasion; the comparison group came after alcohol abstinence and normal rest. Both groups returned approximately one week later in a non-hangover condition.

Participants completed three computer tasks: odd/even number judgments, a version requiring occasional reversal of the response rule, and arithmetic verification. The stimuli remained visible until a response, allowing participants to spend additional time producing an answer.

The investigators assessed drinking history, symptoms and sleep. All participants had a zero reading on the breath instrument on the testing morning. This was a breath result, not direct blood analysis and not a measurement of the previous night’s peak.

Slower responses did not mean more errors

The study reported a pattern of slower response times in the initially hungover group and improvement when that group returned without a hangover. The group-by-time interactions support a difference in the pattern of change across visits, rather than merely observing that everybody performed the test twice.

The accuracy analyses did not reveal significant group differences. Because the tasks had no response deadline, participants could preserve accuracy by responding more slowly. The suggestion that adding deadlines would instead increase errors was a proposed explanation, not an intervention tested in this experiment.

Nor did the study identify a single isolated cognitive deficit. Each task involved several operations, including perception, response selection and memory. A slower response could not be converted directly into a specific amount of working-memory loss.

Sleep changed alongside hangover

The initially hungover group reported approximately 5.10 hours of sleep before the first visit and 7.38 hours before the second. The comparison group’s corresponding averages were approximately 6.97 and 7.08 hours.

Alcohol exposure, symptoms and reduced sleep therefore occurred together in the naturalistic hangover condition. The study did not independently randomize alcohol and sleep loss. It cannot determine how much of an individual’s slowing arose from each component.

Repeated testing adds another issue: familiarity may improve performance. Including a retested comparison group helps assess that problem, but it does not turn the study into an individually calibrated BAC estimator. Participants also chose when to attend, so selection and expectation remain possible influences.

Why this matters for DUI evidence

A comparison across dates should document what differed between them: symptoms, sleep, test conditions, timing, practice and any measured alcohol. A better later score may support the possibility that an earlier state was temporary. It does not specify the earlier alcohol concentration, establish performance while actually driving or identify a legal violation.

Similarly, a zero breath result at a later examination cannot establish that the person had no alcohol earlier. The site’s analysis of retrospective BAC estimation addresses the separate assumptions required to estimate an earlier concentration from suitable concentration-and-time evidence.

The present question also differs from whether hangover affects remembering future actions. Here, the value of the study is its repeated-measurement design and the warning against assigning a change in performance to one uncontrolled factor.

A limited but useful conclusion

The sample consisted of university students, not a representative sample of drivers. There was no on-road or simulator driving outcome, no experimentally controlled previous-night dose, and no measured previous-night blood alcohol curve. Reported drinking amounts were estimates based on self-report.

The study supports checking recent drinking, hangover and sleep when interpreting cognitive comparisons. It does not support treating the difference between two scores as an alcohol meter.

Primary source

Murgia M, Mingolo S, Prpic V, Sors F, Santoro I, Bilotta E, Agostini T. University Students’ Hangover May Affect Cognitive Research. Frontiers in Psychology. 2020;11:573291. https://doi.org/10.3389/fpsyg.2020.573291.