Can an fNIRS Brain Scan Prove Alcohol Impairment?

Answer: A controlled alcohol experiment found that a moderate drinking condition worsened performance on a laboratory response-inhibition task and altered measured prefrontal hemodynamic asymmetry. It did not establish that a wearable optical brain scan can identify a specific person’s blood alcohol concentration, diagnose legal impairment, or determine why a real-world event occurred. The study measured alcohol in breath, converted that result to a blood-alcohol-equivalent display, calculated brain-activity metrics from optical signals, and separately measured task behavior. Those are four distinct evidentiary steps.

The original research is Dubois and colleagues, Change in brain asymmetry reflects level of acute alcohol intoxication and impacts on inhibitory control, Scientific Reports (2023), DOI 10.1038/s41598-023-37305-8. The central forensic question is not whether alcohol can affect the brain; the experiment supports that possibility under its conditions. It is whether optical brain lateralization is a validated substitute for a specimen measurement or a person-specific functional examination. This research does not make that substitution.

What did the experiment actually test?

The investigators used a single-blind, randomized, placebo-controlled, repeated-visit laboratory design. Forty-eight healthy social drinkers were included; 46 completed all three visits, and analysis denominators varied because some visits or recordings were missing. Participants undertook a Go/No-go task before and after beverage administration during placebo, low-dose and moderate-dose sessions. The targeted equivalent blood alcohol concentrations were 0.00%, 0.04% and 0.08%, respectively, but the mean achieved alcohol levels were below their low- and moderate-dose targets.

The experimental task required participants to respond during “Go” trials and withhold a response during “No-go” trials. The principal sensitivity measure, d-prime, summarizes discrimination between signals and non-signals using correct responses and errors. A change in d-prime is task-specific performance evidence. It is not itself a measured ethanol concentration, a roadside driving score, an accident-reconstruction finding, or a clinical diagnosis.

While performing the task, participants wore a time-domain functional near-infrared spectroscopy (TD-fNIRS) headset. The device recorded properties of light traveling through head tissues; its processed signals were used to infer relative changes in oxygenated and deoxygenated hemoglobin and their sum. Researchers then modeled responses in right and left prefrontal regions. These processed responses are hemodynamic proxies for task-associated activity, not direct recordings of neuronal electrical firing.

Was alcohol measured in breath or in whole blood?

Alcohol monitoring used an Intoxilyzer 800 breath instrument. The researchers reported breath-alcohol results in equivalent-BAC terms. This should not be misdescribed as independent measurement of ethanol in a collected venous or arterial whole-blood specimen. The distinction matters whenever a scientific statement is translated into a legal claim about a particular person’s blood.

Baseline breath readings were required to be 0.00 on the instrument. The study’s placebo presentation included juice with vodka sprayed onto the cup lid to support blinding; a displayed zero should not be expanded into a claim that no molecule of ethanol was involved in the placebo presentation. Repeated breath checks informed when the post-beverage task began. These checks describe an experimental exposure protocol, not a validated one-to-one conversion for every individual.

For method-specific context, see Breath Alcohol Measurement and Pharmacokinetics in Forensic Toxicology. Neither general guide substitutes for the instrument records or the protocol of the particular study.

Did alcohol worsen response inhibition?

Yes, on the measured task under the moderate condition. The dose-by-before/after interaction for Go/No-go d-prime was reported as p = 0.003. In the moderate-dose session, post-beverage d-prime was lower than pre-beverage performance (p = 8.13 × 10⁻⁴, with 47 sessions in that within-condition comparison). The researchers did not find a significant comparable change under the low-dose condition. A comparison of post-task performance also distinguished placebo from moderate alcohol.

The result should be described exactly: a group-level effect on laboratory response inhibition, not universal motor slowing. The dose-by-time interaction for reaction time was not significant (p = 0.200), although there was an overall pre/post time effect. It would be inaccurate to say that this experiment established that all participants’ reactions became slower in proportion to alcohol dose. Different behavioral outcome measures answered different questions.

What did optical brain asymmetry show?

During high-inhibition Go/No-go blocks, the study generally found relatively greater activity in right than left prefrontal regions. Alcohol dose altered the strength and timing of that lateralization. For example, an averaged, generalized-linear-model measure of right-minus-left prefrontal activity differed between the low and moderate alcohol conditions (p = 0.028). The moderate condition showed weaker evidence of the typical right-sided difference. Other time-resolved measures distinguished certain placebo-versus-alcohol contrasts.

These are statistical contrasts of processed optical hemodynamic signals. Their units and meaning cannot be conflated with blood ethanol, breath ethanol, heart rate, a chemical biomarker of exposure, or direct proof of incapacity. The processing involved model coefficients, left/right comparisons, direction and angle metrics, and choices about hemoglobin-related signals. Their reliability as an individual forensic classifier was not established.

Some relationships were small: across sessions, change in behavioral performance correlated with the direction measure of lateralization at r = 0.22 (p = 0.00747), while the reported correlation between breath-derived BAC equivalent and performance change was r = −0.29 (p = 0.000524). The authors reported no significant direct association between BAC-equivalent levels and the lateralization metrics in those particular tests (p > 0.1). Thus, brain signal and concentration are neither interchangeable nor necessarily redundant.

Evidence table: measurements, models and defensible conclusions

Why does a brain-signal association not prove individual impairment?

Several inferential steps separate the published group results from an individual forensic opinion. The experiment measured relative optical hemodynamics during a particular screen-based inhibition task; it was not designed as a field sobriety validation or a prospective classifier for legally impaired drivers. The researchers did not establish sensitivity, specificity, likelihood ratios, or a validated individual decision cutoff for a TD-fNIRS test applied to an unknown person. Nor did they test whether the same scan could isolate alcohol from sleep deprivation, medication effects, neurological disease, head motion, skin/scalp circulation changes, or individual baseline differences in casework.

An observed correlation is not a mechanistic proof of why any particular error occurred. A task-related signal may vary with cognitive engagement, baseline performance, attention, effort and hemodynamic physiology. Even the source’s adjusted statistical models explain variation in the experimental outcome; those estimates cannot be exported as a legal algorithm without independent validation and an identified population, endpoint and decision rule.

The study investigators noted that dose levels were below targets, achieved readings overlapped across conditions, roughly half of participants exhibited some placebo expectancy response, and some statistics might not remain significant after more stringent multiple-comparison adjustment. Handedness was not fully modeled because only five participants were left-handed, and alcohol-related circulatory changes are a potential optical-signal confound. These limitations matter particularly when a result is offered as unusually specific or definitive.

How do subjective intoxication, task performance and legal impairment differ?

Participants also rated alcohol’s stimulant and sedative subjective effects and estimated how much they had consumed. Those reports changed with assigned dose. A subjective feeling is a measured report about perception, not the same outcome as a failed No-go trial, a breath instrument reading or a real-world operating error.

The Go/No-go task provides observable performance evidence for attention and response withholding under defined conditions. It does not measure every cognitive or psychomotor function involved in driving, machinery operation or consent-related judgments. Whether a case meets a statutory alcohol threshold is a jurisdiction- and evidence-specific legal issue; the authors’ use of the term “BAC” for breath-derived equivalent readings must not be mistaken for a direct whole-blood assay in a forensic record.

Likewise, a physiological correlate of performance is not proof of causation. An incident may depend on many other contributing circumstances, and a forensic opinion about a specific person requires a case-specific timeline, validated analytical results, clinical and behavioral observations, alternative explanations and articulated uncertainty.

Study-design and generalizability limits

This was a relatively small repeated-measures experiment in screened healthy adult social drinkers. Forty-eight participants contributed data, but paired sample sizes differed across contrasts, and not every participant completed all three visits. Measurements were made near the ascending/peak breath-alcohol phase and in a controlled task environment. The experiment did not establish delayed next-day effects, emergency-department diagnosis, chronic alcoholism effects, accident risk, or whether a handheld version would work as a courtroom-ready device.

The article’s optical approach is promising for investigating how cortical physiology accompanies behavioral control, but it remains a research method. Its evidence cannot be broadened into a claim that brain scanning reliably reads someone’s BAC or distinguishes voluntary intoxication from any other cause of poor inhibitory control.

Forensic review: which primary records answer which question?

Frequently asked questions

Can fNIRS directly measure blood alcohol concentration?

No. The device monitors optical properties from which relative tissue hemodynamic changes are inferred. Ethanol was separately assessed using a breath instrument and expressed as BAC equivalent in the study.

Did the moderate alcohol dose impair a laboratory task?

Yes, moderate-dose Go/No-go discrimination performance deteriorated in the reported within-session comparison. That does not establish a universal impairment threshold for an individual.

Did every alcohol dose slow reaction times?

No. The study reported no significant dose-by-time interaction for reaction time, even though d-prime showed a significant interaction and the moderate condition worsened on that measure.

Could brain asymmetry replace an ethanol test in court?

Not on this evidence. It was neither an assay of ethanol nor validated as an individual forensic classifier or substitute for the specimen-specific tests used in a particular case.

Does a significant brain–behavior correlation mean the brain measurement caused the impairment?

No. Correlation across laboratory observations does not isolate a causal mechanism for a specific person or event. The optical metric is processed and potentially affected by physiological confounders.

Did the placebo reading establish zero ethanol molecules?

No. The placebo protocol used a vodka-sprayed cup lid. Instrument displays and analytical detectability must be distinguished from an absolute molecular claim.

What is the appropriate case-specific conclusion?

The study supports a limited proposition that moderate acute alcohol exposure can coincide with poorer response inhibition and altered prefrontal hemodynamic lateralization in this laboratory population. It does not establish a particular person’s whole-blood alcohol concentration, statutory status, driving ability or causal responsibility.

Primary scientific source and further reading

Dubois J, Field RM, Jawhar S, et al. Change in brain asymmetry reflects level of acute alcohol intoxication and impacts on inhibitory control. Scientific Reports. 2023;13:10278. https://doi.org/10.1038/s41598-023-37305-8.

For interpretation of the underlying specimen and timing questions, see the site’s breath-alcohol method guide, pharmacokinetics guide and evidence library. These links address separate analytical principles; they do not replace the original experimental dataset.