Can Methamphetamine Narrow Driving Gaze Without Increasing Lane Weaving? A 20-Driver Alcohol Crossover

Answer: Yes—within one controlled 20-person simulator experiment, oral methamphetamine changed how participants scanned the roadway without increasing the measured variation in lane position. Methamphetamine alone produced a significantly narrower spatial distribution of visual attention and less orderly gaze transitions compared with placebo. In contrast, a separate low-dose alcohol condition increased the simulator’s lane-position variability. Those findings are different measurements, not contradictory conclusions about an individual’s ability to drive safely. The study neither established a validated eye-tracking test for intoxication nor demonstrated that methamphetamine makes real-world driving safer.

The primary evidence is Hayley and colleagues’ randomized, double-blind, four-condition crossover study, “Acute methamphetamine and alcohol usage alters gaze behaviour during driving” (2024). Its distinguishing contribution was simultaneous recording of eye movements, simulated vehicle control, and directly assayed venous whole-blood methamphetamine and ethanol concentrations. Understanding all three layers is essential before using a gaze pattern, lane-weaving result, or toxicology number in a legal or clinical interpretation.

What exactly did the experiment show?

Participants who received methamphetamine tended to make longer fixations, look within a more concentrated portion of the scene, and exhibit more variable transitions among viewed locations. The researchers quantified the spatial concentration as stationary gaze entropy and the sequence of movements as gaze transition entropy. Methamphetamine alone reduced the first metric and increased the second relative to placebo in the reported comparisons. This suggests a constrained yet less organized sampling pattern during the particular monotonous highway task. It is not a direct measurement of whether a pedestrian, hazard, or vehicle was actually missed.

The alcohol-only condition produced greater standard deviation of lateral position (SDLP) than placebo, methamphetamine alone, and the combination. The alcohol-only versus placebo model-estimated difference was 1.96 centimeters, with a 95% confidence interval from 0.98 to 2.95 centimeters and p < .001. Compared with alcohol alone, the methamphetamine-only and combined conditions had lower observed lane variation. This does not establish a protective treatment effect across complex roadway scenarios. It describes a specific simulator endpoint under a particular dosing schedule.

One apparent paradox therefore disappears when the endpoints are separated: an eye tracker can register altered visual search while a vehicle remains relatively centered in a simple lane. Keeping a simulated car in a lane does not require noticing and correctly prioritizing every possible emerging hazard.

Who participated, and how many independent observations were there?

Forty-one people were recruited. Twenty-one completed the visits and provided usable driving information; one was excluded after a protocol violation was detected during unblinding. The analyzed sample therefore comprised 20 distinct licensed adults, including 12 men and eight women, with a reported mean age of 29.5 years. All reported previous recreational exposure to an amphetamine-type stimulant, but participants were not a representative sample of chronically dependent, sleep-deprived, or injured motorists.

Each person attended four experimental conditions in randomized order with at least a week between visits. Four visits per person are four repeated treatment observations, not 80 independent drivers. The 50-Hz eye tracker and the many five- or ten-minute intervals within each one-hour drive likewise yield nested technical and temporal measurements. Statistical models can account for repeated observations, but their raw row counts must not be mistaken for independent participant counts.

The study excluded various medical conditions and medications, restricted recent alcohol and caffeine use, and screened for recent drug intake on visit arrival. These safeguards strengthen within-person comparisons, while simultaneously narrowing generalizability to ordinary roadside enforcement populations.

What substances and doses were administered?

The four arms were matched placebo; oral methamphetamine at 0.42 milligrams per kilogram; vodka-based alcohol intended to produce a blood-alcohol concentration near 0.04%; and oral methamphetamine with alcohol together. Methamphetamine capsules were prepared in standardized strengths, with weight-adjusted doses rounded and capped at approximately 40 milligrams. The reported average administered methamphetamine dose was 31.5 milligrams. The alcohol dose was weight-based, and the study reports an average administered ethanol amount of approximately 76.74 grams. Those are administered quantities, not measured circulating concentrations or an assurance that every person achieved the nominal BAC target.

The 60-minute highway driving scenario began roughly 1.5 to 2.5 hours after treatment. Participants were instructed to follow the designated lane, approach 100 kilometers per hour, and pass slower simulated traffic. The test did not include the full sensory, social, traffic, or consequence structure of on-road driving. Researchers measured the actual post-dose biological concentrations separately, rather than equating the swallowed dose with the assayed whole-blood value.

It is also important not to substitute findings from d-amphetamine, lisdexamfetamine, or a different methamphetamine-isomer experiment. This primary study addressed oral methamphetamine and ethanol together, with an eye-tracking and lane-keeping question—not a general ranking of stimulant medications or a roadside impairment threshold.

Was the toxicology measurement really whole blood?

Yes. The investigators collected approximately five milliliters of venous whole blood at about 30, 60, and 180 minutes after dosing and stored specimens frozen. For methamphetamine, their laboratory used liquid chromatography with tandem mass spectrometric detection on a triple-quadrupole system, using a deuterated methamphetamine internal standard. The reported whole-blood calibration span was 5 to 1,000 ng/mL, with a 5 ng/mL quantitation limit. For ethanol, the laboratory used a validated gas chromatographic method with flame-ionization detection, internal standardization, and reported quantitative coverage over the relevant forensic blood-alcohol range.

These details matter because legal reporting conventions vary by country and jurisdiction. In this case the biological specimen was actually collected and processed as whole blood; it was not automatically a centrifuged plasma value. The blood concentration unit for methamphetamine was ng/mL. Ethanol was reported as grams per deciliter, with 0.03 g/dL equivalent to 0.03% weight/volume. A previsit breath screening result was not the same analytical sample as the laboratory whole-blood result. Neither analyte should be converted into oral-fluid, breath, or plasma values without study-specific evidence.

Analytical validation supports the identification and quantitation of the chemical in that specimen, subject to the method’s characteristics. It does not by itself validate a person’s driving behavior, nor convert an eye-tracking profile into a measure of impairment.

What did the serial whole-blood concentrations show?

The time course displayed a useful pharmacological contrast. Mean whole-blood methamphetamine was much higher at the 180-minute sample than at the early samples, whereas mean ethanol was lower at 180 minutes than around 60 minutes. Importantly, similar mean methamphetamine concentrations were observed in the methamphetamine-only and combined treatment arms at the later time points. The ethanol-only and combination conditions likewise showed broadly comparable measured ethanol at those times in the original comparisons.

At around three hours, group mean methamphetamine values were about 92–93 ng/mL in both methamphetamine-exposed conditions. Those numbers are actual assay means within this study; they are not a driver-specific concentration-to-impairment equation. A whole-blood result observed in a patient or case cannot be assigned this laboratory group’s average trajectory without knowing the individual dose, route, sampling schedule, previous use, and other circumstances.

Similarly, the ethanol data should not be treated as if every participant maintained 0.04% BAC during the driving hour. The nominal drinking target and measured mean concentration are separate facts. The repeated specimen results also cannot establish an individual’s drinking or methamphetamine-use time merely by reversing the study’s averages.

What is lane weaving, and why did alcohol change it?

SDLP is the standard deviation of the driver’s lateral position over a defined section of a simulated route. It is often used to summarize how steadily a vehicle stays in its lane. In this study, intentional lane changes were removed before computing SDLP, and portions at the beginning and end of driving were excluded to reduce nonrepresentative transitions. Increased SDLP can signify poorer lane control under specified conditions, but SDLP is just one aspect of driving and depends on the roadway task, vehicle interface, traffic context, and analysis decisions.

Alcohol alone increased SDLP by 1.96 centimeters compared with placebo, and by 4.81 and 4.88 centimeters compared with the methamphetamine-only and combination arms, respectively; all these contrasts were reported at p < .001. The investigators also reported greater steering variability with alcohol alone than placebo (p = .018), methamphetamine alone (p < .001), and the combination (p = .03). For average speed, the alcohol condition showed a small reduction relative to placebo, approximately 0.46 km/h (p = .02).

By contrast, methamphetamine alone or methamphetamine with alcohol did not show the same increased lane weaving in this experimental highway setting. Calling that finding a universal improvement in driving would invert its meaning. The study’s gaze abnormalities may have been more sensitive than lane variation to stimulant-related visual-search changes, and a boring open-road simulation may not fully challenge rapid hazard anticipation or decision-making.

What did eye fixation measurements add?

A fixation is an interval during which gaze remains directed toward a particular location. The eye-tracking system estimated both the count of fixations per unit time and their average duration. Methamphetamine alone yielded a higher fixation rate than alcohol alone by a model-estimated 20.29 fixations per minute (p = .001), but this particular rate comparison was not statistically significant against placebo or the combined condition. The difference must be attributed to the pair actually compared—not described as a positive effect versus all other arms.

Average fixation duration increased by a reported 19 milliseconds with methamphetamine alone relative to placebo (p = .035). The researchers did not report corresponding significant increases for every active comparison. Neither a fixation count nor longer fixation duration specifies whether the driver correctly identified a road hazard. Looking repeatedly at lane markings can help maintain position while leaving important peripheral or upcoming events relatively unattended.

The study presented visual heatmaps illustrating how gaze clustered under various drug conditions, but a pictured heatmap represented one participant. It cannot be substituted for the whole sample’s inferential analysis. Its proper role is explanatory: to show what the eye tracker captured, not to prove that every methamphetamine-exposed driver adopted the illustrated pattern.

How do stationary and transition gaze entropy differ?

Stationary gaze entropy (SGE) summarizes how widely gaze positions are distributed among areas of interest. With this study’s normalized index, a reduction indicates a more spatially concentrated scan. Gaze transition entropy (GTE) addresses the variability or unpredictability of moves among locations, rather than the overall footprint of looked-at areas. These indices answer related but nonidentical questions. A person may concentrate on a narrow central region and still switch unpredictably among several locations within it.

Methamphetamine alone significantly lowered SGE relative to placebo (p < .001) and increased GTE relative to placebo (p = .028). The GTE comparison against alcohol alone was also reported significant (p = .009). The combined methamphetamine-plus-alcohol arm did not show the same statistically significant GTE contrast in the comparisons reported. That null finding does not establish pharmacological cancellation: treatment interactions, dose, sample size, and task conditions all matter.

A model of pupil, fixation, or gaze behavior is not a direct measurement of the entire visual field, acuity, reaction to unseen pedestrians, or legal fitness to drive. Nor do these indices independently prove the presence of a particular drug. Fatigue, cognitive workload, scene structure, eyesight, and recording hardware can also influence visual exploration; this experiment was designed to compare its randomized treatment arms, not certify a roadside detector’s specificity.

Which outcomes were statistically significant—and which were not?

The strongest reported lane-control separation involved the alcohol-only condition; stimulant-related differences appeared most clearly in selected gaze measures. Several comparisons did not establish a statistically significant change, and it is just as important to report those nulls accurately. In particular, no significant correlation was identified between measured whole-blood methamphetamine or ethanol concentration and SDLP within the active conditions. That result limits attempts to infer the observed individual lane deviation from a single blood test.

A nonsignificant comparison is not affirmative evidence of equality. Conversely, a significant model contrast cannot, without validation, be promoted into an individual forensic classification rule. The study tested numerous time intervals and ocular/driving variables; confidence in any single isolated numerical contrast should be considered in light of the full pattern and the research design.

Did taking methamphetamine with alcohol “cancel out” alcohol?

No cancellation mechanism was demonstrated. Under the assigned combination, participants’ simulator lane-variation patterns more closely resembled the methamphetamine condition than the alcohol-only condition. That was a group-level behavioral observation under one dose, one treatment schedule, and a restricted simulated roadway. It did not establish that methamphetamine reversed all alcohol effects, improved visual attention, or prevented collisions.

The two drugs exert distinct effects across attention, arousal, motor activity, and judgment, and a stimulant may make someone feel more engaged without making every safety-relevant process more reliable. In fact, methamphetamine alone changed gaze entropy even when SDLP was not elevated. A claim of safety based only on lane keeping overlooks the very visual-search phenomenon the investigators were studying.

At around three hours, whole-blood methamphetamine concentrations were comparable between the single-drug and combined arms, as were ethanol concentrations between alcohol and combined arms at the reported samples. The absence of a large group mean pharmacokinetic difference does not establish matched personal exposure at every moment of the drive. Individual responses, timing, and dose variability remain relevant.

Why did subjective confidence diverge from the simulator?

Participants rated their own driving quality after the session on a scale running from poor to excellent. Mean ratings were approximately 69.95 after methamphetamine alone and 67.60 after the combined treatment, compared with 49.95 after placebo and 37.30 after alcohol alone. The study reported significantly higher self-ratings after methamphetamine-containing conditions than placebo and alcohol alone.

Those ratings measure beliefs about performance, not validated driving competence. A confident participant may have maintained lane position while making less dispersed and more erratic visual fixations. The difference is especially important in expert reporting: subjective perceived control cannot be substituted for objective driving outcomes, and objective lane control cannot establish the completeness of hazard detection.

The study does not establish that any individual knowingly overestimated their ability, lied about an incident, or would have made a particular decision in traffic. The questionnaire was part of the experimental measurement framework, not a forensic credibility evaluation.

Can a whole-blood drug concentration predict gaze or lane control?

Not with a validated individual equation from these data. The paper did find an association between whole-blood methamphetamine and average simulator speed in the combined-drug condition (reported r = −.62, p = .01). That specific association does not establish drug concentration as the cause of a person’s speed or show how to predict an unobserved driver’s outcome. The source found no significant concentration–SDLP association for either tested drug in the active conditions.

Measurement precision and interpretive validity must be distinguished. The laboratory may measure methamphetamine reliably in venous whole blood, yet the relation between concentration and behavior can depend on dose history, tolerance, route, elapsed time, task load, sleep, individual characteristics, and concurrent substances. The present randomized laboratory design was valuable for comparing group conditions, but it was not a prospective validation of a forensic decision threshold for individual intoxication.

A reported concentration after a crash or detention also raises sampling-time questions absent from an instantaneous simulator measurement. The study did not validate retrograde methamphetamine modeling for an individual, an ethanol interaction adjustment for unknown consumption circumstances, or an eye-tracking rule to date last use.

How should an expert separate these evidence layers?

First identify the specimen and assay: the actual venous whole-blood concentration for methamphetamine and ethanol, with collection time and method-specific units. Next describe the research exposure: the controlled oral capsule or vodka amount and the elapsed schedule. Then identify the functional measurement: simulator SDLP and steering variation represent vehicle-control proxies, while gaze duration and entropies represent ocular sampling patterns. Finally assess whether the study actually connected those endpoints to the proposed case-specific conclusion.

In this article, the link from treatment to some simulator and eye-tracking differences is experimentally supported at the group level. The much stronger link from a case’s isolated toxicology result to the precise missed hazard, personal impairment, driver culpability, or crash cause is not established. An eyewitness claim about where someone looked is also not equivalent to a 50-Hz eye-tracking record obtained under controlled laboratory conditions.

Relevant background for interpreting these distinctions includes biological specimen matrices, confirmatory analytical drug testing, and specimen collection and handling. These are general published context resources, not other versions of the same Hayley experiment.

What are the study’s most important limitations?

Small and selected sample: Only 20 healthy, previously stimulant-exposed licensed adults were included in the final analysis. The study cannot characterize older motorists, heavy users, people with sleep deprivation or multiple medical conditions, or drivers facing unexpected high-demand situations with the same certainty.

One dosing regime: Both drugs were given acutely under specific oral schedules. Data do not directly address smoked, intranasal, or injected methamphetamine, larger alcohol exposures, chronic use, withdrawal, or complex uncontrolled polydrug combinations. The estimated nominal target of the alcohol drink is not a measured continuous alcohol clamp.

One type of driving: A one-hour monotonous highway simulation can reveal systematic differences, but it does not reproduce all real-road conflicts, crash consequences, pedestrians, intersections, fatigue accumulation, or divided-attention demands. Keeping lane variation low is not sufficient to demonstrate overall safety.

Equipment and measurement: The cap-mounted eye tracker required scene calibration and was suited to controlled laboratory conditions. Source figures visually illustrate gaze properties, but a single-participant heatmap does not generalize. The reported study did not validate a clinical or law-enforcement diagnostic sensitivity, specificity, or actionable individual cutoff.

Multiple tested outcomes: Model-based contrasts across many gaze, steering, speed, and time measurements require careful accounting for selective interpretation. Not all active-condition comparisons were significant; an isolated statistically significant difference should be interpreted alongside null findings, measurement definitions, and the full causal design.

What can the original study legitimately contribute to forensic toxicology?

It supplies controlled evidence that different psychotropic treatments can affect different layers of a driving-related experiment in distinct ways. Whole-blood drug chemistry identifies measured systemic exposure; gaze entropies summarize where and how eyes explored the roadway; and SDLP describes one aspect of simulated vehicle movement. The methamphetamine group showed restricted and less orderly scanning despite relatively stable lane control, while alcohol alone increased measured lane variation. This is a scientifically useful reminder not to treat a single instrument or physiological proxy as a complete measure of safety.

For evaluating a specific driver, the strongest use of this work is defining the limits of what a blood result, eye movement, or simulator result can support. It cannot turn an unspecified measured whole-blood concentration into a proven individual gaze pattern. Nor can it establish a specific driver noticed or missed an event, possessed a given state of mind, caused a collision, or satisfied a jurisdiction’s impairment standard. Those questions require independent, case-specific evidence and validated inference appropriate to the facts.

Frequently asked questions

1. Did methamphetamine improve driving in this study?

No general claim of safer driving is justified. Methamphetamine-containing conditions did not increase SDLP in the study’s straightforward highway simulation and had lower lane variation than alcohol alone. At the same time, methamphetamine alone significantly changed gaze dispersion and gaze-transition organization. These different endpoints prevent equating stable lane position with intact real-world hazard processing.

2. What does lower stationary gaze entropy mean?

It means measured fixations were spatially concentrated within a more restricted collection of tracked areas during this task. It does not mean the instrument measured the participant’s entire visual field or proved that peripheral hazards were invisible. The interpretation depends on the exact gaze-index definition and scene structure.

3. Did the combined alcohol and methamphetamine condition prove that the stimulant counteracted alcohol?

No. The reported group simulator lane variation in that combined condition was not elevated in the same way as alcohol alone. This experiment did not validate any mechanism or safety prescription for mixing the two agents, especially under higher doses or more demanding traffic conditions.

4. Were the reported methamphetamine results from whole blood or plasma?

From collected venous whole blood assayed with LC–MS/MS. The Table 2 values use ng/mL. They must not be renamed plasma concentrations or converted to saliva or breath concentrations by assuming a universal matrix ratio.

5. Did the research find a reliable whole-blood concentration that predicted lane weaving?

No. The authors reported no significant association between whole-blood methamphetamine or ethanol concentration and SDLP in the active conditions. A separate methamphetamine–speed correlation in the combined condition cannot create a validated individual impairment threshold.

6. Can an eye tracker establish that a driver was impaired or caused a crash?

Not from this study alone. The ocular measures compared 20 people in four controlled laboratory sessions. There were no real-road crash outcomes and no independently validated diagnostic sensitivity, specificity, personal culpability test, or universal impairment cutoff for the eye-tracking metrics.

7. Why can an unchanged lane-keeping measure coexist with altered visual scanning?

Lane keeping and visual search measure different functions. A participant may successfully center a simulator vehicle while reducing the breadth or organization of gaze shifts. Detecting, classifying, and reacting appropriately to an unexpected conflict is a more complex question than the SDLP of a relatively simple highway drive.

Primary study and bottom line

Bottom line: The Hayley experiment demonstrates a task-specific dissociation. Alcohol alone increased simulated lane variation; methamphetamine alone narrowed and changed the organization of gaze without producing the same lane-weaving pattern. Real venous whole-blood methamphetamine and ethanol were analytically measured, yet the investigators did not validate a person-specific drug-concentration, eye-tracking, or crash-causation rule. Report the specimen, the behavioral endpoint, the significant and nonsignificant comparisons, and the limits of individual inference separately.

Primary source: Hayley AC, Shiferaw B, Aitken B, Rositano J, Downey LA. Acute methamphetamine and alcohol usage alters gaze behaviour during driving: A randomised, double-blind, placebo-controlled study. Journal of Psychopharmacology. 2024;38(7):636–646. https://doi.org/10.1177/02698811241261024.