Can Dextroamphetamine Improve Target Detection in Both People and Mice? A 71-Adult, 32-Mouse Study

Short answer: A controlled 2018 experiment found that single oral doses of 10 or 20 mg dextroamphetamine improved several scores on a five-choice continuous performance task in 71 healthy adults. In a separate experiment involving 32 male mice, an injected dose of 0.3 mg/kg improved related measures of detecting the correct target. The most consistent cross-species finding was better identification of target trials, not a universal improvement in inhibition, reaction speed, or all measures of vigilance. Adults showed less variability in hit reaction time; mice did not show that effect. Neither experiment measured a drug concentration in forensic whole blood or established a threshold for individual driving impairment.

The original primary study is MacQueen, Minassian, Kenton, Geyer, Perry, Brigman, and Young, Amphetamine improves mouse and human attention in the 5-Choice Continuous Performance Task, Neuropharmacology 138 (2018):87–96, DOI: 10.1016/j.neuropharm.2018.05.034. This account distinguishes the two randomized experiments, their separate units of analysis, what the task measured, what the drug did not measurably change, and why laboratory attention performance cannot be converted into an individual forensic drug-effect opinion.

What is the narrow scientific answer to the study?

The paper provides evidence for an overlap between the effects of dextroamphetamine on a specified attention task in two species. In the human study, both active-dose groups had better target discrimination, correct target responding, response accuracy, omission rates, and consistency of response timing than placebo. In mice, the middle of three injected doses yielded better target discrimination, hit rate, and response accuracy than saline. These agreements matter because the investigators were trying to validate a behavioral task that could be used across species, not calibrate a drug test for court.

The agreement is incomplete. The human experiment had a significant reduction in reaction-time variability and omission rate, whereas the mouse experiment had no significant drug effect on either endpoint. There was also a high-dose mouse accuracy decrement. A finding that some scores changed in similar directions therefore does not justify saying that all human cognitive effects were reproduced in mice. It supports a narrower interpretation: certain target-detection measures may show comparable pharmacological responsiveness in the two defined task implementations.

How many independent humans were tested?

Seventy-one adults aged 18–35 completed the human five-choice task and contributed data. The randomly assigned, double-blind, single-session groups comprised 29 placebo, 17 dextroamphetamine 10 mg, and 25 dextroamphetamine 20 mg. The study reported 47.9% male participants overall. Doses were supplied orally by an investigational pharmacy; testing began approximately 3.5 hours after ingestion. Unlike a crossover trial, each person participated in only one experimental drug condition for this task. Repeated stimuli and trial blocks inside an individual’s test are not additional independent human participants.

The groups were not perfectly balanced because the placebo and 20 mg conditions were oversampled later in collection. The investigators compared baseline demographics and reported no statistically significant group differences in age, education, sex, race, or ethnicity. Nonsignificance on those comparisons is not proof of equivalence on every unmeasured factor. Medical and psychiatric screening also means these findings should not automatically be transported to patients with ADHD, people using multiple drugs, acutely injured patients, or drivers under field conditions.

How many independent mice were studied?

The mouse experiment involved 32 male C57BL/6J mice, with eight mice per assigned condition: injected saline or dextroamphetamine at 0.1, 0.3, or 1.0 mg/kg. The manuscript reports that all 32 completed pretraining and both successive attention-task training phases. The injection was intraperitoneal, and the mice began the experimental five-choice task ten minutes afterward. This is not equivalent to the oral human dosing protocol. A mass-normalized animal dose cannot be translated into a measured human whole-blood amphetamine concentration.

Each mouse carried out multiple behavioral trials, but those trials remained observations nested within one animal. The mouse analysis treated dose as a between-subject condition, not four independent medication periods for every mouse. The authors examined training comparability before drug challenges, which helps address pre-existing differences in learning the task, but it does not eliminate every species, route, timing, motor, or reinforcement difference.

What is the five-choice continuous performance test?

A continuous performance test presents frequent stimuli requiring a response and less frequent stimuli requiring the subject to refrain from responding. In this five-choice version, the meaning of a target came from its position rather than a letter or number. For the adult participants, one circle above one of five spatial positions meant move a joystick toward that target. Simultaneous circles above all five positions meant withhold a movement. Human participants performed 225 target trials and 45 nontarget trials after a brief practice set; a target appeared for 100 milliseconds, with additional time allowed for the response.

The touchscreen version for mice replaced the joystick with nose or screen contact and provided food pellets as reinforcement. Following training, mice completed a shorter 60-trial procedure with target and nontarget presentations. Mouse stimuli, reinforcement schedules, motor demands, intertrial intervals, and testing duration were not identical to those in the human task. Calling the tasks comparable means that they generate related response categories and sensitivity measures, not that species received the same psychological experience or physical amount of drug.

What do a hit, omission, false alarm, and wrong-location response mean?

A hit is a correct response to a presented target. An omission is failure to respond to a target. A wrong-location response is an attempted response to a target but at the wrong one of five positions. A false alarm is an inappropriate response when the five-position nontarget signal requires withholding action. These distinctions prevent over-reading a single aggregate score. An individual might respond to more true targets without becoming better at resisting false alarms; another might increase response frequency without improving spatial accuracy.

The investigators used signal-detection sensitivity, commonly written d′, to combine hit and false-alarm information, and a separate responsivity index to address response tendency. Neither index is an actual drug concentration or a standardized legal-impairment probability. The published formula for each indicator matters: the paper’s hit-rate definition uses hits and omissions; it does not transform every target trial, wrong-location response, and nontarget error into a single interchangeable percentage.

What improved in the 71 adults?

The overall drug-group effect on signal detection was F(2,68) = 5.427, p = .007, with the paper reporting improvement at both oral active doses relative to placebo. The dose-group comparison for correct target hit rate was F(2,136) = 7.628, p = .001, accompanied by fewer omissions, F(2,68) = 7.350, p = .001. Response accuracy also improved, F(2,136) = 12.037, p < .001. The investigators reported standardized effect-size estimates for these comparisons, but these are group differences in a computerized task, not probabilities that a given person will improve.

A second significant endpoint was hit reaction-time variability: F(2,68) = 12.964, p < .001. A reduced spread of response times means that correct responses were more consistent, which is not the same as a statistically demonstrated reduction in mean response time. The pattern suggests that attention performance is multidimensional. A court or clinical reviewer who says simply that the subjects became faster would miss the fact that mean hit reaction time was not a statistically significant endpoint.

Which human outcomes did not significantly improve?

The researchers reported no significant overall effect on false-alarm rate, F(2,68) = 0.747, p = .478, and no significant overall effect on mean hit reaction time, F(2,68) = 2.112, p = .129. The overall responsivity-index comparison was likewise not significant at the conventional .05 criterion, F(2,68) = 3.007, p = .056. The manuscript does describe a responsivity dose-by-trial-block interaction, F(4,136) = 7.628, p = .001, but that time-block effect should not be silently relabeled as a significant overall dose effect.

The reported premature-response test also failed the conventional threshold, with a source-reported p = .051. Describing p = .051 as a proven decrease in motor impulsivity would reverse the test’s actual classification. Likewise, a nonsignificant change in false alarms is not proof that dextroamphetamine never affects inhibitory control. It means the experiment did not establish that effect with the endpoint, design, and data analyzed here.

Human primary results: measured effects and null findings

These are source-reported group-model comparisons, not a new meta-analysis or pooled analysis across species. The denominators and model degrees of freedom should not be mistaken for additional biological participants. Statistical significance does not itself establish a clinically important effect, a pharmacokinetic explanation, or a particular person’s impairment status.

What changed in the mouse experiment?

With eight mice in each condition, the clearest positive pattern occurred at the intermediate 0.3 mg/kg intraperitoneal dose. Signal detection improved, F(3,28) = 3.438, p = .016, and hit rate increased, F(3,28) = 4.015, p = .017. Response accuracy showed an overall dose effect, F(3,28) = 6.012, p = .002. Crucially, the response was not uniformly favorable at every dose: accuracy increased at 0.3 mg/kg but decreased at 1.0 mg/kg according to the original comparison and source Figure 3.

The observed omissions were lower at the intermediate dose, approximately 18.8% compared with 28.75% under saline, but the omnibus omission analysis was p = .055. It is inaccurate to convert that descriptive difference into a statistically significant omission reduction. Mouse behavior can reflect willingness to approach, motor coordination, task learning, reinforcement, and attentional selection. The observed effects are properly described as performance changes under the conditions tested, not as a measured rise in dopamine or quantified amphetamine exposure in brain tissue.

What failed to change significantly in the mice?

At the group-model level, the mouse data showed no significant drug effect on false-alarm rate, F(3,28) = 0.844, p = .481; mean hit reaction time, F(3,28) = 1.577, p = .218; or hit-reaction-time variability, F(3,28) = 0.252, p = .859. The premature-response comparison was also nonsignificant, F(3,28) = 1.911, p = .151, as was responsivity, F(3,28) = 1.355, p = .278.

These negative tests are important because they limit a superficial claim of cross-species replication. The adults had lower reaction-time variability after active doses; the mice did not. The authors suggested that touching a spatial target after moving through a chamber may recruit additional motor demands compared with directing a joystick. That is a plausible procedural explanation offered in the paper, not a directly tested proof of the neurological basis of the species difference.

Mouse primary results: effective dose and non-effects

The table reports the original 32-mouse, four-group experiment. Eight mice per condition is a small denominator, and there are several endpoint comparisons. Neither a large-looking plotted difference nor a nonsignificant p-value can support an invented individual-diagnosis or criminal-law cutoff.

Are these results evidence of better sustained vigilance?

They are evidence for better performance on some indices of a brief attention task. That is narrower than demonstrating prevention of deteriorating vigilance across a long shift. The human 5C-CPT lasted about eight minutes and did not show an overall progressive decline in performance across the three task blocks. In the placebo group, omissions increased numerically from approximately 2.21% early to 3.01% late, but that comparison was not statistically significant. Therefore, this particular experiment could not demonstrate that dextroamphetamine protected participants against a deterioration that was not firmly established in its placebo group.

The authors contrasted these findings with their separate Conners CPT research, in which longer test duration and a different trial structure exposed some time-dependent changes. The current five-choice study is not a renamed version of that paper. It studied another endpoint architecture, partly overlapping participants, and a distinct animal comparison. That partial overlap also means the two human publications should not be counted as two wholly independent confirmation samples in a literature synthesis.

Does the mouse model establish a human mechanism?

No direct molecular mechanism was measured here. The concordance of hit-rate and sensitivity effects in humans and mice is useful for translational behavioral research. It suggests that researchers can use the mouse touchscreen model to test future pharmacological hypotheses relevant to human target detection. But the study did not directly quantify synaptic dopamine, norepinephrine, dopamine transporter occupancy, cortical signaling, or tissue concentrations after injection. Explanations about catecholamine transporter inhibition come from broader pharmacology and the authors’ discussion, not new directly observed chemical measurements within this experiment.

Nor can the relative animal and human doses be made equivalent by comparing their numbers. Adults received fixed oral 10 or 20 mg doses about 3.5 hours before testing. Mice received 0.1, 0.3, or 1.0 mg/kg intraperitoneally ten minutes before testing. Route, absorption, metabolism, disposition, task length, reinforcement, and apparatus differ. A shared direction of a behavioral endpoint is not proof of matched internal exposure or receptor activation in the two species.

What physical drug specimens did investigators collect?

The article does not report a postdose quantitative assay of dextroamphetamine in human venous whole blood, physically separated plasma or serum, urine, hair, oral fluid, breath, or any mouse blood or brain sample. The tested doses are known because investigators administered them, not because an analytical laboratory measured the resulting concentration in a biological matrix. A administered milligram or milligram-per-kilogram dose is not an ng/mL forensic specimen result.

Behavioral timing is equally different from chemical measurement. Approximately 3.5 hours after oral dosing is the scheduled laboratory task time, not a measured peak blood level. Ten minutes after mouse injection is an experiment interval, not evidence of what a person’s blood concentration would be ten minutes after a disputed event. No validated blood-to-brain model, toxicological concentration-response curve, or legal cutoff was constructed in this study.

What does the study say about impairment or fitness to drive?

Nothing direct. Participants moved a joystick toward brief visual signals; mice used a touchscreen for food rewards. Neither protocol involved on-road driving, a driving simulator, crash outcomes, lane maintenance, divided road-scene attention, field sobriety testing, or police observations. A laboratory improvement in target detection cannot be equated with safely managing complex competing tasks, and a reduced omission rate does not rule out adverse effects in another functional domain.

For legal or occupational interpretation, the first question is what was actually observed in the person or specimen at issue. A measured legal whole-blood drug concentration, if one exists in a particular matter, needs its own analytical validation, specimen chronology, and uncertainty evaluation. This 2018 experiment supplies neither a per-se blood cutoff nor an individualized relationship between a numeric blood result and driving behavior. It would be just as unsound to infer a particular subject was impaired merely because dextroamphetamine was administered as to infer the subject was unimpaired because average target detection improved.

How should an expert use this source without overclaiming?

The scientific use is to identify a defined behavioral effect and its boundaries. Specify the studied population, the experimental chemical, the randomized group size, the tested oral or injected dose, the task, the endpoint, the statistical contrast, and the original nonsignificant results. State explicitly when the endpoint was not measured. Distinguish a laboratory task’s signal-detection parameter from a forensic analytical method’s identification performance. These are very different meanings of “detection.”

When evaluating a case, a reviewer should separately examine the specimen type and analytical method, if any; whether the drug was actually identified and quantified; the time between specimen collection and the event; observed behavior; the possibility of other drugs and sleep loss; and competing causal explanations. The primary paper does not resolve those matters. Its measured task endpoints are informative research observations, not direct case findings.

Useful background on matrix distinctions appears in biological matrices in forensic toxicology. The difference between behavioral identification of targets and chemical identification of drugs is discussed in confirmatory drug testing. For concentration-versus-dose reasoning, consult therapeutic drug monitoring. These are general-context resources, not substitute citations for this original experiment.

Seven frequently asked questions

1. Did dextroamphetamine significantly improve human attention?

It improved several defined five-choice task measures in the randomized groups: signal-detection sensitivity, correct target hit rate, response accuracy, omission percentage, and response-time consistency. Those were group-level results in screened healthy adults after oral 10 or 20 mg dosing. The study did not assess every dimension of attention or real-world functioning. The phrase “improved attention” is defensible only when connected to those particular outcomes and their nonsignificant counterparts.

2. Did it reduce false alarms or speed mean reaction time?

No statistically significant overall drug effect was detected on human false-alarm rate (p = .478) or average hit reaction time (p = .129). Adult hit reaction-time variability was lower, which concerns consistency rather than a significant average-speed improvement. Those are distinct parameters and should never be conflated in expert testimony or public-facing scientific summaries.

3. Were the same doses tested in humans and mice?

No. Adults were given fixed oral amounts of 10 or 20 mg, with the task beginning approximately 3.5 hours later. Mice received intraperitoneal 0.1, 0.3, or 1.0 mg/kg and were tested about ten minutes later. Those regimens differ in route, units, species, and timing. Neither protocol included a quantitative blood-drug assay to establish comparable internal exposures.

4. What happened at the highest dose in mice?

In the mouse experiment, the intermediate 0.3 mg/kg dose improved correct target identification and response accuracy, while the 1.0 mg/kg dose reduced accuracy relative to saline. The original results therefore do not describe a monotonic increase in attentional benefit as dose rises. They also do not establish why the higher dose produced a less favorable task score; the needed drug concentration and mechanistic tests were absent.

5. Can d-prime be interpreted as a drug-test sensitivity or legal threshold?

No. Here d′ is a signal-detection statistic for distinguishing target from nontarget visual stimuli during a behavioral test. It is not the sensitivity of an LC-MS/MS method for amphetamine, a concentration in blood, a toxicology cutoff, or an individual probability of impairment. Chemical confirmation requires a separately validated assay on a specified biological specimen.

6. Does this article prove that amphetamine prevents vigilance lapses over a long shift?

No. The short human task did not show a clear placebo-group vigilance decrement across its three blocks, and the mouse 60-trial procedure did not permit a comparable block-based analysis. A gain on selected target-detection metrics should not be recast as measured prevention of fatigue-related accidents, sleep-loss errors, or extended-shift attention failures. Those require other experiments.

7. Can an expert use the results to declare an individual driver impaired or safe?

No. No road driving, case-specific incident, physical forensic whole-blood quantitation, or individual concentration-impairment cutoff was studied. The results can inform a carefully qualified discussion of experimental attention performance, but the circumstances of any specific person demand their own complete toxicological, behavioral, chronological, and alternative-cause evidence. No single group mean substitutes for that analysis.

Bottom line: cross-species task performance is not a forensic blood-drug inference

This primary study supports a narrow, replicable experimental statement: oral dextroamphetamine improved several target-detection metrics in 71 healthy adults, and an intermediate injected dose improved related metrics in 32 mice. Significant findings coexisted with important null results, different dose-response behavior, and meaningful task-design differences. The data can motivate research on attentional processes across species. They cannot, by themselves, determine a person’s blood concentration, medication exposure at an incident, individual impairment, safety to drive, or causation of a disputed event.

Original primary source: MacQueen DA, Minassian A, Kenton JA, Geyer MA, Perry W, Brigman JL, Young JW. Amphetamine improves mouse and human attention in the 5-Choice Continuous Performance Task. Neuropharmacology. 2018;138:87–96. https://doi.org/10.1016/j.neuropharm.2018.05.034.