Does Amphetamine Increase Motivation Without Proving Better Cognition? What an 18-Mouse Touchscreen Experiment Measured

Answer: In a laboratory experiment using a touchscreen reward task, 1 mg/kg intraperitoneal d-amphetamine increased the average progressive-ratio breakpoint in mice from 49.444 to 80.556 responses (p<0.001). Two dopamine D2-receptor antagonists decreased breakpoint under particular dose-and-timing conditions. These are measurements of operant responding in mice—not measured drug concentrations in blood, improved human cognition, or an individual’s ability to drive. The study also showed that the animal’s willingness to work for milkshake changed when a freely available alternative food could be obtained.

The original investigation was conducted by Christopher J. Heath, Timothy J. Bussey and Lisa M. Saksida and appeared in Psychopharmacology in 2015: Motivational assessment of mice using the touchscreen operant testing system: effects of dopaminergic drugs. A useful forensic reading begins with what the apparatus recorded, then separates the drug administered, the animal’s motor behavior, and the human propositions that were never tested.

What is the main scientific finding?

The investigators adapted two familiar measures of reward-directed behavior to an automated touchscreen. The progressive-ratio task increased the number of responses required to earn each successive portion of strawberry milkshake. The effort-related-choice task let mice work for preferred milkshake while ordinary laboratory chow was freely available. Both produced measurable changes in behavior, supporting the apparatus as an experimental tool.

The most striking pharmacological contrast was bidirectional. Amphetamine increased the effort-related breakpoint, while sulpiride and raclopride, drugs affecting dopamine D2 receptors, decreased it under specified conditions. That comparison supports the task’s sensitivity to dopaminergic manipulation. It does not identify a unique dopamine-mediated cause of every observed mouse response or prove that more responses necessarily signify better learning.

Who and what were actually studied?

The Methods identify 18 male C57BL/6 mice purchased at approximately eight to ten weeks of age. These were animals, not adult or pediatric human volunteers. The same animals underwent successive training and drug-testing sessions, so each daily performance score is a repeated measurement rather than an additional independent subject. The report states that three animals were euthanized after developing mild rectal prolapse during the overall work; the later effort-choice analyses therefore have smaller denominators than the initial drug experiments.

The figure captions are important here: the progressive-ratio pharmacology figures show n=18, whereas effort-choice Figure 6 specifies n=16 for the chow-versus-milkshake workload comparisons and n=15 for the condition without free chow. Treating those 18, 16 and 15 values as one undifferentiated denominator would distort the strength and scope of the evidence. Likewise, the many touches recorded on a screen are not independent animals.

How did the touchscreen measure effort?

The mouse approached a central illuminated region of the screen. Infrared sensors registered a response without the animal necessarily pressing a physical key. Correct target responses earned approximately 20 microliters of strawberry milkshake. The ratio of work to reward was increased deliberately. Under the linear progressive-ratio-four schedule, completing one trial required one response, then the next reward required five responses, then nine, thirteen, and so forth. The session could terminate after five minutes without a response or reward-magazine entry.

The principal endpoint was breakpoint: the response requirement completed in the last successfully finished trial. It is a behavioral measure, dependent on motivation, reward valuation, motor capacity, attention to the cue, and the task’s termination rule. It is not a psychometric intelligence score. A mouse that reaches a higher breakpoint has expended more of the particular instrumental response under these conditions; it has not necessarily formed a more accurate memory or made a more rational decision.

Before pharmacology, mice learned fixed-ratio requirements and rapidly acquired the progressive-ratio schedule. The original Results report breakpoint stabilization across the last five of seven initial progressive-ratio sessions; post-hoc differences among those later sessions were nonsignificant. That training phase matters because the drug contrast was not simply between animals that had never learned how the device worked.

What doses were given, and what was chemically measured?

Three compounds were administered by intraperitoneal injection, not by oral dosing, inhalation or a transdermal patch. The tested amphetamine doses included 0.1 and 1.0 mg/kg, with behavior recorded after a 20-minute injection-to-test interval. Sulpiride was examined at different doses and 20- or 40-minute intervals; the later decisive breakpoint decrease occurred with 50 mg/kg at 40 minutes. Raclopride was examined at 0.03 and 0.3 mg/kg with a 20-minute interval. Drug comparisons were separated by at least eleven days of washout between compounds.

No biological drug concentration was quantified. The source did not perform confirmatory mass-spectrometric analysis of amphetamine, sulpiride or raclopride in mouse whole blood, plasma, serum or brain. There were no human blood specimens, no human therapeutic levels, and no pharmacokinetic time curve from measured samples. The numeric mg/kg values describe doses administered to mice. Converting them into human blood ng/mL, a legal impairment threshold or a time-of-use estimate would require unperformed measurements and a separate validated model.

Evidence table 1: the three main drug comparisons

The vehicle results are not interchangeable baselines. They come from different experimental sequences, compounds and test intervals. This table is a structured comparison of three reported contrasts, not a dose-response equivalence analysis. Nor should the point estimates be interpreted as rates of neural dopamine release: the values are the mean number of required touchscreen responses at the last earned reward.

What did 1 mg/kg amphetamine change?

In the n=18 progressive-ratio series, drug condition significantly affected breakpoint, F(2,34)=27.40, p<0.001. The increase from 49.444 to 80.556 responses at 1.0 mg/kg is approximately 31.1 additional responses, or roughly 63 percent relative to that vehicle mean. The lower 0.1 mg/kg dose did not show the same reported significant breakpoint effect. That is an important distinction: the experiment did not establish a uniformly positive stimulant effect across every administered dose.

The investigators also evaluated reward-collection latency, post-reinforcement pause, infrared beam breaks, off-target touches, total response time and inter-response interval. Amphetamine’s reward-collection latency test was nonsignificant (p=.127), as was the post-reinforcement pause (p=.299). Rear-zone infrared movement was not significantly changed by amphetamine (p=.410), while the front-zone beam-break rate rose at the higher dose. The latter finding fits greater activity around the active screen, not necessarily generalized locomotor hyperactivity everywhere in the chamber.

Non-target touches increased at 1.0 mg/kg, from approximately 0.027 to 0.041 per second (post-hoc p=.030). This is not simply an error count in a human attention test. The four unused touchscreen apertures were physically close to the central target, and the authors discuss the possibility of near-miss touches while animals are working vigorously. The study cannot isolate a single psychological mechanism from these screen contacts alone.

Did sulpiride reduce responding?

Yes, but timing was decisive. When sulpiride was administered 20 minutes before testing, the researchers did not identify a breakpoint reduction even across relatively high investigated doses. At a 40-minute delay, the contrast at 50 mg/kg showed a mean breakpoint decline from 37.670 to 31.000, with p=.009. The lower tested 25 mg/kg dose did not show a separately reported significant breakpoint decrease. Calling sulpiride categorically effective without specifying the tested interval would omit the study’s most consequential negative results.

At 40 minutes, reward-collection latency (p=.860) and post-reinforcement pause (p=.909) were both nonsignificant. Rear-zone beam-break rate was also nonsignificant (p=.423). The front-zone beam-break measure declined with 50 mg/kg (p=.008), consistent with less time or activity at the response panel. This distribution of effects may support reduced task-associated activity, but it does not rule out all changes in motor behavior, reward processing or other pharmacological pathways.

The source also reports a magazine-entry main effect p=.048 for sulpiride, but no significant individual post-hoc treatment comparison. A reader should not silently promote that omnibus result into a definitive conclusion that either dose increased or decreased magazine behavior. These within-task null results are necessary for an accurate pharmacological interpretation.

What did raclopride show that sulpiride did not?

Raclopride 0.3 mg/kg, tested after 20 minutes, significantly lowered breakpoint from 40.778 to 29.222 (p=.001); the overall treatment effect was F(2,34)=11.67, p<0.001. But its movement profile was broader. The higher dose decreased infrared beam-break rates in the front zone (p=.010), rear zone (p=.001), and reward-magazine area (p=.001). Reduced non-target touches were also reported (p=.005).

This combination prevents an overconfident statement that raclopride selectively reduced motivation while leaving general motor output intact. The authors themselves identify nonspecific locomotor suppression as a plausible contributor. Reward-collection latency showed a significant overall effect (p=.023), yet the pairwise tests reported did not establish a significant difference between either individual dose and vehicle. Total response time and inter-response interval produced trend-level comparisons around p=.078, not conventionally significant main effects.

Thus both D2-antagonist experiments support pharmacological sensitivity of the assay, but the two drugs did not create identical evidence about the distinction between motivation and motor function. An expert analysis should not flatten these different motor findings into one asserted receptor-specific behavioral mechanism.

What is the effort-related-choice experiment?

The second paradigm tested whether an animal would perform repeated screen touches for a highly preferred reward when a less preferred food was available without labor. The mouse could earn strawberry milkshake through instrumental responding or consume freely scattered standard chow. That design tests a cost-benefit tradeoff: what happens to preferred-reward pursuit as the effort needed to obtain it rises?

The researchers raised the fixed response requirement from 16 to 32 and then 40 responses per reward. They measured actual chow mass consumed in grams and milkshake volume earned in microliters, not a dopamine concentration. As the required screen touches increased, animals obtained less milkshake and consumed more freely available chow. Removing chow at the demanding ratio caused more screen work again, helping distinguish a feasible behavioral alternative from loss of ability to perform the learned task.

Evidence table 2: effort requirements changed observed food choices

The original repeated-measures analysis reported an effect of work requirement on chow consumption, F(2,30)=57.138, p<0.001, and on earned milkshake, F(1.113,16.702)=133.187, p<0.001. Every reported adjacent workload comparison was significant in the original post-hoc analysis. In a distinct test at fixed ratio 40, removal of free chow increased completed work, t(14)=-9.330, p<0.001. These are behavioral choices under a reward design, not estimates of human capacity to perform a job or drive.

Why does reward preference matter?

Without showing that milkshake was preferred, a declining amount of work for milkshake could be misread. The authors therefore conducted a choice assessment in which mice could consume chow and either water or milkshake without an operant requirement. Animals consumed significantly less chow when milkshake was available than when water was the alternate drink (p<0.001), and greater milkshake than chow in a direct preference contrast (p=.019). These findings establish the relevant direction of preference under the experiment’s food-restriction and testing conditions.

Preference is still context-dependent. These mice were food-restricted to approximately 85–90 percent of baseline free-feeding weight, and milkshake was used as the reinforcer. A shift toward freely available chow is therefore not a universal measurement of reward preference, loss of pleasure or human psychological motivation. It is the specific behavioral pattern elicited by a selected incentive and effort schedule.

Does a higher breakpoint mean improved cognition?

No direct cognitive enhancement was established by the progressive-ratio outcome. The mice were trained to execute a familiar response at one illuminated location. The test primarily measured continued responding as the price of another reward rose. It did not administer a standardized human attention battery, measure episodic memory, determine reasoning accuracy, or test ability to respond safely to traffic hazards. The result could be one input into experimental studies that later include a cognitive test; it is not itself proof of better cognition.

This distinction matters whenever a drug is said to improve performance. One type of performance can improve because the subject works harder, approaches the target more frequently, or perseveres longer. Another type might require fewer mistakes, appropriate stopping, flexible choice, or speed-accuracy balance. The study helps separate the ability or willingness to perform a task from the cognitive property an investigator actually intends to measure.

Were locomotion and incentive effects completely separated?

Not completely. The apparatus offered additional infrared beams and response-timing measures so the researchers could examine where the animal moved and how it performed its responses. Under amphetamine, the rear zone and reward timing remained largely unchanged while front-zone activity increased. Under sulpiride, several motor indices were unaffected but screen-proximal activity declined. Under raclopride, significant reductions extended across front and rear movements and reward-magazine entries.

Those differences strengthen interpretation compared with looking at breakpoint alone, but they do not create a validated causal decomposition of the response into exact percentages attributable to motivation, alertness, sedation, reward preference, or muscular activity. The authors did not chemically sample brain dopamine during each touch or independently manipulate every potential mechanism. The most defensible inference is that the touchscreen assay detects drug-sensitive reward-directed behavior while remaining susceptible to motor and task-structure influences.

What can forensic toxicologists legitimately take from an animal task?

The study provides a concrete demonstration that drug action can change one component of goal-directed behavior without creating a uniform improvement across all performance variables. It also illustrates the value of looking for internally discordant outcomes. If an expert cites only the strong amphetamine breakpoint increase, but ignores the nonsignificant reward latency or the separate effects of an antagonist on movement, the summary omits constraints important to interpretation.

Equally important, a nominal injected mg/kg dose is not a forensic concentration. Interpreting a human toxicology result requires actual specimen identity, sampling time, validated analyte method, identification criteria, uncertainty, and the clinical or legal question. The site’s separate guide to biological matrices explains why tissue and blood compartments cannot be swapped. The discussion of confirmatory drug testing distinguishes evidence for chemical identity from behavior, and therapeutic drug monitoring addresses the additional need for an interpretable timed concentration.

What limitations control extrapolation to a human case?

First, only male inbred laboratory mice were studied. Second, the animals experienced extensive operant training and food restriction, which differ fundamentally from untrained human behavior under ordinary conditions. Third, the drugs were delivered by intraperitoneal injection, with tested intervals of 20 or 40 minutes, not through the common clinical routes of administration. Fourth, the experimental groups were modest in size and partly overlapping across stages, with attrition before the choice-task work.

Fifth, the study reported group averages and within-animal comparisons, not a validated rule for predicting the state of an individual animal, much less a person. Sixth, no analytically measured drug concentration permits pharmacokinetic interpretation. Finally, no human driving, aviation, occupational, eyewitness or incident-causation outcome was collected. These are unmeasured outcomes, not statistically nonsignificant trials of those propositions.

The original article’s practical value is methodological: it teaches readers not to mistake vigorous responding for all-purpose competence, not to conflate reward valuation with cognitive accuracy, and not to treat movement controls as optional. It does not supply a lawful inference of individual impairment or nonimpairment from the presence of amphetamine in any person’s blood.

How should an expert report describe the study?

A reliable synopsis identifies the actual source, the n=18 progressive-ratio cohort, the later n=16/15 choice denominators, the animal intraperitoneal exposures, and the exact recorded behavior. It reports the significant breakpoint contrasts alongside the negative and mixed motor-control findings. It also states that drugs were not quantified in any biological sample and that the research did not test a person’s toxicology or driving. This preserves the experiment as supportive background science without promoting its endpoints into clinical or forensic facts.

For further examination, the primary source is Heath, Bussey and Saksida, Psychopharmacology (2015), DOI 10.1007/s00213-015-4009-8. Read its Figures 3–7 with the accompanying statistical tests, especially the differing motor patterns under amphetamine, sulpiride and raclopride, and the decline in earned milkshake as work requirements rise.

Frequently asked questions

1. Did amphetamine increase motivation in this experiment?

It increased an operational measure of reward-directed effort, the progressive-ratio breakpoint, at 1.0 mg/kg in mice. This supports altered task performance and is consistent with greater willingness to work for reward under the protocol. It does not isolate motivation from every other psychological or motor process.

2. How many independent animals contributed?

Eighteen mice entered the initial touchscreen pharmacology series. The subsequent workload comparisons included 16 mice, and the separate no-free-chow comparison included 15. Repeated touches and multiple testing sessions are observations on the same animals, not additional subjects.

3. Is a progressive-ratio breakpoint a drug blood concentration?

No. Breakpoint is the last completed response requirement in a reward task. The source contains administered doses in mg/kg but no chemical measurements of amphetamine or either antagonist in whole blood, plasma, serum or any other drug-assay matrix.

4. Why did raclopride require additional caution?

Its higher dose decreased movement measures in more than one part of the chamber as well as reducing breakpoint. Nonspecific motor suppression could therefore contribute to less responding. A selective conclusion about motivation alone would exceed the observations.

5. Did sulpiride work equally well at 20 and 40 minutes?

No. The authors found no breakpoint effect at their initial 20-minute administration-to-test interval. Their significant 50 mg/kg breakpoint decrease was reported after a 40-minute interval. The timing difference prevents a blanket conclusion across administration schedules.

6. Did the experiment establish that amphetamine makes people drive better or worse?

No humans were studied, and no driving simulator or on-road test was administered. A trained mouse’s willingness to touch a screen for milkshake is not a validated surrogate for a driver’s divided attention, hazard response, or person-specific impairment.

7. What is the strongest use of this study in a forensic discussion?

It illustrates why a measured endpoint must be defined precisely and why movement, reward preference, drug dose and task design should be assessed separately. It cannot replace an individual’s laboratory records, clinical evidence, or a validated human performance study.