Can D-Amphetamine Improve Zebrafish Maze Alternation While Reducing Flexibility? A 36-Fish Study

Short answer: In a controlled experiment involving 36 adult zebrafish initially allocated to three groups, repeated exposure to d-amphetamine increased the proportion of alternating turns in a free-movement Y-maze, a behavioral pattern previously associated with working-memory performance in this model. But the same amphetamine-exposed fish also showed less varied search behavior and, after a further drug challenge, lower locomotor activity than controls. Nicotine produced a different profile: greater alternation after challenge without the same reported loss of flexibility. These are experimentally observed fish behavior differences, not measurements of drug concentrations in human blood or evidence that an individual was cognitively unimpaired.

The primary source is Madeleine Cleal, Barbara D. Fontana, and Matthew O. Parker, “The cognitive and behavioral effects of D-amphetamine and nicotine sensitization in adult zebrafish,” Psychopharmacology 238, 2191–2200 (2021), doi:10.1007/s00213-021-05844-5. The distinction between increased repetition of a useful maze strategy and flexible adaptation is the central forensic-scientific lesson.

What question did the researchers actually test?

The investigators asked whether repeated, intermittent exposure to two stimulants changed cognitive strategy and locomotion in adult zebrafish. An animal can turn alternately left and right more often without necessarily becoming more adaptable. If it keeps repeating the same route when circumstances change, more alternation may be accompanied by reduced behavioral flexibility. The authors therefore monitored not just one summary measure of apparent working memory but also repeated-turn sequences, overall movement, and the way strategy use changed over the course of a testing hour.

This was a primary experimental investigation, not a retrospective review of human arrests or toxicology results. The authors used a free-movement pattern Y-maze that records spontaneous arm entries. They did not interrogate witnesses, test drivers, or analyze legal specimens. Its value is mechanistic and methodological: it shows how choosing a single behavioral endpoint can obscure a competing effect in another domain.

How many animals were studied, and what counts as an independent observation?

The initial study population was 36 approximately three-month-old AB wild-type zebrafish, distributed equally between sexes. Researchers randomly allocated 12 fish to a d-amphetamine group, 12 to a nicotine group, and 12 to a water-control group. The design repeatedly followed the same fish through baseline testing, chronic exposure, withdrawal, and challenge. Four observations from one fish are four visits from one independent animal, not four independent fish.

The figure describing the principal behavioral outcomes reports analyzed group sizes of 11 controls, 11 amphetamine fish, and 12 nicotine fish. The Methods state that extreme studentized-residual observations were removed from subsequent modeling, but the figure’s smaller counts should not simply be reclassified as confirmed deaths or clinical attrition. The statistical models also include repeated time bins and arm turns. A denominator such as 683 in a model F statistic is not a population of 683 independently randomized animals. Both the initial allocation and the narrower figure-specific denominators matter to an accurate reading.

What concentrations and routes of exposure were used?

Fish were exposed to 25 micromolar d-amphetamine or 5 micromolar nicotine in treatment water. Control animals received water without the corresponding stimulant. These numbers describe nominal aquarium-water exposure concentrations. They are not oral doses in milligrams, injected doses in milligrams per kilogram, or analytically confirmed concentrations inside the fish’s blood, brain, or other tissues.

The investigators used 30-minute exposures on each of 14 consecutive days, followed by two days without drug. The same treatment was then administered again as a 30-minute challenge. The distinction between water concentration and absorbed internal dose is especially important: laboratory exposure conditions do not identify the mass of drug absorbed by an individual fish, let alone establish any human-equivalent toxicological concentration.

What exactly did the free-movement Y-maze measure?

Each fish explored a three-armed apparatus for one hour. Automated recording produced a time-ordered sequence of left and right choices. Investigators examined four-choice patterns, sometimes called tetragrams. Alternations are LRLR and RLRL; repetitions are LLLL and RRRR. Changes in the relative frequency of these patterns provide evidence of how consistently an animal searches.

The percentage of alternations was interpreted as a working-memory-related measure because earlier pharmacological experiments with this maze found reductions after memory-disrupting treatments. Nevertheless, that interpretation is not equivalent to demonstrating that an animal remembered a human-like event or that its entire cognitive repertoire improved. Repetition counts help identify competing stereotyped behavior, and the number of total turns serves as a separate measure of locomotion. Six consecutive ten-minute portions of each exploration hour were also analyzed to evaluate how strategy changed over time.

What was the experimental sequence?

First came a one-hour baseline Y-maze session before treatment. Second, after 14 daily half-hour exposure sessions, fish completed the maze immediately after the final exposure. Third, after 48 hours in home tanks without treatment, they completed another maze session. Fourth, following a renewed half-hour exposure to the originally assigned treatment, they completed a final maze assessment. This structure provides both within-animal comparison and contrasts among treatment groups.

The experimenters were blinded to treatment assignment during the behavioral assessments. The investigators analyzed treatment stage, treatment and their interactions with linear mixed models, followed by multiple-comparison procedures described as Bonferroni corrected. Their conclusions depend on modeled changes in behaviors across the stages, not on serial drug concentration-time curves. The paper’s conclusions should therefore be framed as changes in recorded task behavior after assigned environmental exposure.

What happened to maze alternations after d-amphetamine?

In amphetamine-treated fish, the proportion of alternations increased after the 14-day treatment course compared with the original baseline. The reported baseline comparison had a 95% confidence interval for the change of 12.24 to 30.60 percentage points, with p < 0.001. The authors characterized this as an increase of about eleven percentage points in alternation use during the chronic-exposure stage.

After the two-day drug-free interval, the difference from baseline no longer met the stated significance criterion: the reported confidence interval extended from −0.141 to 16.75 and p = 0.06. This is an important negative result. It is not evidence that the effect was definitely absent after withdrawal, but neither is it a significant baseline difference under the reported analysis.

Following a fresh d-amphetamine challenge, alternation use again increased over baseline, with a reported 95% confidence interval of 18.87 to 37.48 and p < 0.001. The authors described an increase of roughly nineteen percentage points from baseline at challenge. These are relative behavioral proportions obtained in the maze, not percentages of individuals experiencing a clinical improvement.

Did nicotine produce the same behavioral effect?

Nicotine-treated fish also showed higher alternation use at the challenge stage. The investigators reported a rise from approximately 25% at baseline to approximately 36% after challenge, with a 95% confidence interval for the change of 1.889 to 20.376 percentage points and p = 0.01. That is evidence of a challenge-associated increase within this experimental group.

The nicotine group did not show a significant baseline comparison at the preceding chronic-exposure stage (p = 0.09) or during withdrawal (p = 0.21). Thus a headline saying that nicotine significantly enhanced every stage of working memory would misstate the actual sequence of findings. The authors also reported no significant treatment-stage effect on repetitive tetragrams within the nicotine group (p = 0.29), in contrast to the amphetamine-associated change after challenge.

Evidence table 1: Which key changes were significant, and which were not?

The confidence intervals above concern a modeled change in behavioral proportion. They are not confidence intervals around measured drug concentrations. The nonsignificant comparisons are included deliberately because the time course does not support an unqualified claim of continuous or universal cognitive enhancement.

Why did increased alternation not mean better cognitive flexibility?

A familiar task can reward persistent use of a particular strategy. A fish that frequently alternates directions may look successful on an alternation-based measure while becoming less able to switch to other exploration patterns. The researchers specifically examined whether alternation frequency varied through six successive ten-minute blocks within the hour. In amphetamine-treated fish, there was a significant treatment-stage effect on the use of alternations, but no statistically significant effect of time within the testing hour (p = 0.22) or treatment-stage-by-time interaction (p = 0.18) under the reported model.

The authors interpreted this pattern alongside the high challenge-stage alternation levels as reduced flexibility or a more persistent stereotyped search strategy. That is a behaviorally specific inference rather than a direct measurement of executive function in a person. The same study reported a decrease in repeated-turn tetragrams in the amphetamine group after challenge (reported confidence interval −10.64 to −2.67; p < 0.001). Less repetition of LLLL or RRRR does not contradict more perseveration in alternation: the repeated strategy can itself be alternating LRLR rather than always turning the same direction.

What did the locomotor measurements show?

Total turns during the one-hour exploration provided a separate index of locomotor activity. For the overall mixed model, treatment had a significant effect (F(2,122) = 4.47, p = 0.01), as did the treatment-by-stage interaction (F(3,122) = 4.0, p < 0.01). The treatment-stage main effect alone was not significant (F(3,122) = 1.15, p = 0.33).

After challenge, amphetamine-treated fish made substantially fewer turns than water-treated controls; the authors reported p < 0.001 for this comparison. The paper described nicotine challenge as producing increased locomotion relative to baseline with reported p = 0.03. Because the numerical interval printed for that nicotine comparison crosses zero, the precise magnitude and uncertainty warrant additional scrutiny rather than an emphatic numerical claim. The safest conclusion is that the two drugs produced distinguishable locomotor profiles in this animal model, with amphetamine challenge linked to hypolocomotion against controls.

Evidence table 2: Why did the study’s three behavioral readouts matter?

The three metrics answer different questions. Choosing only the increasing alternation percentage would conceal a simultaneous reduction in movement and a narrower behavioral repertoire. Choosing only the reduced movement would miss the alternation change. An adequate scientific interpretation integrates them without promoting one proxy into a general human impairment assay.

Were amphetamine or nicotine actually measured in blood or tissue?

No quantitative drug assay of an authentic fish or human biological matrix was reported as an outcome of this experiment. The treatment concentrations of 25 µM and 5 µM were laboratory exposure specifications in aquarium water. Behavioral tracking recorded choices and arm entries. The study did not report measured d-amphetamine, nicotine, metabolites, or internal drug concentration-time curves in fish blood, whole brain, liver, or other tissues.

Consequently, it would be a category error to convert the stated water concentrations into a human venous whole-blood amphetamine number, a urine drug-test cutoff, an oral-fluid level, or an estimate of someone’s last use. Even an independently validated plasma assay in a different experiment would not automatically supply a conversion to forensic whole blood. Readers seeking an explanation of why a specimen’s physical matrix matters can consult the published guide on biological matrices in toxicology.

What does a significant group effect say about individual fish or people?

A significant treatment comparison establishes evidence that the studied group patterns differ under the stated analysis. It does not mean every amphetamine-exposed fish alternated more, or that any single fish was correctly classified by its maze performance. The models account for repeated observations, and individual variability and limited group size remain relevant. Some contrasts failed to meet the statistical threshold, despite changes in their point estimates.

More importantly, animal behavior is separated from human legal proof by a long inferential chain. The design did not assess real-world driving, workplace vigilance, eyewitness identification, accident reconstruction, or a person’s mental condition at a specified clock time. A litigant cannot take p < 0.001 for fish maze behavior and conclude a particular driver was impaired—or unimpaired—after amphetamine exposure. That conclusion would require case-specific evidence and appropriate validation not supplied here.

Why is this a useful example of behavioral endpoints disagreeing?

The study is useful precisely because the measures do not all point in the same direction. One operational indicator associated with working memory improved under repeated amphetamine conditions. An indicator of flexible strategy use did not show the same favorable pattern. Activity declined against controls after challenge. These are not mutually exclusive observations; they reflect different components of task behavior and illustrate the hazards of using “cognition” as though it were one uniform scalar quantity.

A second lesson concerns comparator drugs. The nicotine group increased alternations after challenge but did not show the same significant change in simple repetitive tetragrams. These outcomes are not proof that nicotine is cognitively safer in humans or that one stimulant is therapeutically preferable. They are observations in a particular fish strain, schedule, water concentration, and maze. Translational models can suggest testable hypotheses without providing direct human efficacy or forensic impairment validation.

How should analytical confirmation be distinguished from behavioral inference?

A behavior recording is an authentic observation of movement, but it is not chemical confirmation of which substance is present in a physical biological specimen. In a real forensic toxicology case, drug identity and concentration require an appropriately collected specimen and validated analytical methods. A nominal exposure concentration in fish water does not satisfy those requirements. For general context see confirmatory drug testing; that guide addresses a different stage of evidence assessment than the Y-maze outcome in this paper.

Similarly, therapeutic drug monitoring relies on matrix-specific assays and a clinical indication, neither of which this animal study established. The published overview of therapeutic drug monitoring explains the distinct use of quantified patient samples. Mentioning these methods does not imply that the Cleal study measured plasma, serum, or therapeutic blood concentrations. It did not.

What methodological limitations affect interpretation?

The initial randomized groups were small, and the principal figure included fewer analyzed fish in two arms than the allocation count. Only one exposure concentration per stimulant was used; the investigators did not present a concentration-response series that establishes a monotonic effect or a toxic threshold. The study was repeated-measures and included several behavioral domains and comparisons; no broad cross-domain multiplicity adjustment beyond the stated post hoc framework should be presumed.

The Y-maze links alternation patterns to working-memory-related performance, but there was no independent human-style declarative memory test. Fish water uptake and individual drug concentrations were not measured. Sex was balanced at assignment but the paper does not establish a robust sex-specific exposure-response law. The two drugs were studied at differing molar environmental concentrations and cannot be described as equipotent doses. Nor was the 48-hour withdrawal an analogue of a legally defensible timeline for human use.

There is also a useful source-precision caution: some narrative statements about the nicotine timing and its locomotor confidence interval are difficult to reconcile with adjacent reported statistics. A responsible synthesis privileges the expressly identified baseline, exposure, withdrawal, and challenge comparisons and does not silently repair an inconsistent confidence interval with an invented number. Where the primary report does not provide a clean estimate, uncertainty remains.

What can an expert fairly say about this paper?

An accurately bounded statement is that, in adult zebrafish subjected to 14 days of waterborne d-amphetamine or nicotine treatment followed by withdrawal and challenge, the researchers observed increases in Y-maze alternating-turn strategies and drug-specific changes in movement and flexibility. At its strongest, the experiment supports the proposition that a single “improved” behavioral endpoint can coincide with less favorable changes in other tasks or behavioral dimensions.

It does not establish the amount of d-amphetamine absorbed by an individual animal, does not directly quantify drugs in any legally relevant human blood specimen, and is not a validated instrument for assessing personal intoxication, criminal intent, driving capacity, memory credibility, or causation of a particular incident. Those questions require separate human, analytical, and case-specific foundations. The original study can be checked directly at doi:10.1007/s00213-021-05844-5.

Scientific FAQs

1. Does increased Y-maze alternation prove that d-amphetamine improves all memory?

No. Alternating-turn patterns are an operational working-memory-related measure in this particular zebrafish apparatus. The result does not establish generalized memory improvement, human autobiographical recall, or better testimony accuracy. The investigators simultaneously recorded outcomes suggesting less flexible search behavior.

2. Were 36 fish analyzed in every graph?

Thirty-six were initially randomized, twelve per group. The main behavioral figure lists eleven water controls, eleven amphetamine fish, and twelve nicotine fish. A figure-specific analyzed denominator should not be replaced by the randomized count or multiplied by the number of tests, periods, and recorded turns.

3. Did alternation remain significantly elevated after 48 hours without d-amphetamine?

The comparison to baseline during the withdrawal stage was not statistically significant at the authors’ stated criterion (p = 0.06). This is different from the significant increase after the 14-day course and after drug challenge. Nonsignificance does not establish exact equality.

4. Were the 25 µM amphetamine and 5 µM nicotine values blood-test results?

No. They were nominal water-exposure concentrations used to treat zebrafish. The experiment did not report quantitatively assayed drug in human venous whole blood, fish plasma, tissue, or urine and cannot be used to infer those concentrations from its behavioral results.

5. Why can an animal alternate more while being less flexible?

Alternation measures one recurring search sequence. A fish can use that sequence with increased frequency while reducing other exploration patterns. Behavioral flexibility asks whether strategies shift appropriately; the existence of one frequently repeated useful pattern does not itself answer that question.

6. Did nicotine and d-amphetamine have identical effects on movement?

No. The researchers described reduced total maze turns after amphetamine challenge compared with controls and a different locomotor pattern after nicotine challenge. The two substances were not administered at matched absorbed doses, so the experiment cannot establish comparative human safety or impairment from these differences.

7. Can these results establish DUI impairment, recent use, or accident causation?

No. The study used fish exposed through water and measured maze behavior, without chemically measured human whole-blood concentrations, actual driving, individual toxicokinetic curves, or case-specific event reconstruction. Its legitimate role is limited to demonstrating behavioral complexity and generating mechanistic hypotheses that need separate human validation.