Does D-Amphetamine Raise Cortical “Up-State” Activity—and Why Did Prazosin Block It in Rats?

Short answer: In an experiment involving invasive recordings from anesthetized male rats, intravenously administered d-amphetamine increased the fraction of time that medial prefrontal cortical electrical activity occupied a so-called Up state. The response depended on the animal’s initial cortical state. Blocking D1-like or D2-like dopamine receptors reduced, but did not consistently eliminate, amphetamine’s electrical effect; blocking α1 adrenergic receptors with prazosin prevented a statistically demonstrable response in the tested pretreatment protocol. Dopamine receptor agonists acting together partly reproduced the pattern, whereas L-DOPA did not. These findings concern rat brain local-field potentials, not a measured human blood-amphetamine concentration, a validated test of alertness, or proof about a driver or litigated event.

The original primary study is Shen and Shi’s 2021 experiment, “Amphetamine Promotes Cortical Up State in Part Via Dopamine Receptors”, published in Frontiers in Pharmacology. Its value lies in separating several related but different questions: whether amphetamine changes an electrophysiological pattern; whether dopamine and noradrenergic receptor manipulations alter that pattern; and whether any such result predicts observable human performance. The experiment addressed the first two, not the third.

What is a cortical Up state?

Groups of cortical neurons can alternate between periods of comparatively greater electrical activity and periods of relative inactivity. The active periods are called Up states; intervening Down states are associated with reduced neuronal firing and synaptic activity. Up/Down cycling is often discussed in relation to slow-wave sleep, anesthesia, and certain quiet-wakefulness conditions. It is not itself a laboratory measurement of consciousness, competence, or drug intoxication.

The researchers placed an electrode into the medial prefrontal cortex and recorded extracellular local-field potentials. They mathematically reversed the displayed polarity so that the selected upward deflections represented Up-state transitions. The principal endpoint, designated T_UP, was the percentage of analyzed recording time classified as Up state. A second endpoint, P_SO, represented the absolute mean spectral power between 0.3 and 1.5 Hz—an electrophysiological measure of slow oscillation, expressed in the study’s voltage-squared units. Increasing T_UP and decreasing P_SO are related but not interchangeable observations.

Neither endpoint is the quantity of amphetamine in blood. Neither is an assay for the neurotransmitter dopamine. These distinctions are essential when a mechanistic neuroscience paper is cited in forensic or clinical discussions.

Which animals, drug, and recording technique did the investigators use?

The subjects were male Sprague–Dawley rats weighing approximately 250–400 grams, aged about 2.5–5 months. The rats were maintained under chloral hydrate anesthesia: an initial 400 mg/kg intraperitoneal dose, followed by intravenous supplemental dosing as described in the original protocol. Body temperature was controlled near 36–37°C. A glass electrode recorded local-field potentials from the medial prefrontal cortex, with signals filtered over approximately 0.1–200 Hz and digitized at 20 kHz.

D-amphetamine and the receptor-active comparison drugs were administered through a lateral tail vein. Thus, the operative experimental exposure was an intravenous dose in milligrams per kilogram—not an oral prescription, smoked drug, or measured human venous whole-blood concentration. The researchers used successive doses in certain protocols, and repeated recordings after each dose occurred within the same animal. A two-minute electrophysiological recording and the many samples used to calculate its spectral characteristics are not new independent rats.

The large baseline-response analysis contained 126 rats. Smaller mechanistic subexperiments used separately reported sample counts. Those subgroup counts cannot safely be added to 126 as though every animal entered only one nonoverlapping experiment; the article does not provide a complete unique-animal accounting across all illustrated protocols. The defensible biological denominator must be attached to the specific source comparison.

Why did a rat’s initial cortical state matter?

The investigators stratified the 126-rat baseline analysis by the percentage of time already spent in the Up state before amphetamine. Ten rats began below 25%; 89 began between 25% and 45%; and 27 began above 45%. The baseline classification was an electrophysiological state definition, not a diagnosis of impaired attention or a clinically validated severity threshold.

Intravenous d-amphetamine increased Up-state time across these categories, but animals in the lowest-baseline stratum showed a weaker dose-response relationship than animals with intermediate or higher starting values. The reported model comparison between the lowest and intermediate strata gave an intercept F(1,130) = 90.2, p < .001 and a slope F(1,324) = 18.37, p < .001. The corresponding intercept comparison against the highest-baseline stratum was also significant, F(1,67) = 37.13, p < .001, but the slope comparison was not: F(1,95) = 0.037, p = .847.

That last null finding matters. The paper supports baseline-dependent differences, but not the claim that every possible aspect of the response curve differed significantly between every pair of strata. The investigators also found no statistically significant suppression of the slow-oscillation power measure in the lowest-baseline group, even though significant suppression was observed in the intermediate and higher groups.

Figure 1 contains an illustrative individual rat whose Up-state percentage rose from 38% to 95% and then 98% after cumulative 1 and 2 mg/kg amphetamine, respectively. Those numbers describe that example animal; they should not be misrepresented as the group means for 126 rats or as a human concentration-response result.

First evidence table: baseline electrical state versus the amphetamine response

How did the researchers separate dopamine from noradrenergic actions?

D-amphetamine changes several monoaminergic signaling systems. The investigators reasoned that a dopamine contribution was plausible because an earlier experiment had shown strong involvement of α1 adrenergic receptors, whereas dopamine transporter blockade also affected Up-state time. Instead of merely attributing the electrical result to a named neurotransmitter, this study used receptor-selective pharmacological challenges to test how the response changed.

The main comparisons employed SCH23390, a D1-like receptor antagonist; raclopride, a D2-like receptor antagonist; and prazosin, an α1 adrenergic receptor antagonist. The investigators administered antagonists before or after amphetamine in distinct protocols. That temporal distinction matters: preventing a response from developing is not the same experimental question as reversing an already established response.

Drug-induced changes in local-field potential patterns can support receptor involvement. They do not directly measure dopamine release, norepinephrine release, synaptic receptor occupancy, transporter concentrations, or precise receptor location. Systemic drug injections also complicate attribution of a response to a single anatomical nucleus.

What happened after D1-like receptor blockade?

In the pretreatment experiments, SCH23390 at 0.1 mg/kg did not itself significantly change baseline Up-state time: 45.3% ± 2.3% before versus 45.0% ± 3.4% afterward, F(1,6) = 1.161, p = .323. The study also reported no significant standalone effect on slow-oscillation power, p = .514. Amphetamine still promoted the Up state after this antagonist was present, but its response curve was smaller than the baseline-adjusted control curve.

The reported comparison with control rats included a T_UP intercept test of F(1,20) = 23.37, p < .001 and a dose-response slope test of F(4,81) = 3.957, p < .01. A statistically supported attenuation is not equivalent to total abolition. The residual amphetamine response under D1-like blockade is among the paper’s reasons for describing dopamine’s role as partial rather than exclusive.

When SCH23390 was delivered after amphetamine, it also partly reversed an existing Up-state response, with dose-dependent differences in the comparison. That sequence cannot prove that D1 receptor activation is the sole requirement for wakefulness. It says that altering a receptor-sensitive pharmacological pathway affected an electrical signal in anesthetized rats.

Did D2-like receptor blockade behave identically?

No. Raclopride’s effects were more complicated. It could independently increase baseline Up-state time: the article reported 35.5% ± 1.7% before versus 47.8% ± 4.6% after raclopride, F(1,11) = 18.79, p < .001. Its standalone comparison for slow-oscillation power was nonsignificant, F(1,11) = 0.788, p = .394. That baseline change means a comparison of raw post-treatment percentages without accounting for starting state could confuse two separate drug actions.

After baseline-adjusted comparisons, pretreatment with raclopride attenuated the amphetamine effect, particularly for higher doses in the illustrated protocol. The source reported a T_UP curve intercept difference of F(1,35) = 19.07, p < .001 and a slope difference of F(1,129) = 13.85, p < .001. When given after amphetamine, however, raclopride partially reversed responses in both lower- and higher-dose protocols, with a larger reversal in the lower-dose group.

The authors acknowledged that the pre- and post-treatment patterns did not entirely agree. Their possible explanation—that initiation and maintenance of the electrical state may recruit different processes—remains an interpretation to be tested, not a directly observed switch in a single known receptor circuit. Combining SCH23390 and raclopride pretreatment attenuated certain amphetamine doses but did not suppress every intermediate-dose effect consistently.

What made prazosin different?

Prazosin blocks α1 adrenergic receptors and, in this model, affected the amphetamine response more strongly than the selective dopamine antagonists. The 0.1 mg/kg prazosin pretreatment slightly reduced baseline Up-state time, from 39.8% ± 2.0% to 35.1% ± 2.2%, F(1,7) = 8.534, p < .05. Its baseline slow-oscillation-power comparison was nonsignificant, p = .482.

With prazosin pretreatment in place, the investigators did not detect a statistically significant amphetamine-driven increase in T_UP, F(5,17) = 2.231, p = .097, or a significant decrease in P_SO, F(5,8) = 0.964, p = .492. The researchers interpreted the pattern as prevention of the usual response in that protocol. Scientifically, those null comparisons should be reported as absence of a demonstrated change under the tested conditions; they are not an equivalence trial proving an effect of exactly zero.

In another experiment, prazosin administered after amphetamine substantially reversed the established Up-state increase in both a low-dose group and a high-dose group. The source described sample sizes of 14 and 11 rats in those respective post-treatment protocols. Additional amphetamine injections after prazosin did not restore the typical response in the high-dose group. This contrast supports an important α1-sensitive component in the particular anesthetized preparation, not an instruction to administer prazosin for intoxication or clinical stimulant effects.

Second evidence table: receptor interventions, genuine nulls, and the measured endpoint

Could dopamine receptor stimulation by itself reproduce amphetamine’s effect?

The study’s agonist experiments argue against a simple one-receptor explanation. A D1-like agonist, SKF38393, delivered over a 2.5–10 mg/kg range in 11 rats did not significantly increase Up-state time: F(3,31) = 1.13, p = .352. At the highest tested dose, it modestly reduced slow-oscillation power, but that isolated spectral result was not the same as a full amphetamine-like shift.

A D2-like agonist, quinpirole, in six rats had the opposite direction of effect on T_UP, reducing it across the tested range; the reported test was F(4,24) = 25.52, p < .001. Those findings show why the phrase “dopamine stimulation increases cortical arousal” is too crude for this specific preparation.

After pretreating six rats with quinpirole, the investigators administered SKF38393. Under that co-activation protocol, the D1-like agonist increased T_UP, F(4,24) = 6.248, p < .001, and decreased P_SO, F(2,24) = 6.142, p < .001. Nevertheless, those responses were weaker than the amphetamine comparison. The result is consistent with coordinated receptor involvement, but it does not identify every presynaptic step or establish a chemically measured quantity of released neurotransmitter.

Why did L-DOPA fail to behave like amphetamine?

L-DOPA is a biosynthetic precursor of dopamine and, indirectly, norepinephrine. In a separate seven-rat group, intravenous L-DOPA over a 20–80 mg/kg dose range reduced Up-state time, F(3,21) = 7.557, p = .001, instead of increasing it. Slow-oscillation power also declined slightly. In short, a precursor that can feed monoamine synthesis did not replicate the effects of a transporter-acting stimulant under the tested experimental conditions.

This was a negative mechanistic comparison, not a clinical trial of L-DOPA for attention-deficit/hyperactivity disorder or stimulant overdose. Dose, release dynamics, transport, receptor population and timing could all influence differences between these interventions. The paper proposed neurobiological explanations but did not directly establish which one caused the L-DOPA result.

Why do electrophysiological sample counts require care?

The independent unit in an animal experiment is ordinarily the animal when treatment assignments and biological responses are tied to each animal. The 126-rat baseline analysis is much stronger numerically than a single representative trace, but each rat contributed repeated 2-minute recordings. Recording at 20 kHz produces thousands of electrical samples per second; it does not transform each sample, spectral window or dose point into an independent biological subject.

The original analysis used mixed-effects models with individual-animal intercepts to account for repeated observations. The authors also adjusted for baseline T_UP because starting cortical state influenced dose response. Different receptor experiments reported different n values, sometimes with related rats participating in more than one sequence. Unless a unique-allocation ledger is available, it would be misleading to sum every figure’s n to advertise a larger independent cohort.

A forensic expert evaluating such literature should distinguish the denominator of a physiological comparison, the number of repeated electrode measurements used to construct it, and the number of truly independent biological preparations. Failing to separate these levels inflates apparent precision and can make mechanistic data look like robust clinical validation.

Did the researchers measure dopamine, norepinephrine, or amphetamine concentrations?

No direct quantitative drug-chemistry result establishes a human whole-blood concentration here. The study administered known nominal doses of receptor-active compounds and recorded their effects on a local-field potential. It did not supply a calibrated human venous whole-blood amphetamine or methamphetamine value, a separated plasma or serum concentration, an oral-fluid result, a urine metabolite level, or a paired pharmacokinetic curve. Likewise, changing a dopamine-receptor-sensitive physiological signal is not the same as directly assaying dopamine or norepinephrine in synaptic fluid.

That separation prevents two forensic errors. First, an intravenous rat dose expressed as mg/kg cannot be directly converted into a human blood ng/mL value. Species differences, route, distribution, drug stereochemistry, time and analytical matrix all affect exposure. Second, an electrical signal altered by a pharmacological blocker is not an analytic confirmation of the blocker’s exact receptor occupancy or a demonstration that amphetamine levels fell.

The relevant parent compound was d-amphetamine. It must not be casually relabeled methamphetamine, dextroamphetamine-plus-levoamphetamine combination therapy, or lisdexamfetamine. Those agents may share downstream systems yet differ in administration, metabolism, time course, active species and clinically studied effects.

Does a longer Up state mean an awake or safer subject?

The rats were anesthetized throughout electrophysiological recording. The researchers did not wake them for a controlled attention test and did not quantify independent measures of sustained attention, divided attention, driving, flight performance, memory formation, reaction to road hazards, interpersonal aggression or a clinical diagnosis. A cortical state associated in other research with arousal cannot be treated as a validated individual classifier for those outcomes.

Even inside the preparation, the two electrical markers did not always move in tandem. The lowest-baseline group showed increased Up-state time but no statistically established reduction in slow power. Quinpirole and L-DOPA changed certain readouts in directions that differed from amphetamine. This complexity argues against reading a single electrophysiological percentage as a general-purpose “impairment score.”

The findings may guide mechanistic research on stimulant effects and cortical networks. They do not establish that amphetamine necessarily improves wakefulness in a specific patient, that prazosin reverses clinical intoxication, or that a particular driver was either impaired or fit to drive at a stated blood concentration.

What experimental limitations matter most?

Anesthesia is a central limitation because cortical network dynamics under chloral hydrate may differ from unmedicated wakefulness. Only male rats were used, and the study’s electrophysiological endpoints were specialized invasive local-field potentials. Baseline differences affected dose response and required statistical adjustment. Some receptor antagonists shifted baseline physiology even before amphetamine was administered. Thus, their intervention effects cannot automatically be interpreted as a single clean receptor-specific subtraction.

Another limitation is pharmacological selectivity and anatomical localization. The drugs were injected systemically. Their effects do not prove where within the brain each relevant action occurred, nor do they distinguish all possible receptor subtypes and indirect feedback loops. Moreover, prazosin’s prevention of a statistically detectable electrical change was not designed as a human placebo-controlled efficacy or equivalence trial.

The smaller receptor-challenge groups, repeated dose measurements, several dose-response contrasts and method-dependent transformations also warrant caution. A p value does not state the probability that a biological mechanism is true. A null p value does not prove exact equivalence. The source’s own discussion identifies differences between pre- and post-treatment raclopride results, which are reasons for additional research rather than details to omit.

How should forensic testimony characterize this paper?

A scientifically precise summary would say that d-amphetamine changed local cortical field-potential patterns in an anesthetized rat model; baseline state predicted the size or pattern of several responses; D1-like and D2-like receptor interference partly attenuated them; α1 blockade had a stronger effect; and selective agonists plus L-DOPA demonstrated that simple dopamine stimulation was not sufficient to reproduce every outcome.

The testimony should also state what was not tested. This was not a human psychomotor impairment study, a courtroom credibility study, an emergency-treatment protocol, or a direct comparison between blood concentration and behavioral risk. It did not assay a case specimen or evaluate a specific intoxication event. Individualized conclusions about medication effect, timing, driving, criminal intent or cause of an accident require a separate evidentiary foundation.

The distinction between a measured neural signal, a pharmacological intervention, and an inferred clinical effect is the decisive boundary. Describing that boundary does not diminish the research; it identifies exactly what evidence it contributes.

Frequently asked questions

1. Was this a study of methamphetamine users?

No. The animals received intravenous d-amphetamine. No humans were enrolled, and methamphetamine was not the agent evaluated. Related stimulants should not be substituted for the exact studied substance.

2. Does “126 rats” mean 126 rats for every receptor comparison?

No. The 126-rat denominator belongs to the baseline-dependent amphetamine analysis, split into groups of 10, 89 and 27. Smaller groups were used for antagonist and agonist experiments. Repeated recordings, dose points and sampled electrical windows are not new animals.

3. Did prazosin prove that norepinephrine causes all amphetamine effects?

No. Prazosin prevented a detectable Up-state response under the experimental protocol, consistent with an important α1-sensitive component. The study did not quantify neurotransmitter release or prove an exclusive pathway in every brain region or species.

4. Did the dopamine blockers completely stop amphetamine?

No. SCH23390 and raclopride attenuated certain responses, but residual effects remained and the pattern depended on dose and timing. Their results differed from the stronger prazosin intervention.

5. Was the absence of an L-DOPA Up-state increase a nonsignificant result?

No. L-DOPA significantly decreased Up-state time, F(3,21) = 7.557, p = .001. That is a measured directionally opposite result, not a nonsignificant absence of any change.

6. Can these results be used to calculate a human blood amphetamine level?

No. The experimental dose was expressed in rat mg/kg, while the principal measurements were invasive cortical electrical recordings. There was no measured human venous whole-blood or plasma dose–concentration relationship and no validated forensic conversion.

7. Can this study establish a person’s alertness, DUI impairment, or accident cause?

No. The experiments did not observe human alertness tests, driving, collisions, individual behavior or case specimens. Mechanistic rat electrophysiology can inform hypotheses, but it cannot replace individual evidence or validate a clinical or legal impairment threshold.

Primary research and final evidence boundary

Source: Shen G and Shi W-X (2021), Amphetamine Promotes Cortical Up State in Part Via Dopamine Receptors, Frontiers in Pharmacology 12:728729, DOI 10.3389/fphar.2021.728729. This article is an original scientific interpretation of the published animal experiment. It reports the actual electrical, pharmacological and statistical endpoints without converting them into an unmeasured human blood concentration, diagnosis, driving determination or case-specific causal opinion.