Psychosis: The Utility of Ketamine as a Pharmacological Model of Psychotic-like Symptoms in Rodents: A Review of Dosage Regimens

Schizophrenia is characterized by a constellation of positive, negative, and cognitive symptoms [29,30,31]. Positive symptoms, those that represent the addition of abnormal behaviors, include hallucinations, delusions, formal thought disorder, and bizarre behavior [31,33,34]. In contrast, negative symptoms reflect the loss or diminution of normal function, such as avolition, apathy, alogia, and anhedonia [31,35]. Cognitive symptoms are also central to psychotic disorders and involve disorganized thinking, memory impairments, attentional deficits, and difficulty maintaining focus [32]. Notably, some of these cognitive symptoms overlap with those seen in attention-deficit/hyperactivity disorder (ADHD), which is often comorbid with psychosis [30]. Among these, deficits in attention are particularly debilitating, as they disrupt the ability to complete tasks and to maintain coherent cognitive engagement or verbal expression.

Selective and sustained attention deficits are hallmarks of schizophrenia, bipolar disorder, and ADHD [27,29]. The American Psychiatric Association [36] identifies impaired sustained attention, impulsiveness, and hyperactivity as the core triad of ADHD. In bipolar disorder, distractibility during manic episodes is so prominent that it is included as a diagnostic criterion in the DSM-V [36]. To explain the diverse symptomology observed in psychosis, two major neurobiological frameworks have emerged as follows: the dopamine hypothesis and the glutamate hypothesis. Recent research increasingly supports a convergent model, integrating these mechanisms and offering new insights into the pathophysiology of psychotic disorders.

Several rodent models have been developed to emulate psychotic-like symptoms using a range of methodological approaches, including genetic, neurodevelopmental, lesion-based, and pharmacological manipulations. These models aim to elucidate the neural mechanisms underlying cognitive dysfunctions commonly associated with schizophrenia, bipolar disorder, and ADHD, particularly deficits in learning and memory, executive function, and attention. Among these, pharmacological models offer a practical and scalable platform for investigating the neurobiology of psychosis. This review focuses specifically on the use of ketamine in rodents as a pharmacological model for inducing psychotic-like symptoms.

Ketamine (2-chlorphenyl-2-methylamino-cyclohexanone, KET) is a PCP derivative initially developed as a safer anesthetic with analgesic and amnesic properties [79,80]. Beyond its clinical applications, KET has gained recognition in psychiatric research for its ability to model psychotic symptoms, and more recently, as a fast-acting anti-depressant [81,82]. First synthesized from PCP in 1962 to minimize PCP’s neurotoxicity [80], KET has since become widely used in both human and animal studies to induce symptoms relevant to schizophrenia and bipolar disorder.

In human studies, acute administration of KET transiently increases positive (e.g., hallucinations and delusions), negative (e.g., blunting of affect), and cognitive symptoms (e.g., impaired memory and attention) in both individuals with schizophrenia and healthy controls [47,50,83,84,85]. While schizophrenic patients typically experience auditory hallucinations, healthy participants receiving acute KET more often report visual hallucinations [47,86]. Interestingly, chronic PCP users report both auditory and visual hallucinations [49]. However, acute KET models are limited in their ability to simulate long-term changes in brain structure and function associated with chronic psychosis [87,88].

In contrast, sub-chronic (<3 mos) and chronic (>3 mos) KET use in humans has been associated with more persistent schizophrenia-like symptoms, including delusions, cognitive impairments, depression, and disassociation [49,50,64,84,86,89]. Studies have demonstrated that chronic usage of KET or PCP induces chemical and structural changes in brain region connectivity similar to those seen in schizophrenic individuals. These changes were seen in both human studies and animal models [47,90,91,92].

In rodents, chronic KET administration produces prolonged behavioral and cognitive changes consistent with schizophrenia, including impairments in hippocampal memory, working memory, and fear learning [89,93], mirroring those observed in schizophrenic patients [94,95,96]. These deficits in KET-treated rodents are typically assessed using tasks such as the novel object recognition task [89,97,98] and trace fear conditioning protocols [24,99,100,101].

Like KET, PCP administration in animals also induces changes in brain structure that resemble those seen in schizophrenia, including reductions in gray matter volume and alterations in neurotransmitter signaling [27,49]. Functional neuroimaging studies have linked KET’s NMDA receptor antagonism to disruptions in prefrontal and limbic circuits. For instance, the NMDA receptor blockade is associated with negative symptoms [102], while increased glutamate/glutamine levels and altered blood flow patterns in the prefrontal cortex are correlated with positive symptoms [47,77,85,103,104]. Reductions in thalamic n-acetyl aspartate [92], prefrontal grey matter volume [90], and frontal cortex blood flow [84,105] have all been observed in chronic KET or PCP users, mirroring abnormalities in patients with psychosis [55].

Despite its widespread application in preclinical studies of psychosis-like symptomology, KET administration protocols vary widely in terms of dose, route, frequency, and duration. Previous reviews on KET have addressed several topics in KET research including its efficacy for anesthesia, analgesia [106,107], influencing depression [108,109,110,111] and fear memory [112,113], and its effects on the dopaminergic system [114,115]. Further, while there are numerous reports of the efficacy of KET to elicit psychotic- and schizophrenia-like behavioral and neurological symptoms in rodents, the dosing protocols reported vary markedly. Therefore, this review preferentially focuses on comparing acute, sub-chronic, and chronic KET regimens in rodent models to help researchers determine the most appropriate dosing strategy to meet specific experimental goals. While definitions of “sub-chronic” and “chronic” vary across studies, we define sub-chronic administration as treatment lasting ≤30 days. Notably, dose equivalency between mice and rats is approximately 2:1, respectively [116,117]. The review highlights findings from classic and contemporary studies across several behavioral tasks, emphasizing the need to align KET protocols with the cognitive, affective, or neurochemical dimensions of interest.

Electrophysiological studies provide critical insight into the acute effects of KET on excitatory neurotransmission within circuits implicated in psychotic-like behavior. Razoux et al. [120] administered an acute IP dose of 25 mg/kg KET to Wistar rats and observed a transient increase in the amplitude of prefrontal cortex-evoked excitatory postsynaptic potentials (EPSPs) in the shell region of the nucleus accumbens. This increase was observed between 10 and 40 min post-injection and coincided with the onset of hyperlocomotion, stereotypic behavior, and ataxia—features of the KET “alpha phase”. These results support the model proposed by Takahata & Moghaddam [158], which suggests that the behavioral effects of NMDA antagonists like PCP and KET are mediated via disrupted glutamatergic projections from the prefrontal cortex to subcortical structures, including the nucleus accumbens and ventral tegmental area. Concomitantly, reduced glutamatergic signaling from the hippocampus to nucleus accumbens may further contribute to this dysregulation [158,159].

KET and PCP bind within the ion channel pore of NMDA receptors, a site accessible under depolarized conditions when the Mg²⁺ block is relieved. Consequently, KET preferentially impacts NMDA receptors on tonic firing GABAergic interneurons over burst firing pyramidal neurons due to their sustained depolarization [160,161,162]. This leads to disinhibition of excitatory pyramidal neurons within cortical and limbic circuits, potentially underlying KET-induced behavioral phenotypes that resemble psychosis. These disruptions in cortico–basal ganglia signaling are hypothesized to underlie core symptom domains of schizophrenia, including impaired salience attribution, disorganized thinking, and aberrant motivation [163].

An early report [128] revealed that systemic administration of KET (80 mg/kg IP), repeated daily, induced high-amplitude slow-wave bursting and spontaneous seizure-like activity in EEG recordings from the amygdala and hippocampus of freely moving Sprague–Dawley rats. A similar pattern of results were found after daily 30 mg/kg administration that was repeated for months [128]. A more comprehensive view of KET’s neurophysiological impact has emerged through system-level analyses using techniques such as local field potential (LFP) and multi-structure microelectrode array recordings. In anesthetized rats, sub-anesthetic doses of KET reduced firing rates in reticular thalamic and prefrontal cortical neurons, decreased delta (1–4 Hz) power, and elevated beta (12–30 Hz) and gamma (30–100 Hz) power [130]. In contrast, freely moving rats given 50 mg/kg KET IP showed increased firing rates in the reuniens thalamus and CA1 region of hippocampus, along with elevated delta and gamma power in both hippocampus and thalamus [131]. These effects of KET mirror EEG abnormalities, particularly increased delta and gamma oscillations, commonly observed in schizophrenia patients [164].

Nasretdinov et al. [132] utilized chronic, large-scale electrode arrays in Sprague–Dawley rats to assess neural dynamics across prefrontal cortex, striatal, basal ganglia, thalamic, and limbic regions. Following acute KET administration (25 or 50 mg/kg IP), they observed significant increases in average LFP power across these structures, coupled with a pronounced reduction in functional connectivity, measured as correlated activity between electrode pairs, both within and between brain regions. These findings suggest that while KET enhances localized neuronal activity, it simultaneously disrupts network-level coherence across cortico–basal ganglia–thalamic circuits. Such desynchronization may contribute to cognitive and sensorimotor deficits characteristic of psychotic disorders.

A broad array of behavioral tasks has been employed to validate KET-induced rodent models of schizophrenia and bipolar disorder. Although rodents cannot perform standardized cognitive batteries employed in human studies, many behavioral tasks have strong ecological validity and can be effectively modified for preclinical models. As Powell and Miyakawa [28] noted, one symptom commonly observed in schizophrenia is “psychomotor agitation”, which manifests in rodents as hyperactivity and stereotypic movements. The open field task is widely used to measure changes in locomotor activity and is among the most frequently employed tasks to establish construct validity in KET-based rodent models of psychosis. Thirteen of the 21 reviewed studies used the open field task and consistently reported increased locomotor activity following both acute and sub-chronic KET administration (see Table 1).

Beyond locomotion, sixteen studies employed additional tasks to assess cognitive impairments, particularly those involving hippocampal and prefrontal cortex function, brain regions known to be disrupted in schizophrenia and bipolar disorder. For example, novel object recognition and trace fear conditioning tasks are frequently used to assess long-term hippocampal memory in rodents and both have demonstrated cross-species validity [28,165,166,167,168,169,170]. Tasks such as the association mismatch task [171] assess spatial memory, while working memory, one of the most reproducible cognitive deficits in schizophrenia [172,173], is examined using tasks like the Y-maze, radial-arm maze, and the delay match-to-sample (DMTS) tests. The discrete paired-trial variable-delay T-maze task closely mirrors human delayed non-match to sample tasks and is predictive of prefrontal cortex-related working memory deficits [174].

Sensorimotor gating, another domain disrupted in psychosis, is often tested using pre-pulse inhibition of the startle reflex, a reliable correlate of human deficits [175,176]. Impaired latent inhibition, another hallmark of schizophrenia patients [177,178], is well modeled using established rodent protocols [179,180].

To assess attention and executive function, studies frequently employ the five-choice serial reaction time task (5-CSRTT), adapted from sustained attention tasks in humans [181]. Similarly, the attentional set-shifting task parallels the Wisconsin Card Sorting Test, measuring cognitive flexibility [182,183]. Trace fear conditioning protocols can also be used to measure attention in rodents [184].

Negative symptoms such as anhedonia, behavioral despair, and deficits in social interaction are assessed in rodents using tasks like thermal nociception [185], forced swim test [186], sucrose preference [187], elevated plus maze [188], neophagia [103], and tests of social interaction and social memory.

Together, these behavioral protocols offer robust and diverse metrics for evaluating KET-induced psychosis-like states in rodents. While this list is not exhaustive, it reflects the core behavioral assays used across the 28 representative studies reviewed here and highlights the translational relevance of rodent behavioral phenotyping in preclinical psychiatric research.

A total of 28 rodent studies were analyzed to assess the behavioral and neurobiological effects of differing KET dosing regimens. Thirteen studies used rats, with acute doses ranging from 4 mg/kg subcutaneous to 25 mg/kg IP, and sub-chronic schedules ranging from 25 mg/kg to 80 mg/kg IP once daily for up to three months. One study employed intravenous administration with washout periods of 0 to 60 min. Rat strains included Lister, Wistar, Sprague–Dawley, and Long–Evans.

Eight studies used mice, with acute doses ranging from 5 mg/kg to 100 mg/kg IP; and sub-chronic regimens from 4 mg/kg to 100 mg/kg, administered over 5 to 21 days with washout periods of 0 min to 14 days. Mouse strains included C57BL/6J, 129SvPasIc, C57BL/6JOlaHsd, C57Bl6N, and Swiss–Kunming (see Table 1). Dose equivalents between mice and rats are generally estimated at a 2:1 ratio, respectively [116,117].

Behavioral effects of KET depend not only on dose but also on timing. The pharmacodynamic profile includes the half-life phase (5–15 min) characterized by ataxia or cataplexy; the alpha phase (15–40 min) marked by hyperlocomotion and stereotypy [129]; and the beta phase when behavior is returning to baseline or showing sedative-like effects. Studies indicate that the alpha phase aligns with behavioral testing in many acute dosing models and may not fully capture persistent psychosis-like symptoms. Thus, sub-chronic regimens with post-administration washout periods are more effective in modeling enduring cognitive impairments [89,97,189]. Washout periods are especially critical for attention impairments, and both low doses with long washouts and moderate doses followed by short washouts yielded significant deficits [126,137,189]. For working memory, even low dose KET with short washouts (e.g., 2 days) resulted in long-lasting deficits [125]. Anxiety and depression-like symptoms were sensitive to both dose and washout length [93,136,140], reinforcing the need for precise protocol design.

While IP administration is preferred for practicality, analyses of blood serum levels indicate that this route produces peaks and troughs in plasma levels unlike IV infusions, which better mimics human pharmacokinetics [129]. Some studies suggest that more frequently IP dosing (e.g., twice daily) helps sustain bioavailability and enhance translational relevance [129,189,190].

The behavioral impairments induced by sub-chronic KET correlated with broad neurobiological changes. For example, (i) elevated levels of glutamate/glutamine and reduced expression of NMDA receptor subunits [118,119,120]; (ii) elevated oxidative stress markers and protein damage in key brain regions, prefrontal cortex, hippocampus, striatum, and amygdala, mirroring findings in bipolar disorder [124,151]; (iii) sex-dependent increases in hippocampal 5-HT turnover and synaptic proteins [140]; and (iv) decreased c-Fos-, BDNF, and parvalbumin-positive interneurons in cortical and hippocampal regions [97,133], consistent with postmortem studies of psychosis patients [155,157].

While some behavioral tasks for rodents are well suited for translation of results to human behavior, others face limitations due to being based on face or predictive validity. Throughout this review, we have purposely referred to the ability of ketamine to induce behavioral changes in rodents that are psychotic-like or schizophrenia-like since the altered behaviors in rodents bear some resemblance to the human disordered behavior, but it is not a perfect match. Human task parameters may not be able to be implemented as they are non-existent in rodents, such as guilt, self-confidence or delusions, while others like punishment and probabilities, cannot be matched between species [191]. Hyperlocomotion is used extensively as a measure of manic-like behavior, as the core positive symptoms of schizophrenia; however, hyperlocomotion must also be recognized as nonspecific arousal state that likely interferes with the expression of other high-order behaviors. For example, hyperlocomotion or persistent arousal will complicate the interpretation of reduced social interaction or impaired social or object recognition memory. The challenge is even more concerning when one considers how to interpret ketamine’s effects in a test in which intact memory is inferred from a lack of movement, such as the conditioned freezing response in fear conditioning. Thus, it is critical, where possible, to include measures of higher order behaviors that are less confounded by movement differences.

The forced swim test, which is sensitive to drugs that are effective anti-depressants in humans, has good reliability and predictive validity, but poor construct validity [192]. The common measure of immobility after struggling to escape the forced swim test is considered to reflect despair but could also represent the rapidly learned response to stress, or the shift to a passive coping mechanism. Some have argued that likening immobility to behavioral despair borders on anthropomorphism. Either way, numerous trials of putative anti-depressant drugs that demonstrate promising results in rodents in the forced swim test go on to fail in human trials; thus, there are significant concerns regarding the reliability of the forced swim test as a screen for depressive-like behavior [193,194].

Tests designed to measure anxiety in rodents carry similar interpretational challenges in relating rodent behavior to that of the human feeling of anxiety. The most commonly used tests are those in which administration of drugs that reduce anxiety in humans are also effective in changing approach-avoidance behavioral profiles in rodents [195]. Tasks measuring anxiety-like behavior in rodents are challenged by the need to distinguish fear, the escape response to immediate danger, from anxiety, the response to potential threats or danger [196]. Tests such as novelty-suppressed feeding may potentially confuse anxiety and anhedonia. Often tasks that measure cognition, such as set-shifting and reversal learning, and tasks sensitive to stress and aging, such as open field, are susceptible to non-cognitive factors such as age, anxiety, and motor deficits, that mask true cognitive impairments [197]. These limitations should be considered in the experimental design and interpretation of differing KET dosing schedules. To capture stable not just transient states, the use of multiple tests for the same function should be considered.

A comprehensive review of KET metabolism and pharmacokinetics [198] revealed the following: KET has low binding to plasma proteins [199] and is widely distributed due to its liposolubility [198]. KET is metabolized to norketamine, an active metabolite, mainly hydroxylized in 6-hydroxy-norketamine and excreted in urine and bile following glucuronoconjugation [198]. The pharmacokinetics are not well understood. Other lesser metabolites are also produced, including 4-OH-norketamine, 6-OH-norketamine, and 5-OH-norketamine [198]. Norketamine is found in blood within 2–3 min and peaks in approximately 30 min, remaining more than 5 h, following a KET IV bolus [200,201]. Sites of metabolism include the liver, kidneys, intestine, and lungs, particularly in animals [202]. The elimination clearance of KET is 1000–1600 mL/min or 12–20 mL/min/kg [203] compared to norketamine [204,205]. The half-life is 2–3 h. Clearance of KET may be 20% high for females than males [206]]. Due to the accumulation of norketamine, less KET is needed over time when IV administered by [204,206]. The combination of the rapid half-life of KET and the lasting presence of its active metabolite may provide support for the contention that twice daily administration of sub-chronic KET provides lasting suppression of NMDA receptor activity which effectively induces persistent cognitive deficits even after washout of KET.

When administered IV, KET reaches its receptors at a high rate with a transfer half-life of less than 1 min, while intramuscular administration has a bioavailability of approximately 93%, with a plasma peak of 5 min [198]. Oral administration bioavailability is approximately 20% [207] with a concentration peak in 20–30 min [208]. Bioavailability of intrarectal administration is approximately 25% and intranasal is about 50% [200].

Limitations of pharmacokinetic data can limit the interpretation of dosing comparisons. As stated, the pharmacokinetics of norketamine are not well understood. While KET bioavailabilty after IV administration is high, it is also eliminated rapidly. However, because norketamine accumulates once a level of KET is reached, there is less need for more frequent KET. A dosing regimen of KET twice per day, or bolus doses, may keep KET levels high enough to allow the accumulation of norketamine without the need for IV administration.

Here, we note that some dosing regimens produce anti-depressant-like effects with psychosis-like effects produced in others. Anti-depressant mechanisms may confound or interact with psychosis-like effects. These potential interactions may be a result of shared pathways involving synaptic plasticity, the glutamatergic system, and dopamine signaling [209,210]. Further, depression and psychosis are known to both involve abnormal synaptic plasticity [209,210]. Depressant drugs, like KET, block NMDA receptors, which induces anti-depressant effects and psychosis-like symptoms in healthy individuals [209]. In the glutamatergic system, the “glutamate burst” is associated with depressant action. A surge in extrasynaptic glutamate, followed by AMPA receptor activation seen in anti-depressants like KET, appears to be a key mechanism of the therapeutic effect of anti-depressants [210]. In psychosis, disruptions in glutamate signaling are central to the pathophysiology. Disruptions in the glutamate pathway could exacerbate psychosis [210].

The dopaminergic system is also affected by changes in glutamate [210] induced by depressants, and such changes can indirectly modulate dopamine release causing psychosis-like symptoms in some individuals [210].

It is of interest to note that both Manooki et al. [133] and Thelen et al. [124] observed sex differences in the expression of depressive-like behaviors in rats and mice, respectively. However, the effect of sub-chronic KET in the forced swim test was species-specific: the male mice exhibited anti-depressant-like effects, with reduced immobility in the forced swim test, while female mice displayed increased depressive-like behavior, with greater immobility. In contrast, the female Wistar rats exhibited reduced immobility compared to the male rats; that is, the females spent more time struggling, trying to escape, and less time immobile than the control rats [133]. Females also displayed behavior indicating heightened anxiety compared to males. This difference in affective behavior in females may be due to an increased sensitivity to the disruption of the glutamate pathway or possibly due to the higher clearance rate of KET in females compared to males.

No new data were created or analyzed in this study. Data sharing is not applicable to this article.