The dopamine hypothesis remains the most enduring and clinically productive framework for understanding antipsychotic drug action, even as its original formulation has been substantially revised in light of neuroimaging, genetic, and pharmacological evidence accumulated over six decades.
The original dopamine (DA) hypothesis, formulated in the early 1960s, arose from two convergent observations. First, chlorpromazine and reserpine, independently observed to reduce psychotic symptoms, both acted to reduce dopaminergic neurotransmission, the former by receptor blockade and the latter by depleting presynaptic DA stores. Second, amphetamine and other DA-releasing agents produced a paranoid psychosis in healthy individuals that was clinically indistinguishable from acute schizophrenia, and exacerbated symptoms in patients with established schizophrenia.1 These observations led to the straightforward formulation that schizophrenia resulted from excess dopaminergic activity, and that effective antipsychotics worked by reducing it.
The discovery in the 1970s that the clinical potency of first-generation antipsychotics correlated almost precisely with their affinity for the D2 receptor subtype, across a range of six orders of magnitude in binding affinity, provided strong mechanistic support for this framework.2 The correlation between D2 binding affinity and clinical dose was one of the most striking structure-activity relationships in clinical psychopharmacology and remains foundational to understanding why high-potency agents such as haloperidol require doses measured in single-digit milligrams while low-potency agents such as chlorpromazine require hundreds of milligrams to achieve equivalent antipsychotic effect.
The revised or "updated" dopamine hypothesis, which emerged through the 1990s and 2000s, is more anatomically and functionally specific. Positron emission tomography (PET) imaging studies demonstrated that patients with acute schizophrenia show elevated presynaptic DA synthesis capacity and increased amphetamine-stimulated DA release in the striatum, consistent with mesolimbic dopaminergic hyperactivity driving positive symptoms such as hallucinations and delusions.3 However, the same imaging literature also demonstrated that the prefrontal cortex in schizophrenia is characterized by reduced DA activity, consistent with hypofunctionality of the mesocortical pathway and providing the neurobiological substrate for negative symptoms and cognitive deficits.3
This anatomically differentiated model, with concurrent mesolimbic hyperdopaminergia and mesocortical hypodopaminergia, has important pharmacological implications. Blocking D2 receptors throughout the brain addresses the mesolimbic excess but worsens the already-deficient mesocortical activity, which is one reason that first-generation antipsychotics (FGAs), which are non-selective D2 blockers, often worsen negative symptoms and cognitive function rather than improving them. The challenge of designing agents that suppress mesolimbic excess without further compromising mesocortical function is central to the rationale for the second-generation antipsychotics (SGAs) discussed in subsequent modules.
The dopamine hypothesis also does not fully account for the role of glutamate in schizophrenia pathophysiology. Phencyclidine (PCP) and ketamine, which are N-methyl-D-aspartate (NMDA) receptor antagonists, produce a syndrome resembling schizophrenia more completely than amphetamine alone, including negative symptoms and cognitive disorganization in addition to positive symptoms.4 The glutamate-dopamine interaction model proposes that NMDA receptor hypofunction on GABAergic interneurons in the prefrontal cortex disinhibits mesocortical and mesolimbic dopamine neurons, producing both the hyperdopaminergia of the mesolimbic system and the hypodopaminergia of the prefrontal cortex through a single upstream mechanism. This framework helps explain why purely dopaminergic interventions remain incomplete treatments, and why glutamate-modulating agents continue to be explored as adjuncts.4
Five dopamine receptor subtypes (D1 through D5) have been identified, all belonging to the G protein-coupled receptor (GPCR) superfamily. They are grouped into two families based on their primary signaling mechanism: the D1-like family (D1 and D5), which couples to Gs proteins and stimulates adenylyl cyclase to increase intracellular cyclic adenosine monophosphate (cAMP); and the D2-like family (D2, D3, and D4), which couples to Gi/Go proteins and inhibits adenylyl cyclase, reducing cAMP.1 This functional dichotomy is pharmacologically consequential because the two families have opposing effects on neuronal excitability and synaptic plasticity in the circuits relevant to psychosis, reward, and motor control.
D1 receptors are the most abundantly expressed dopamine receptors in the brain, with highest density in the striatum (caudate nucleus and putamen), nucleus accumbens, and prefrontal cortex (PFC). In the PFC, D1 receptor activation plays a central role in working memory and executive function by modulating the persistent firing of layer V pyramidal neurons that underlies working memory maintenance.5 The relationship between D1 activation and PFC cognitive function follows an inverted-U dose-response curve: at low levels of DA input, D1 stimulation is insufficient to sustain the persistent neuronal firing that underlies working memory; at moderate levels, D1 stimulation is optimal and working memory performance peaks; at high levels, excessive D1 activation suppresses neuronal firing through inhibitory interneurons and degrades the same cognitive processes it supports at moderate concentrations. This inverted-U relationship explains why both the hypodopaminergia of schizophrenia and the high-dose stimulant-induced DA excess produce cognitive deficits by falling on opposite sides of the optimum. D5 receptors have a distribution broadly overlapping D1, with expression in the hippocampus and hypothalamus, but their specific clinical role in psychosis pharmacology is less clearly delineated.1
The D2 receptor is the primary pharmacological target of all currently approved antipsychotics. It is expressed at high density in the striatum, nucleus accumbens, substantia nigra, ventral tegmental area (VTA), and pituitary lactotrophs, with lower but functionally significant expression in the PFC.1,2 D2 receptors exist in two affinity states regulated by G protein coupling (high-affinity D2High and low-affinity D2Low), and the proportion in the high-affinity state may be dysregulated in schizophrenia. D2 receptors also function presynaptically as autoreceptors on dopaminergic neurons, where their activation inhibits DA synthesis and release through a negative feedback loop. The clinical relevance of presynaptic D2 autoreceptor blockade is that antagonists at these autoreceptors increase DA release acutely, partially countering their own postsynaptic blocking effect, a phenomenon that complicates the initial clinical response and contributes to the delay between receptor occupancy and full antipsychotic effect.2
PET occupancy studies established that antipsychotic clinical response correlates with striatal D2 occupancy in the range of 65 to 80%, and that occupancy above approximately 80% is associated with a steep rise in extrapyramidal side effects (EPS) without additional antipsychotic benefit.11 This therapeutic window for D2 occupancy is fundamental to understanding why dose escalation beyond the clinically recommended range produces more adverse effects without improving efficacy, and why maintaining occupancy below the EPS threshold while remaining above the efficacy threshold is the primary challenge in antipsychotic dose optimization.
D3 receptors are concentrated in limbic regions, particularly the nucleus accumbens shell and olfactory tubercle, with relatively limited striatal expression compared to D2.1 Their limbic predominance makes them theoretically attractive targets for agents aiming to modulate reward and motivational circuitry with reduced EPS risk. Cariprazine, a partial agonist with preferential affinity for D3 over D2, exploits this distribution to address negative symptoms and motivational deficits, and its clinical profile in negative symptom management is discussed in detail in CNS-Antipsy-04. D4 receptors are expressed at relatively low density overall, with concentration in frontal cortex, amygdala, hippocampus, and midbrain. Clozapine has substantially higher D4 affinity relative to D2 compared to most other antipsychotics, a property originally proposed to contribute to its superior efficacy and lower EPS burden, though the precise clinical significance of D4 blockade remains an area of ongoing investigation.1
The anatomical specificity of DA neurotransmission, organized into four distinct projection systems each originating in midbrain nuclei, is the key to understanding both the therapeutic effects and the adverse effect profile of antipsychotic agents. Blockade of D2 receptors is not a single event with a single consequence: it simultaneously affects all four pathways to varying degrees, with divergent clinical implications for each. Rational antipsychotic pharmacology requires mapping the expected effects of a given agent across all four circuits.1,3
The mesolimbic pathway originates in the VTA of the midbrain and projects to limbic structures including the nucleus accumbens, amygdala, hippocampus, and related areas of the temporal and cingulate cortex. This pathway subserves reward processing, motivational salience, and emotional memory.3 In schizophrenia, mesolimbic DA hyperactivity is thought to produce aberrant assignment of salience to neutral stimuli, a process that is experienced subjectively as heightened significance, reference, or persecution, and which is expressed clinically as delusions and hallucinations. Blockade of D2 receptors in the mesolimbic pathway is the primary mechanism by which antipsychotics suppress positive symptoms. The fact that mesolimbic D2 blockade is the therapeutic target makes the development of pathway-selective agents desirable but pharmacologically difficult, as no currently available agent achieves truly selective mesolimbic blockade.
The mesocortical pathway also originates in the VTA but projects primarily to the PFC, as well as to other association cortex regions including the anterior cingulate and the orbitofrontal cortex. This pathway regulates working memory, executive function, attention, and the planning and initiation of goal-directed behavior.3,5 In schizophrenia, the mesocortical pathway is characterized by DA hypoactivity, reflecting reduced D1 receptor stimulation in the PFC. The clinical manifestations of this hypodopaminergia constitute the negative symptom and cognitive deficit dimensions of schizophrenia: blunted affect, alogia (poverty of speech), avolition (reduced motivation and goal-directed behavior), anhedonia (diminished capacity for pleasure), and deficits in working memory and attention. Because these symptoms arise from insufficient rather than excess DA activity in this pathway, D2 blockade by antipsychotics does not directly address them and may worsen the functional state of the mesocortical system. This is a critical pharmacological reason why negative symptoms and cognitive deficits often respond poorly to antipsychotic treatment, particularly with FGAs, and why they represent the dimension of schizophrenia most responsible for long-term functional disability.
The nigrostriatal pathway originates in the substantia nigra pars compacta and projects heavily to the striatum, comprising the caudate nucleus and putamen (the dorsal striatum). This is the pathway whose degeneration produces the motor deficits of Parkinson disease (PD). In the context of antipsychotic pharmacology, D2 blockade in the nigrostriatal pathway mimics the functional state of DA deficiency and produces EPS: acute dystonia, akathisia, drug-induced parkinsonism, and, with prolonged exposure, tardive dyskinesia (TD).1 The nigrostriatal pathway is the anatomical basis for the inverse relationship between antipsychotic EPS risk and D2 receptor affinity relative to serotonin 2A (5-HT2A) blockade that forms the core rationale for the SGA class. EPS produced by nigrostriatal blockade are dose-dependent and are observed across all antipsychotics capable of achieving meaningful D2 occupancy, though the threshold for their appearance varies considerably by agent and is influenced by receptor binding kinetics, as discussed in Section 5.
The tuberoinfundibular pathway originates in the hypothalamic arcuate nucleus and projects to the median eminence of the hypothalamus, where released DA enters the hypothalamo-hypophyseal portal circulation and reaches the anterior pituitary. DA acting at D2 receptors on pituitary lactotroph cells tonically inhibits prolactin secretion.1 Blockade of these D2 receptors by antipsychotics removes this tonic inhibition, producing hyperprolactinemia. The clinical consequences of antipsychotic-induced hyperprolactinemia include galactorrhea, amenorrhea and menstrual irregularity, sexual dysfunction, gynecomastia in men, and with sustained elevation, reduced bone mineral density. Among the four pathways, the tuberoinfundibular system does not exhibit tolerance to D2 blockade in the way that the nigrostriatal system partially does over time, which is why hyperprolactinemia tends to persist as a chronic adverse effect of prolactin-elevating antipsychotics rather than attenuating with continued treatment. Agents with relatively lower affinity for pituitary D2 receptors, or those that preferentially spare this pathway due to their additional receptor actions, produce less prolactin elevation, a clinically meaningful distinction when prescribing for young patients over extended time periods.
Schizophrenia is a syndrome characterized by three overlapping but neurobiologically distinct clinical dimensions: positive symptoms, negative symptoms, and cognitive impairment. Understanding these dimensions as separate treatment targets, each with its own neurobiological substrate and differential pharmacological responsiveness, is essential for realistic treatment planning and for recognizing the limits of currently available agents.6
Positive symptoms represent distortions or excesses of normal mental function and include hallucinations (most commonly auditory), delusions, disorganized thinking, and disorganized or bizarre behavior. The term "positive" refers to the presence of abnormal experiences rather than to their clinical desirability. Positive symptoms are most directly linked to mesolimbic DA hyperactivity and are, as a group, the dimension most responsive to antipsychotic treatment.1,6 The majority of patients with schizophrenia who receive adequate antipsychotic therapy achieve substantial reduction in positive symptom burden, though complete remission is achieved in a minority. Persistent positive symptoms despite adequate D2 occupancy define treatment resistance, a clinical category that applies to approximately 20 to 30% of patients12 and for which clozapine has the strongest evidence base.7 The acute episode in a first-episode patient, such as the clinical situation motivating this discussion, typically responds well to antipsychotic treatment, with positive symptom resolution within weeks of initiating therapy.
Negative symptoms represent reductions in or absence of normal mental function and are grouped into five core domains using the BNSS (Brief Negative Symptom Scale) and similar rating instruments: blunted affect (reduced emotional expressivity), alogia (diminished verbal output and spontaneity), avolition (reduction in motivation and self-initiated activity), anhedonia (reduced experience of pleasure from activities previously found rewarding), and asociality (diminished interest in social interaction).6 These domains are clinically heterogeneous; some patients present with predominantly affective flattening and alogia, others with severe avolition and asociality.
Negative symptoms are further classified as primary or secondary. Primary negative symptoms are intrinsic to the illness itself, reflecting the mesocortical DA deficit. Secondary negative symptoms arise from other identifiable causes: EPS-related akinesia mimicking avolition and blunted affect; depressive symptoms; positive symptom burden producing social withdrawal; or sedation from medication. This distinction is clinically consequential because secondary negative symptoms are potentially remediable by addressing their cause, while primary negative symptoms require pharmacological strategies targeting the underlying mesocortical deficit.
Primary negative symptoms represent the dimension of schizophrenia for which current pharmacotherapy is least effective and which carries the greatest burden for long-term functional outcomes. A patient who achieves full remission of positive symptoms but retains significant avolition and anhedonia may be unable to return to employment, maintain relationships, or live independently. Among currently available antipsychotics, cariprazine and lurasidone have the most robust evidence for benefit in primary negative symptoms, with cariprazine having the advantage of a prospective head-to-head trial demonstrating superiority over risperidone specifically on negative symptom scales.8 Clozapine may also reduce negative symptoms to a greater degree than other agents in treatment-resistant patients, though the evidence is confounded by its simultaneous reduction in positive symptoms and medication-induced EPS, both of which would also reduce secondary negative symptoms.
Psychosocial interventions are an essential component of negative symptom management and should be initiated as early as possible following stabilization. Supported employment (particularly the Individual Placement and Support model), social skills training, cognitive remediation therapy, and family psychoeducation have evidence bases for improving functional outcomes in patients with prominent negative symptoms, and their effects are additive to, rather than substitutable for, pharmacological treatment.9 For the patient in his post-acute stabilization phase, a structured combination of the most evidence-supported antipsychotic choice for negative symptoms alongside early psychosocial rehabilitation engagement is the most defensible clinical approach.
Cognitive deficits in schizophrenia are pervasive, affecting multiple domains including attention, working memory, processing speed, verbal learning and memory, and executive function. They are present before the first psychotic episode, worsen at illness onset, and represent a strong independent predictor of functional outcomes separate from positive and negative symptom severity.6 The MATRICS (Measurement and Treatment Research to Improve Cognition in Schizophrenia) consensus battery identified seven cognitive domains as the primary targets for treatment development in schizophrenia. No currently approved antipsychotic has demonstrated robust pro-cognitive effects beyond what is attributable to reduced positive symptom burden and sedation reduction. Anticholinergic medications used to manage EPS from FGAs further impair cognition and worsen this dimension. Cognitive remediation approaches, particularly those incorporating strategy learning and goal-oriented practice, show the most consistent evidence for improving functional cognition in schizophrenia.9
Antipsychotic agents are not selective D2 antagonists. All approved antipsychotics bind to a constellation of receptor types with varying affinity, and the clinical profile of each agent, including its therapeutic advantages, adverse effects, and optimal clinical application, is substantially determined by its full binding profile rather than its D2 affinity alone.1,2 Understanding this multi-receptor pharmacology is prerequisite to rational agent selection and adverse effect anticipation.
5-HT2A receptors are expressed at high density in the prefrontal cortex, where serotonin tonically suppresses DA release via inhibitory interneurons. Blockade of cortical 5-HT2A receptors by antipsychotics disinhibits DA release in the PFC, partially restoring mesocortical DA activity.1 This mechanism is the primary pharmacological rationale for the SGA class and is why agents with a high 5-HT2A to D2 affinity ratio, such as quetiapine, olanzapine, and clozapine, produce fewer EPS and may have more favorable effects on negative symptoms and cognition than FGAs. In the nigrostriatal system, 5-HT2A blockade similarly disinhibits DA release, reducing the degree of effective D2 blockade in the striatum and thereby raising the EPS threshold. The practical consequence is that an SGA achieving 80% D2 occupancy in the striatum produces fewer EPS than an FGA at equivalent occupancy, because the concurrent 5-HT2A blockade partially restores nigrostriatal DA tone.
H1 receptor blockade is the primary mechanism of antipsychotic-induced sedation and contributes substantially to weight gain, probably through effects on hypothalamic appetite regulation.1 Agents with high H1 affinity, including clozapine, olanzapine, and quetiapine, produce the most pronounced sedation and are among the agents most associated with weight gain and metabolic complications. In clinical practice, H1 blockade is not inherently undesirable: sedation can be therapeutically useful in acute agitation, and the timing of dosing can be adjusted to minimize functional impairment. However, persistent sedation in the maintenance phase impairs daytime functioning, reduces treatment adherence, and compounds the cognitive deficits already present in schizophrenia.
Anticholinergic effects from muscarinic M1 receptor blockade produce the peripheral and central adverse effects familiar from this class: dry mouth, urinary retention, constipation, blurred vision, tachycardia (peripheral), and cognitive impairment and sedation (central).1 Low-potency FGAs, particularly thioridazine and chlorpromazine, and among SGAs, clozapine, carry the highest anticholinergic burden. Central M1 antagonism contributes to worsening of the cognitive deficits already present in schizophrenia and is one reason that the routine prophylactic use of anticholinergic agents such as benztropine to prevent EPS from FGAs carries a real cost in cognitive terms. When EPS management is required, anticholinergics should be used at the lowest effective dose for the shortest necessary period, with preference for switching to a lower-EPS agent when feasible.
Alpha-1 (alpha-1 adrenergic) receptor blockade produces orthostatic hypotension, reflex tachycardia, and contributes to sexual dysfunction and nasal congestion.1 Low-potency FGAs and clozapine carry significant alpha-1 blocking activity. Orthostatic hypotension is particularly consequential in elderly patients and in patients receiving concurrent antihypertensive therapy, where the additive effect can produce clinically significant falls and syncope. Dose titration is the primary management strategy; starting at low doses and increasing gradually allows partial tolerance to develop and minimizes symptomatic hypotension. QTc prolongation, mediated through cardiac hERG (human ether-a-go-go-related gene) potassium channel blockade, is a separate cardiac adverse effect associated most prominently with thioridazine, pimozide, and ziprasidone and is discussed in detail in CNS-Antipsy-05.
Antipsychotic agents are conventionally divided into first-generation antipsychotics (FGAs), also called typical or conventional antipsychotics, and second-generation antipsychotics (SGAs), also called atypical antipsychotics. This classification, while widely used, is an oversimplification that reflects a historical distinction more than a pharmacologically homogeneous categorical difference.1,2
FGAs are characterized by potent D2 receptor antagonism with relatively limited activity at other receptor types. They are further classified by potency, which in this context refers to the milligram dose required for antipsychotic effect rather than intrinsic pharmacological efficacy at the D2 receptor. High-potency FGAs, including haloperidol, fluphenazine, and trifluoperazine, achieve D2 blockade at doses of 2 to 20 mg per day and carry a high risk of EPS due to their tight D2 binding and limited offsetting receptor activity. Low-potency FGAs, including chlorpromazine and thioridazine, require doses of 200 to 1000 mg per day and produce less EPS (due in part to their concurrent anticholinergic and antihistaminergic activity) but substantially more sedation, orthostatic hypotension, and anticholinergic adverse effects.1 Mid-potency agents such as perphenazine represent an intermediate profile. All FGAs carry significant risk of TD with prolonged use, reflecting sustained nigrostriatal D2 blockade and the resulting dopamine receptor supersensitivity that develops over time.
Despite their adverse effect burden, FGAs retain clinical utility. Cost remains a practical consideration in resource-limited settings and for uninsured patients, where generic haloperidol or perphenazine represents a fraction of the cost of branded SGAs. Long-acting injectable formulations of haloperidol decanoate and fluphenazine decanoate provide reliable plasma concentrations in patients with adherence difficulties, a group representing a substantial proportion of those with chronic schizophrenia. The CATIE (Clinical Antipsychotic Trials of Intervention Effectiveness) study, which compared perphenazine directly against several SGAs in a real-world study population, found that perphenazine performed comparably to the SGAs on the primary measure of all-cause discontinuation, a result that moderated early enthusiasm for the SGAs as categorically superior to all FGAs.10
SGAs are defined operationally by a lower propensity for EPS and TD at clinically effective doses compared to FGAs, and mechanistically by their combined D2 and 5-HT2A antagonism (or, in the case of partial agonists such as aripiprazole and cariprazine, by partial D2 agonism alongside 5-HT2A antagonism). The 5-HT2A to D2 binding affinity ratio is higher in SGAs than in FGAs, and this ratio is a reasonable, though imperfect, predictor of EPS liability across agents.2 SGAs are a pharmacologically heterogeneous group: clozapine, the prototypical atypical antipsychotic, is a multi-receptor agent with relatively modest D2 affinity but high affinity for D4, D1, 5-HT2A, H1, M1, and alpha-1 receptors; aripiprazole is a partial D2 agonist with a mechanistically distinct profile from all other antipsychotics; lurasidone has a relatively clean receptor profile compared to olanzapine or quetiapine; and ziprasidone carries QTc liability that distinguishes it within the class. Treating SGAs as a uniform group for purposes of adverse effect prediction or efficacy comparison is pharmacologically unjustified.
The most clinically significant metabolic adverse effects, including weight gain, glucose dysregulation, and dyslipidemia, are concentrated in the SGA class, particularly with clozapine and olanzapine. These metabolic effects, largely absent from most FGAs, represent a major long-term safety concern and have substantially complicated the narrative of SGAs as categorically preferable to FGAs. The choice between agents within and across classes should be individualized based on the patient's symptom profile, prior treatment response, comorbidities, and specific adverse effect risks, rather than on a default preference for one class over the other.
Stahl SM. Stahl's Essential Psychopharmacology: Neuroscientific Basis and Practical Applications. 4th ed. Cambridge: Cambridge University Press; 2013:129–237.
Seeman P, Lee T, Chau-Wong M, Wong K. Antipsychotic drug doses and neuroleptic/dopamine receptors. Nature. 1976;261(5562):717–719.
doi:10.1038/261717a0Howes OD, Kapur S. The dopamine hypothesis of schizophrenia: version III — the final common pathway. Schizophr Bull. 2009;35(3):549–562.
doi:10.1093/schbul/sbp006Javitt DC, Zukin SR. Recent advances in the phencyclidine model of schizophrenia. Am J Psychiatry. 1991;148(10):1301–1308.
doi:10.1176/ajp.148.10.1301Goldman-Rakic PS. The cortical dopamine system: role in memory and cognition. Adv Pharmacol. 1998;42:707–711.
doi:10.1016/s1054-3589(08)60846-7Tandon R, Nasrallah HA, Keshavan MS. Schizophrenia, "just the facts" 4. Clinical features and conceptualization. Schizophr Res. 2009;110(1–3):1–23.
doi:10.1016/j.schres.2009.03.005Kane J, Honigfeld G, Singer J, Meltzer H. Clozapine for the treatment-resistant schizophrenic: a double-blind comparison with chlorpromazine. Arch Gen Psychiatry. 1988;45(9):789–796.
doi:10.1001/archpsyc.1988.01800330013001Nemeth G, Laszlovszky I, Czobor P, et al. Cariprazine versus risperidone monotherapy for treatment of predominant negative symptoms in patients with schizophrenia: a randomised, double-blind, controlled trial. Lancet. 2017;389(10074):1103–1113.
doi:10.1016/S0140-6736(17)30060-0Cella M, Roberts S, Pillny M, et al. Psychosocial and behavioural interventions for the negative symptoms of schizophrenia: a systematic review of efficacy meta-analyses. Br J Psychiatry. 2023;223(1):8–17.
doi:10.1192/bjp.2023.21Lieberman JA, Stroup TS, McEvoy JP, et al. Effectiveness of antipsychotic drugs in patients with chronic schizophrenia. N Engl J Med. 2005;353(12):1209–1223.
doi:10.1056/NEJMoa051688Kapur S, Zipursky R, Jones C, Remington G, Houle S. Relationship between dopamine D2 occupancy, clinical response, and side effects: a double-blind PET study of first-episode schizophrenia. Am J Psychiatry. 2000;157(4):514–520.
doi:10.1176/appi.ajp.157.4.514Smart SE, Kepinska AP, Murray RM, MacCabe JH. Predictors of treatment resistant schizophrenia: a systematic review of prospective observational studies. Psychol Med. 2021;51(1):44–53.
doi:10.1017/S0033291719002083