Induced pluripotent stem cell-derived neuronal models in psychiatry: Mechanistic insights and translational horizons

Abstract

Human induced pluripotent stem cell (iPSC)-derived neuronal models have become an important platform for investigating psychiatric disorders by enabling experimental access to human neurons carrying patient-specific genetic risk. Advances in reprogramming, neural differentiation, single-cell profiling, electrophysiology, and genome editing have facilitated the accessibility of knowledge which was once unavailable. Psychiatric disorders pose particular challenges for in vitro modeling due to their polygenic architecture, developmental origins, and clinical heterogeneity. This review summarizes the current state of iPSC-derived neuronal technologies, summarizes key findings across major psychiatric disorders, and discusses methodological limitations and emerging translational directions. We emphasize quantitative functional phenotyping and patient stratification as central principles for future progress in precision psychiatry. The future of precision psychiatry relies on two central principles; analyzing patient specific neuronal characteristics with an objective, comparable readout, and classifying patients into biologically meaningful subgroups which share common features such as mechanism, biomarker profile, or functional and structural signature.

Key Words: Induced pluripotent stem cells; Schizophrenia; Bipolar disorder; Major depressive disorder; Autism spectrum disorder; Attention-deficit hyperactivity disorder; Precision psychiatry; In vitro disease modeling

Core Tip: Human induced pluripotent stem cell-derived neuronal models are transforming psychiatric research by enabling direct investigation of patient-specific neural phenotypes in a human genetic context. Beyond disease modeling, emerging evidence suggests that quantitative functional phenotyping can identify biologically meaningful patient subgroups and treatment-response signatures. This review synthesizes findings across major psychiatric disorders and proposes a conceptual shift from reproducing diagnostic categories in vitro toward mechanism-based stratification. Such approaches may bridge the gap between psychiatric genetics and clinical decision-making, advancing the development of precision psychiatry.

INTRODUCTION

Psychiatric disorders are leading contributors to global disease burden, yet their underlying biological mechanisms remain incompletely understood. Large-scale genetic studies demonstrate extensive polygenicity across disorders including schizophrenia, major depressive disorder (MDD), bipolar disorder (BD), autism spectrum disorder (ASD), and attention-deficit/hyperactivity disorder (ADHD), with many risk variants affecting gene regulation rather than protein-coding sequence[1-3]. Access to living human brain tissue is extremely limited, and animal models often fail to capture human-specific aspects of neurodevelopment and gene regulation. Induced pluripotent stem cell (iPSC) technology and in vitro neural differentiation enable the generation of patient-specific human neurons, providing a complementary experimental system for postmortem studies and animal models. iPSC-derived neurons allow longitudinal interrogation of neurodevelopment, synaptic maturation, and functional responses to pharmacological perturbation, while preserving human genetic architecture increasingly implicated in psychiatric disease.

IPSC-DERIVED NEURONAL TECHNOLOGIES

iPSC-derived neuronal technologies have become a central platform for human disease modeling as they retain donor-specific genetic architecture while enabling scalable generation of human neural cell states that are otherwise difficult to access experimentally[4]. The application of these approaches across major psychiatric disorders is summarized in Table 1. Typically, many workflows begin with directed neural induction that recapitulates core steps of neuroectoderm specification, most commonly through dual SMAD inhibition to efficiently generate PAX6+/SOX1+ neural progenitor cells (NPCs) that can be expanded and differentiated[5]. From this progenitor stage, subtype specification can be guided using developmentally directed patterning to enrich for forebrain-relevant excitatory and inhibitory neuronal identities, supporting disease-relevant comparisons at defined developmental stages[6]. Because neuronal identity alone is insufficient to establish model fidelity, contemporary pipelines typically incorporate orthogonal quality control across cell state; marker and transcriptomic profiling, increasingly at single-cell resolution and function such as patch clamp, calcium imaging, and multielectrode array-based network phenotyping to benchmark excitability, synaptogenesis, and emergent circuit activity[7,8]. As no single platform captures all aspects of neuronal dysfunction, the choice of phenotyping approach should be guided by the biological question and the intended translational application. The major strengths and limitations of commonly used platforms are summarized in Table 2. In parallel, 3D organoid and assembloid systems extend 2D differentiation by enabling multicellular organization and inter-regional interactions; at the same time, they underscore the importance of managing batch effects and reproducibility through standardized workflows and molecular benchmarking[9-11]. Importantly, 3D complexity should not be treated as intrinsically superior to 2D neuronal cultures. Organoids and assembloids are most informative when the phenotype under study depends on multicellular organization, developmental patterning, migration, glial-neuronal interactions, or inter-regional connectivity. In contrast, for donor stratification based on cell-autonomous excitability, synaptic density, bioenergetics, or acute pharmacological response, simpler 2D or induced-neuron systems may provide cleaner and more scalable readouts. In 3D systems, apparent disease effects can be confounded by differences in organoid size, maturation, regional identity, cellular composition, necrotic core formation, and batch-to-batch variability. Therefore, stratification-oriented organoid studies should incorporate multiple organoids per donor, independent differentiations, randomized batch design, predefined morphological and molecular QC thresholds, and single-cell or spatial transcriptomic benchmarking against reference developmental datasets. Under these conditions, 3D systems can complement 2D assays by testing whether cell-autonomous phenotypes persist, amplify, or reorganize within multicellular circuit-like contexts. A complementary and increasingly widely adopted strategy bypasses these staged developmental trajectories by using transcription factor-based “forward programming” rather than prolonged progenitor-mediated differentiation. Whereas directed differentiation models neurodevelopmental progression through intermediate neural progenitor states, rapid transcription factor-driven reprogramming (notably NGN2) compresses fate acquisition into a short, more standardized induction window, often yielding relatively synchronous neuronal populations that are well suited to higher-throughput experimental designs[12,13]. This distinction is especially relevant when the primary objective is scalable generation of functionally mature neurons for quantitative phenotyping (e.g., electrophysiology and network assays), systematic genetic perturbation, or screening, rather than interrogation of early lineage decisions supporting the use of induced-neuron paradigms alongside conventional iPSC-to-neuron differentiation frameworks.

Table 1 Summary of induced pluripotent stem cell-derived neuronal findings in psychiatry.

Disorder	Primary model	Key reported phenotypes	Ref.
Schizophrenia	Forebrain/cortical neurons; NPC → neurons; 22q11.2DS neurons; single-cell RNA-seq	Neurite/synaptic and connectivity deficits; catecholamine-release abnormalities; WNT/progenitor dysregulation; 22q11.2DS shows mitochondrial and developmental transcriptomic defects; single-cell heterogeneity shifts	[15-22]
Bipolar disorder	Neurons (electrophysiology); NPCs; cortical spheroids	Hyperexcitability with lithium-linked rescue; functional subgrouping; NPC calcium-entry and differentiation changes; lithium effects in spheroids; mitochondrial/bioenergetic abnormalities	[23-28]
Autism spectrum disorder	Neurons (syndromic/idiopathic); CNV/risk-gene cohorts; neuron-astrocyte co-culture; MEA/patch; iPSCs (NO assay)	Synapse/network phenotypes; 16p11.2 dosage effects; CNTN5/EHMT2 network hyperactivity; convergent expression modules; astrocyte modulation; nitric-oxide signature in iPSCs	[29-38]
Major depressive disorder	Serotonergic/forebrain neurons; cortical neurons; NPC → neurons; GABA interneurons; dopaminergic neurons	SSRI-resistance phenotypes including serotonin-evoked hyperactivity; antidepressant-response signatures; bioenergetic/electrophysiology changes; HTR2C-linked interneuron activity with rescue; ketamine plasticity pathways	[1,39-45]
ADHD	NSCs/NPCs; telencephalon organoids; cortical neurons; dopaminergic neurons (PARK2 CNVs)	Progenitor proliferation deficits; early cortical organization changes in organoids; glutamatergic and calcium-network hypoactivity; stress-modulated bioenergetic deficits	[47-52]
Anorexia nervosa	Neuronal cultures	Limited evidence; transcriptomic dysregulation signals; major gap in functional phenotyping	[53]
OCD	iPSC line resources	Mostly line derivation/validation; limited neuronal phenotyping	[54]

NPC: Neural progenitor cell; RNA-seq: RNA sequencing; CNV: Copy number variation; MEA: Multielectrode array; iPSCs: Induced pluripotent stem cells; NO: Nitric oxide; GABA: Gamma-aminobutyric acid; ADHD: Attention-deficit/hyperactivity disorder; SSRI: Selective serotonin reuptake inhibitor; NSCs: Neural stem cells; OCD: Obsessive-compulsive disorder.

Table 2 Functional phenotyping platforms for induced pluripotent stem cell-derived neuronal models in psychiatry.

Platform	Primary readout	Strengths	Limitations	Best suited for
Patch clamp	Single-cell electrophysiology (action potentials, synaptic currents, ion-channel function)	Highest functional resolution; direct measurement of neuronal physiology	Low throughput; technically demanding	Mechanistic studies and validation of ion-channel or synaptic phenotypes
MEA	Network activity, burst dynamics, synchrony	Non-invasive, longitudinal recordings; suitable for moderate-throughput screening	Limited single-cell resolution; influenced by culture density and maturity	Drug screening, network phenotyping, subgroup identification
Calcium imaging	Single-cell and network activity dynamics	Higher throughput than patch clamp; enables population-level analysis	Indirect measure of electrical activity; lower temporal resolution	Functional comparison between donor groups and pharmacological studies
High-content imaging	Neurite morphology, synaptic markers, cell survival	Scalable and highly multiplexed	Primarily structural rather than functional	Morphological phenotyping and toxicity/rescue assays
Single-cell/spatial transcriptomics	Cell identity, pathway activity, cellular composition	High molecular resolution; identifies disease-associated cell states	Destructive, expensive, not a direct functional assay	Molecular characterization, organoid quality control, pathway discovery

MEA: Multielectrode array.

IPSC-DERIVED NEURONAL MODELS IN PSYCHIATRIC DISORDERS

Schizophrenia

Schizophrenia is a highly heritable, clinically heterogeneous psychiatric disorder in which risk is distributed across many common variants of small effect, alongside rarer alleles, consistent with a markedly polygenic architecture[3]. Genetic discovery efforts have now implicated large sets of associated loci and convergent biological themes, reinforcing the view that schizophrenia reflects distributed perturbations of neurobiological systems rather than a single pathogenic pathway[2]. In parallel, converging evidence supports a neurodevelopmental framework, in which altered synaptic development and refinement (including excitation-inhibition balance) interacts with later maturational processes and dopaminergic dysfunction to produce the clinical syndrome[14].

Schizophrenia is the most extensively studied psychiatric disorder using iPSC-derived neuronal models. Brennand et al[15] first reported reduced neurite outgrowth, altered synaptic protein expression, and impaired connectivity in schizophrenia patient-derived neurons. Additionally, schizophrenia iPSC-derived neurons showed elevated catecholamine secretion and a higher fraction of tyrosine hydroxylase-positive neurons relative to controls[16]. Subsequent studies identified dysregulated WNT signaling, altered neural progenitor proliferation, and reduced synaptic maturation and excitatory transmission[17-19]. iPSC-derived neurons from 22q11.2 deletion syndrome patients with schizophrenia showed bioenergetic impairment, including reduced ATP and decreased oxidative phosphorylation activity, reduced mitochondrial DNA (mtDNA)-encoded protein products as well as developmentally patterned transcriptomic dysregulation organized into co-expression networks affecting cell-cycle survival and signaling pathways[20,21]. Single-cell transcriptomic analyses further revealed increased heterogeneity within schizophrenia-derived neuronal cultures, with shifts in developmental state distributions rather than uniform disease-specific signatures[22]. Collectively, these findings support a model in which schizophrenia risk manifests as altered neurodevelopmental trajectories.

BD

BD is among the most heritable major psychiatric illnesses. At the genetic level, BD shows a strongly polygenic architecture: Large GWAS implicate dozens of loci, with risk alleles enriched in brain-expressed and synaptic signaling genes, particularly in neuronal populations relevant to cortical and hippocampal circuitry[23]. This polygenicity together with substantial clinical heterogeneity (e.g., BD I vs BD II, episode polarity, age of onset) and confounding by treatment exposure creates a core challenge for mechanism: Disease liability is distributed across many variants of small effect, making it difficult to isolate causal pathways without human, genotype-preserving experimental systems.

In this context, iPSC-derived neuronal models are well suited to BD because they retain each donor’s genomic background and enable standardized differentiation into disease-relevant neural states, allowing cellular phenotypes to be quantified under controlled conditions. A seminal study demonstrated that neurons derived from BD patients exhibit intrinsic hyperexcitability and, critically, that this phenotype can be selectively normalized by lithium in neurons derived from clinically lithium-responsive individuals providing one of the clearest examples of an iPSC-neuron assay tracking a clinically meaningful treatment predictor[24]. A follow-up analysis reported that BD patient-derived neurons can separate into subgroups with distinct electrophysiological properties, consistent with the idea that iPSC-neuron readouts may delineate biologically meaningful heterogeneity within BD beyond DSM-based diagnostic[25].

More recent work has expanded BD modeling upstream to earlier developmental stages and into 3D systems to capture circuit-level and treatment-duration effects. In NPCs, BD-derived lines were reported to show dysregulated store-operated Ca²⁺ entry alongside accelerated differentiation, aligning with longstanding evidence that calcium signaling is a recurrent theme in BD genetics[26]. In parallel, iPSC-derived cortical spheroids have been used to probe lithium mechanisms in a networked context, identifying transcriptional and functional effects of prolonged lithium exposure including effects consistent with modulation of excitability and neuroprotective programs while emphasizing that treatment duration can materially alter observed cellular outcomes[27].

Additionally, iPSC-derived BD neurons revealed evidence for mitochondrial and bioenergetic abnormalities including altered oxidative phosphorylation, high-energy phosphate measures, mitochondrial morphology/mtDNA findings, and mitochondrial gene/protein expression changes with downstream links to oxidative stress/inflammation and potential therapeutic leverage via mitochondria-targeted interventions[28].

ASD

ASD is a genetically complex neurodevelopmental condition in which liability reflects contributions from both rare, often large-effect variants [including de novo disruptive mutations and pathogenic copy number variations (CNVs)] and a substantial burden of common polygenic variation, with convergent enrichment in pathways related to chromatin/gene regulation and synaptic/neurotransmission biology[29-31].

Early work in monogenic ASD-related syndromes provided proof-of-principle that patient-derived neurons can reveal disease-relevant cellular defects and pharmacologically tractable phenotypes e.g., Rett syndrome iPSC-derived neurons showing abnormal neuronal maturation and function in vitro[32]. More broadly, the field has emphasized synaptic development as a central mechanistic hub, including work explicitly framing iPSC approaches for interrogating synaptogenesis-related phenotypes relevant to ASD[33].

More recent ASD iPSC studies increasingly leverage genetically defined cohorts especially CNVs and high-confidence risk genes to reduce heterogeneity and strengthen causal inference. A prominent example is 16p11.2 copy-number variation: IPSC-derived neurons from deletion vs duplication carriers show contrasting cellular phenotypes consistent with dosage-sensitive neurodevelopmental mechanisms[34]. At the level of neuronal function, iPSC-derived neurons from individuals carrying ASD-associated haploinsufficiency in CNTN5 or EHMT2 were reported to develop hyperactive neuronal networks, linking specific ASD risk genotypes to measurable circuit-level phenotypes (multielectrode array/patch clamp)[35]. Transcriptome-centric studies further suggest that, even in idiopathic/normocephalic ASD cohorts, iPSC-derived neuronal cells can reveal co-expression modules consistently associated with ASD, supporting the use of these systems to identify convergent molecular programs across genetically diverse individuals[36].

An influential study using iPSC-derived neurons and astrocytes from non-syndromic ASD donors examined how astrocytes modulate neuronal connectivity and network properties, reinforcing glial contributions as a mechanistic layer that can be directly tested in human patient-derived cultures[37].

Finally, a more explicitly translational direction couples ASD-relevant cellular abnormalities to quantifiable biomarker-style readouts, exemplified by a recent study using a nanosensor-based electroanalytical assay to measure nitric oxide (NO)-associated signals in patient-iPSCs, identifying a discriminable NO signature between an autism sample and comparator lines and highlighting NO-linked redox/inflammatory biology as a candidate measurable phenotype with potential diagnostic or stratification relevance[38].

MDD

MDD is a complex disorder in which large meta-analyses have identified dozens to > 100 independent associated variants and highlighted enrichment in brain-relevant annotations, reinforcing a model in which many variants of small effect collectively shift risk through distributed molecular pathways rather than single-gene causality[1,39].

A particularly influential line of work has leveraged clinically annotated cohorts stratified by antidepressant response to interrogate serotonergic mechanisms: IPSC-derived serotonergic circuitry of forebrain neuron paradigms have revealed cellular phenotypes linked to selective serotonin reuptake inhibitor (SSRI) resistance, including neurite and gene-expression differences consistent with altered serotonergic signaling[40] and a serotonin-induced hyperactivity phenotype in neurons from SSRI non-remitters driven by upregulated excitatory serotonergic receptor signaling relative to remitters and controls[41].

In addition to SSRI responsiveness, iPSC platforms are increasingly used to discover treatment-response-associated cellular signatures and candidate biomarkers. In a translational psychiatry study, patient-derived cortical neurons generated from clinically characterized donors were used to probe antidepressant response biology, identifying response-associated differences spanning neuronal morphology and synaptic connectivity and transcriptional changes[42]. Complementing these pharmacology-linked readouts, induced NPCs and iPSC-derived neurons from MDD patients have been reported to exhibit altered bioenergetics together with measurable electrophysiological differences, supporting a mechanistic link between mitochondrial energetic state and neuronal function in at least a subset of patients[43].

The literature also extends into cell-type-specific and circuit-relevant models that target mechanistic hypotheses beyond excitatory cortical neurons. In an EMBO Molecular Medicine study focused on severe MDD with suicidal behavior, iPSC-derived GABAergic interneurons displayed altered morphology, increased firing, and weakened calcium signal propagation, with transcriptomic analyses implicating reduced HTR2C (5-HT2C) expression; pharmacologic or genetic targeting of 5-HT2C was reported to restore the neuronal activity phenotype, providing a concrete example of a cell-type-specific, human model supporting target validation[44]. Finally, iPSC-derived dopaminergic neurons have been used to probe mechanisms relevant to rapid-acting antidepressants: Ketamine was shown to enhance structural plasticity through AMPA receptor-driven brain-derived neurotrophic factor and mammalian target of rapamycin signaling in human iPSC-derived dopamine neurons, linking a clinically relevant compound class to conserved plasticity pathways in a human neuronal system[45]. Recent studies suggest that biologically distinct MDD subtypes can be identified based on functional connectivity and transcriptomic profiles[46]. Integrating these patient subgroups with iPSC-derived neuronal phenotypes may help determine whether imaging-defined subtypes converge on distinct cellular mechanisms, thereby improving patient stratification and precision psychiatry.

ADHD

ADHD is a highly heritable neurodevelopmental condition whose genetic ground is dominated by common-variant polygenicity, with GWAS meta-analyses identifying risk loci enriched in brain-expressed regulatory annotations and implicating neurodevelopmentally expressed genes and neuronal cell types. Early large case-control GWAS identified the first genome-wide significant loci for ADHD, supporting a distributed architecture rather than single-gene causality[47]. More recent, larger meta-analyses expanded these findings (e.g., 27 loci; prioritized risk genes enriched in early brain development and associations with specific neuronal subtypes including dopaminergic neurons), sharpening hypotheses around cell-type- and developmental-stage specificity[48].

A pilot study comparing iPSC- and neural stem cell (NSC)-derived lines from male children and adolescents with ADHD as well as genetically predisposed individuals defined via polygenic risk scores, reported reduced proliferation at the NSC stage in the ADHD group, with no detectable differences at the iPSC stage - consistent with the idea that measurable phenotypes may emerge during lineage commitment rather than pluripotency[49]. Complementing 2D precursor assays, telencephalon organoids derived from an ADHD individual showed altered early cortical layer-like organization, including reduced growth of layer structures relative to controls, supporting the feasibility of probing early corticogenesis-relevant phenotypes in 3D[50].

McNeill et al[51] (2026) established iPSC-derived cortical neurons from adult ADHD patients and controls and reported impaired glutamatergic development along with a hypoactive synaptic signaling phenotype by calcium imaging; reduced signaling strength and frequency, pointing at glutamatergic pathways as potential mechanistic and therapeutic targets. Finally, beyond neurodevelopmental and synaptic phenotypes, ADHD patient-derived iPSC dopaminergic neurons carrying PARK2 CNVs exhibit bioenergetic dysfunction; ATP and basal respiration deficits, modulated by cellular stress, supporting mitochondrial energy metabolism as a mechanistically testable axis in genetically defined ADHD subgroups[52].

GAPS AND OPPORTUNITIES: PSYCHIATRIC DISORDERS WITH LIMITED IPSC EVIDENCE

Several psychiatric disorders remain underrepresented in iPSC-based disease modeling, largely due to challenges in assembling sufficiently powered, clinically deep cohorts and the need to align differentiation paradigms with the most relevant developmental windows and cell types. This gap is well illustrated by anorexia nervosa (AN) and obsessive-compulsive disorder (OCD), where the iPSC literature is comparatively sparse and still dominated by early-stage efforts. For AN, one of the few primary mechanistic studies used patient-derived iPSC neuronal cultures to identify disease-associated transcriptional dysregulation and nominate candidate pathways for follow-up, and recent field syntheses explicitly emphasize that only a few iPSC-derived AN models exist to date[53]. For OCD, published iPSC work has frequently taken the form of resource generation (e.g., establishment and validation of patient-specific iPSC lines such as TUSMi004-A) rather than systematic neuronal phenotyping, underscoring how much mechanistic space remains to be explored using disease-relevant neural differentiation and functional assays[54].

TRANSLATIONAL AND PRECISION PSYCHIATRY IMPLICATIONS

Across psychiatric disorders, convergent cellular themes include altered neurodevelopmental timing, synaptic maturation deficits, and changes in intrinsic excitability. These insights have motivated efforts to use patient-derived neuronal assays as functional models linking genetic variation to disease mechanisms and treatment response. Within this context, Laifenfeld et al[4] reviewed blood-derived neuronal models as a potential precision psychiatry approach, framing such systems as complementary to clinical assessment and genetic information rather than diagnostic replacements.

METHODOLOGICAL CONSIDERATIONS AND LIMITATIONS

Key limitations of psychiatric iPSC models include limited cohort sizes (often underpowered for polygenic disorders), small effect sizes, and polygenic heterogeneity, where diverse combinations of common and rare variants can yield partially convergent or even divergent cellular phenotypes across patients. Additional constraints include developmental state dependence; phenotypes may only emerge in specific maturation windows or cell types, loss of age- and exposure-associated signatures during reprogramming, and batch effects arising from line-to-line variability, differentiation efficiency, and culture conditions.

Furthermore, cellular reprogramming may erase age- and exposure-associated molecular signatures. Consequently, iPSC-derived neuronal models primarily reflect genetic susceptibility rather than the cumulative effects of aging and environmental exposures. Model fidelity is further challenged by incomplete representation of circuit- and systems-level complexity, including long-range connectivity, neuromodulatory inputs, myelination, vascular and immune components, and physiologic endocrine-metabolic context, as well as limited capture of gene-environment interactions (stress hormones, inflammation, medications, DNA damage) that are central to psychiatric pathophysiology. Best practices emphasize isogenic controls, multimodal phenotyping, replication across independent lines and differentiations, transparent reporting of QC metrics, and cautious interpretation of disease relevance and translational claims.

CONCLUSION

iPSC-derived neuronal models have matured into a rigorous experimental platform for psychiatric research, enabling direct interrogation of human neurodevelopmental trajectories, synaptic and circuit-relevant dysfunction, and inter-individual variability in a genotype-preserving context. Although these systems do not recapitulate the full clinical, environmental, and systems-level complexity of psychiatric illness, they offer a uniquely tractable bridge between psychiatric genetics and mechanism by allowing convergent biological programs such as excitability-inhibition balance, synaptogenesis and connectivity, transcriptional regulation and epigenetic modulation, and cellular bioenergetics to be measured quantitatively in various defined human neural cell states (Figure 1).

Figure 1 Patient somatic cells are reprogrammed into induced pluripotent stem cells and differentiated into neuronal subtypes relevant to psychiatric disease. Quantitative molecular and functional phenotyping is used to identify convergent cellular mechanisms and individual-specific drug responses, supporting translational and precision psychiatry applications. iPSCs: Induced pluripotent stem cells.

A key conceptual shift emerging across disorders is the move from categorical “disease-in-a-dish” ambitions toward stratified, mechanism-first modeling. Stratification can be based on complementary biological features rather than diagnostic categories alone, including genetic architecture (e.g., polygenic risk or rare variants), quantitative functional phenotypes, or clinically defined treatment response groups. In highly polygenic disorders, multi-donor cohorts are required to identify convergent biological mechanisms across diverse genetic backgrounds, whereas isogenic controls are primarily suited to dissect the effects of individual variants. Together, these complementary approaches improve the identification of pathway level mechanisms beyond conventional case control comparisons.

In highly polygenic disorders, common-variant liability acts alongside rare variants in an additive manner, motivating experimental designs that preserve donor genotype while enabling mechanistic deconvolution of convergent pathways and subgroup-aware analyses. In practice, the strongest translational impact is likely to arise from quantitative functional phenotyping electrophysiology, calcium/network assays, synaptic measures integrated with molecular profiling, rather than from attempts to reproduce full diagnostic syndromes in vitro. This framework aligns particularly well with precision psychiatry goals: IPSC-derived assays can nominate actionable axes for example, excitability and calcium signaling programs, support target validation, and provide stratified experimental readouts that map onto treatment-response biology.

The field’s next phase will be defined by scaling and integration: Larger and genetically stratified cohorts, systematic incorporation of glia and multicellular formats (organoids/assembloids), and standardized perturbation paradigms that model hormonal, metabolic, and pharmacologic influences. These priorities are especially important for disorders with sparse iPSC evidence such as AN where iPSC approaches are positioned to illuminate mechanisms that are difficult to resolve in animal models or peripheral tissues, including reward-feeding circuitry-relevant programs, cell-type-specific gene regulation, and epigenetic modulation under controlled experimental stressors. Overall, the most durable contribution of psychiatric iPSC models will be to convert genetic risk into testable cellular hypotheses and scalable functional readouts that support target discovery, patient stratification, and individualized therapeutic prediction.