NMDA receptor–dependent structural plasticity links early life stress to persistent aggression Authors: Jessica T. Jacobs, Jacob C. Nordman Published: 2026 ABSTRACT Early life stress (ELS) is a major risk factor for reactive and pathological aggression that can persist long after the original stress exposure. Across species, ELS produces enduring alterations in the neural circuits that regulate social behavior and aggression, yet the molecular, structural, and circuit mechanisms that stabilize these changes remain poorly defined. Emerging evidence implicates N-methyl-D-aspartate receptor (NMDAR)-dependent structural plasticity in aggression circuits as a mechanism linking early adversity to persistent aggressive behavior.In this review, we present a hypothesis-driven narrative that synthesizes clinical, postmortem, imaging, and preclinical evidence to define a testable framework for future research. We focus on how ELS may recruit NMDAR signaling to reshape excitatory connectivity within aggression circuits, with particular emphasis on the posterior ventral medial amygdala to ventrolateral ventromedial hypothalamus pathway. Within this pathway, preclinical studies suggest that ELS increases dendritic spine density, promotes excitatory shaft synaptogenesis, and biases synaptic input toward proximal dendritic locations predicted to exert stronger effects on neuronal firing. We propose that this location-specific remodeling may increase circuit excitability, thereby lowering the threshold for aggressive responding.We further examine the translational potential of targeting NMDAR-dependent plasticity, focusing on the clinically available antagonists ketamine and memantine. By integrating molecular, structural, circuit-level, and clinical perspectives, we propose that NMDAR-dependent control of synapse formation and location represents a plausible mechanism through which ELS persistently biases neural circuits that regulate aggression. Targeting these processes may provide new strategies for reducing pathological aggression following ELS. Introduction: early life stress, aggression, and neural circuit adaptations Childhood represents a period of heightened neural plasticity during which environmental experiences exert disproportionate and lasting effects on brain development. Adverse childhood experiences—including abuse, neglect, and exposure to violence—are highly prevalent and strongly associated with long-term psychiatric vulnerability (Abate et al., 2025; Association, 2013; Caspi et al., 2002; Herrenkohl et al., 2022; Widom, 1989; Widom and Brzustowicz, 2006). One of the most robust and societally consequential outcomes of early life stress (ELS) is a persistent increase in pathological aggression, often described as the “cycle of violence” (Widom, 1989). Consistent with this framing, individuals exposed to childhood maltreatment are significantly more likely to engage in reactive and impulsive forms of aggression later in life, with elevated rates of violent offending and recidivism (Hunt et al., 2019; Milaniak and Widom, 2015). A central challenge is to understand how transient stress exposure during development produces enduring changes in neural circuits that bias behavior toward aggression. Experience-dependent synaptic plasticity provides one mechanism through which stress can rapidly alter circuit function, changing the strength of excitatory inputs within stress- and aggression-related pathways (Chang et al., 2015; Chen and Baram, 2016; Danese and McEwen, 2012; Hodges et al., 2019; McEwen and Morrison, 2013; Mikics et al., 2018; Nordman et al., 2020a). However, persistent behavioral change likely requires mechanisms that stabilize these functional alterations over longer time scales. Emerging evidence suggests that N-methyl-D-aspartate receptor (NMDAR)-dependent structural plasticity can stabilize these synaptic and circuit changes through alterations in dendritic spine density and morphology, synapse number and size, and the spatial organization of synapses along dendrites (Bartsch et al., 2023b; Chen and Baram, 2016; McEwen and Morrison, 2013; Nava et al., 2017; Radley et al., 2004; Vyas et al., 2002). Such structural remodeling could stabilize stress-induced increases in circuit excitability, thereby lowering the threshold for aggressive responding long after the initial stress exposure has ended. In this review, we synthesize clinical, postmortem, circuit-level, and rodent evidence to develop a testable framework by which ELS engages NMDAR-dependent structural plasticity within aggression-related circuits, producing persistent and location-specific changes in excitatory connectivity that bias behavioral responses toward aggression. While ELS induces broad and heterogeneous effects across the brain, we focus specifically on how these mechanisms operate within defined aggression pathways, with particular emphasis on the posterior ventral medial amygdala (MeApv) to ventrolateral ventromedial hypothalamus (VmHvl) pathway. ELS-induced structural plasticity in aggression circuits Early life stress induces lasting changes in the neural architecture of social circuits that regulate aggression and related behaviors. While these effects are widespread, a subset of structural alterations within prefrontal-amygdala and amygdalo-hypothalamic pathways appear particularly relevant for stress-induced aggression. Across human and rodent studies, ELS consistently alters connectivity, dendritic structure, and synapse organization in these circuits, suggesting a conserved mechanism through which stress biases behavior toward aggression. Throughout this review, we focus on the role of stress in promoting maladaptive aggression, defined as aggression that is persistent, poorly regulated, disproportionate to the eliciting stimulus, or expressed outside of its typical social context. However, stress-induced aggression can also serve important adaptive functions related to defense, competition, or territoriality. Accordingly, not all increases in aggression resulting from stress are inherently pathological (Nelson and Trainor, 2007). Examining how ELS-induced structural plasticity alters circuit function may therefore provide a framework for identifying mechanisms that bias adaptive aggression toward more persistent and poorly regulated forms of aggression. Human studies of ELS-induced circuit remodeling Direct evidence for ELS-induced synaptic remodeling in humans remains limited, as synaptic structure cannot be measured directly in living individuals. However, postmortem studies suggest that ELS is associated with alterations in dendritic and spine architecture in regions implicated in behavioral regulation and aggression. Reductions in mature dendritic spines have been reported in the prefrontal cortex (PFC) of individuals with a history of ELS and post-traumatic stress disorder (PTSD) (Kaul et al., 2020; Young et al., 2015), while hippocampal spine density negatively correlates with lifetime stress exposure (Soetanto et al., 2010). These findings are consistent with volumetric imaging studies showing reduced gray matter in prefrontal and limbic regions (Lupien et al., 2009; Teicher et al., 2016), and collectively suggest that ELS may induce structural changes that impair regulatory control over stress-responsive circuits involved in aggression. Complementing these findings, diffusion tensor imaging (DTI) studies suggest that ELS is associated with alterations in white matter pathways linking the PFC and amygdala. One of the most consistently affected tracts is the uncinate fasciculus, which connects ventral prefrontal regions with the amygdala—key nodes for regulating emotional reactivity. Individuals exposed to childhood maltreatment frequently exhibit reduced fractional anisotropy in this tract (Hanson et al., 2015; Huang et al., 2012), and these alterations are associated with increased trait anger and aggression (Kim et al., 2019). Such findings suggest that impaired structural connectivity between regulatory and threat-processing regions may contribute to disinhibited aggressive responding. Functional magnetic resonance imaging (fMRI) studies further demonstrate that ELS alters activity and connectivity within these circuits (Gee et al., 2013; Hart and Rubia, 2012; Nelson et al., 2025). Individuals with high ELS exposure show heightened amygdala reactivity to emotional stimuli and reduced functional coupling between the amygdala and medial PFC (mPFC) (Cisler, 2017; Herzog and Schmahl, 2018). Importantly, reductions in amygdala–mPFC connectivity in children with childhood adversity are associated with increased aggression and impulsivity (Park et al., 2018), consistent with a model in which weakened top-down control permits exaggerated threat responses and reactive aggression. Abnormal functional connectivity within the default mode network is also associated with aggressive behavior in individuals who have experienced childhood maltreatment (Li et al., 2025; Zhang et al., 2022), and increased default mode network connectivity is predictive of conduct problems in individuals who have experienced ELS (Liu et al., 2025). Taken together, these studies suggest that ELS alters connectivity within circuits involved in emotional regulation and aggression. However, these imaging findings should be interpreted cautiously. Diffusion tensor imaging and fMRI can identify large-scale differences in white matter organization, regional activity, and functional connectivity, but they cannot determine whether ELS produces synapse-level structural remodeling. Diffusion tensor imaging measures water diffusion within tissue, which can be influenced by multiple features of white matter organization, while fMRI measures blood-oxygen-level-dependent activity rather than neuronal or synaptic structure directly. Accordingly, these studies provide evidence for circuit-level alterations following ELS, but do not directly demonstrate synaptic remodeling. This limitation highlights the importance of integrating human imaging findings with postmortem studies and mechanistic work in rodent models. Rodent models of ELS-induced structural plasticity Rodent models provide direct evidence that ELS induces structural remodeling within defined aggression circuits. A particularly well-characterized substrate is the MeApv–VmHvl pathway, which plays a central role in experience-dependent escalation of aggressive behavior (Mojahed et al., 2024; Nelson and Trainor, 2007; Nordman et al., 2020a, 2020b). Animals exposed to early life stressors such as chronic social isolation and foot shock exhibit increased dendritic spine density in the MeApv–VmHvl pathway (Bartsch et al., 2023b). These structural changes correlate with enhanced excitatory synaptic transmission and increased aggression, suggesting that ELS promotes aggression through coordinated structural and functional plasticity (Nordman et al., 2020a, 2020b). Complementary evidence indicates that activation of inhibitory somatostatin-expressing neurons within the MeA reduces aggression and rescues social deficits following ELS (Mañas-Ojeda et al., 2025), further supporting a role for excitatory MeA signaling in stress-induced behavioral adaptations. Although the MeApv–VmHvl pathway provides one of the clearest examples of ELS-induced structural remodeling linked to aggression, stress-related aggression likely emerges from adaptations across a broader network of interconnected circuits. Studies of peripubertal stress, adolescent social isolation, and traumatic stress implicate prefrontal–amygdala pathways in persistent aggression. Peripubertal stress produces hyperactivity within the amygdala, reduced orbitofrontal activity, and abnormal aggression, consistent with impaired orbitofrontal regulation of amygdala function (Marquez et al., 2013). Similarly, adolescent social isolation is associated with reduced activity of prefrontal pyramidal neurons and increased excitability of basolateral amygdala principal neurons during aggressive behavior, suggesting a shift in circuit balance toward amygdala-driven responding (Tan et al., 2021). Additional evidence implicates hypothalamic, bed nucleus of the stria terminalis (BNST), and hippocampal pathways in stress-related aggression. Social isolation-induced aggression has been linked to enhanced AMPA receptor-dependent transmission within a glutamatergic posterior amygdala–VmHvl pathway that promotes attack behavior (Zha et al., 2020). Similarly, traumatic stress potentiates MeApv–BNST projections and experimental induction of synaptic potentiation within this pathway is sufficient to increase attack behavior (Nordman et al., 2020a). Stress-induced aggression has also been associated with enhanced ventral hippocampal signaling through hypothalamic pathways, further supporting the involvement of distributed limbic–hypothalamic networks in aggressive behavior (Chang and Gean, 2019). Although these studies primarily examine synaptic plasticity, such changes are thought to contribute to the induction and stabilization of structural remodeling. Collectively, this work indicates that stress-induced aggression reflects plasticity across corticolimbic, extended amygdala, hippocampal, and hypothalamic circuits. However, whether structural remodeling within these pathways directly contributes to persistent aggression remains largely unresolved, making the MeApv–VmHvl pathway a particularly valuable model for investigating how NMDAR-dependent structural plasticity contributes to long-term aggressive behavior. Importantly, the behavioral consequences of ELS can diverge by sex. In our work, male mice exposed to chronic social isolation and foot shock during adolescence exhibit persistent increases in aggression, whereas females exposed to similar stressors do not show increased aggression (Aholt et al., 2026). Extending these observations, other studies have found that adolescent social isolation increases aggression in males, whereas females exhibit alterations in social behavior rather than increased aggression (Wang et al., 2022). These findings suggest that ELS can produce sex-specific behavioral outcomes, raising the possibility that stress engages distinct patterns of circuit remodeling across males and females. One potential explanation for these divergent behavioral outcomes is that stress engages distinct patterns of structural remodeling across males and females. Early life stress increases dendritic arborization and alters spine density within the infralimbic cortex of adolescent females, while producing little or no comparable structural change in males (Farrell et al., 2016). Similar sex-dependent effects have been reported following prenatal and chronic stress, which produce different patterns of dendritic remodeling in the prefrontal cortex and hippocampus of males and females (Galea et al., 1997; Garrett and Wellman, 2009; Murmu et al., 2006). Collectively, these findings suggest that developmental stress does not engage a uniform pattern of structural plasticity across sexes. Instead, stress may recruit distinct forms of circuit remodeling in males and females, potentially contributing to divergent behavioral outcomes following ELS. Taken together, these studies suggest that ELS induces structural plasticity across multiple brain regions. Emerging findings indicate that stress-induced behavioral adaptations likely reflect plasticity across distributed neural circuits, with the MeApv–VmHvl pathway providing some of the most direct evidence linking structural remodeling to persistent aggression (Fig. 1).Fig. 1Proposed role of NMDAR-dependent structural plasticity in the MeApv–VmHvl pathway in ELS-induced aggression. ELS increases the number of MeApv–VmHvl synapses, which may contribute to persistent aggression by altering excitatory connectivity within the circuit. Proposed role of NMDAR-dependent structural plasticity in the MeApv–VmHvl pathway in ELS-induced aggression. ELS increases the number of MeApv–VmHvl synapses, which may contribute to persistent aggression by altering excitatory connectivity within the circuit. NMDAR-dependent mechanisms of structural plasticity in aggression circuits The mechanisms through which transient stress exposure produces persistent structural remodeling remain incompletely understood. Among the candidate mechanisms, NMDARs are particularly well positioned to link stress exposure to long-term changes in synaptic and structural plasticity because of their central role in activity-dependent circuit modification (Danese and McEwen, 2012; McEwen and Morrison, 2013). In aggression circuits, NMDAR activation may provide a key molecular link between stress exposure and long-lasting circuit remodeling. Role of NMDARs in structural plasticity Activity-dependent synaptic and structural plasticity relies heavily on signaling through NMDARs (Squire and Kandel, 2009; Willard and Koochekpour, 2013). Unlike AMPA receptors, NMDARs require both glutamate binding and postsynaptic depolarization for activation, allowing them to function as molecular coincidence detectors that link neuronal activity to long-term circuit modification. At the structural level, NMDAR activation regulates dendritic spine dynamics through calcium-dependent signaling pathways that control actin remodeling (Bloodgood and Sabatini, 2009; Chen et al., 2011). Strong activation promotes spine growth and synapse stabilization, whereas weaker or prolonged signaling can drive spine shrinkage or elimination (Lee et al., 2009; Lisman et al., 2002; Lisman and McIntyre, 2001; Noguchi et al., 2016; Rust, 2015). Disruption of NMDAR signaling impairs these processes, leading to synaptic instability and loss (Alvarez et al., 2007; Ultanir et al., 2007), underscoring their essential role in maintaining and reshaping excitatory connectivity. NMDAR-dependent plasticity in aggression circuits following ELS NMDAR-dependent plasticity has emerged as a key mechanism linking experience to long-lasting changes in aggression-related circuits. Across both stress and winner-effect models, persistent increases in aggression are associated with NMDAR-dependent potentiation and structural remodeling within hypothalamic aggression pathways (Bartsch et al., 2023b; Nordman et al., 2020a; Stagkourakis et al., 2020; Yan et al., 2024). Work from our lab demonstrates that prior attack experience potentiates synaptic strength in the MeApv–VmHvl pathway through NMDAR-dependent long-term potentiation (Nordman et al., 2020a). In this pathway, optogenetic induction of LTP is sufficient to increase aggression, whereas NMDAR blockade prevents LTP-induced increases in aggression, providing direct evidence that NMDAR-dependent synaptic plasticity can drive persistent changes in aggressive behavior (Nordman et al., 2020a). Complementing these findings, ELS also induces structural remodeling within the MeApv–VmHvl pathway, including increased dendritic spine density and shaft synaptogenesis (Bartsch et al., 2023b). Pharmacological blockade of NMDARs during stress exposure prevents these structural changes, indicating that NMDAR activation is required for stress-induced remodeling of this aggression circuit. Taken together, evidence from both stress and winner-effect models supports a role for NMDAR-dependent structural remodeling in persistent aggression. Regional specificity of NMDAR-dependent plasticity following ELS Broader studies of ELS provide important mechanistic insight into how NMDAR-dependent plasticity becomes dysregulated across neural circuits. Early life stress alters NMDAR-mediated structural plasticity in the hippocampus, PFC, and amygdala, but these effects are regionally distinct. In the hippocampus and PFC, stress commonly induces dendritic atrophy and spine loss, whereas in the amygdala, ELS promotes dendritic growth and synaptogenesis (Bartsch et al., 2023b; Brunson et al., 2005; McEwen and Morrison, 2013; Radley et al., 2004). These opposing structural effects highlight the regional specificity of stress-induced plasticity. Consistent with this broader principle, stress also produces regionally distinct changes in NMDAR subunit composition, which may influence how plasticity is induced, stabilized, or constrained across circuits. During typical development, GluN2B-containing receptors predominate early and support prolonged calcium signaling and heightened synaptic plasticity, whereas maturation is accompanied by a shift toward GluN2A-dominant receptors that promote faster signaling kinetics and synaptic stabilization (Paoletti et al., 2013; Yashiro and Philpot, 2008). Under conditions of chronic stress, alterations in NMDAR subunit composition have been reported across multiple brain regions (Chang et al., 2015; Roceri et al., 2002; Wieck et al., 2013). GluN2B-dominant receptor profiles can be maintained or reintroduced in regions such as the amygdala and hippocampus, prolonging a highly plastic state that facilitates synaptic strengthening and circuit hyperexcitability (Chang et al., 2015; Roceri et al., 2002; Suvrathan et al., 2014). In contrast, stress-induced shifts toward GluN2A dominance in the PFC may prematurely stabilize synapses and limit adaptive plasticity (Wieck et al., 2013). In aggression-related circuits, such shifts in NMDAR subunit composition may bias the induction and stabilization of stress-induced structural plasticity, potentially contributing to persistent aggressive behavior following ELS. These observations suggest that NMDAR-dependent plasticity may influence not only the magnitude of structural remodeling, but also the specific organization of excitatory connectivity within aggression circuits. Understanding how stress shapes the structural architecture of these circuits is therefore critical for linking molecular plasticity to persistent behavioral change. Structural organization of NMDAR-dependent plasticity in aggression circuits Early life stress-induced NMDAR dysregulation may reshape excitatory circuitry not only by altering synapse number, but also by changing the spatial organization of synaptic inputs along dendrites. This distinction has important functional consequences for how neurons integrate input and generate output, particularly within aggression-related circuits. Traditionally, excitatory synapses are thought to occur primarily on dendritic spines, whereas dendritic shafts are more commonly associated with inhibitory inputs. However, recent evidence shows that excitatory synapses can also form directly on dendritic shafts, and these synapses have been documented across multiple brain regions and experimental contexts (Cooke and Woolley, 2005; Harris and Weinberg, 2012; Nava et al., 2017). Importantly, stressful experiences can promote the formation of excitatory shaft synapses. For example, acute stress increases excitatory shaft synapse number in the medial prefrontal cortex, suggesting that stress-induced synaptogenesis is not restricted to dendritic spines (Nava et al., 2017). Functionally, spine and shaft synapses differ in how they influence neuronal activity. Dendritic spines compartmentalize electrical and biochemical signaling, whereas shaft synapses contact the dendritic shaft directly and may exert stronger local depolarizing effects (Harnett et al., 2012; Yuste, 2013). In addition, synapses located closer to the soma, such as on proximal spines, are more effective at driving action potential generation than distal inputs (Ferreira et al., 2020; Parvez et al., 2010). As a result, the spatial distribution of synapses along dendrites can strongly influence the impact of excitatory input on neuronal firing. NMDAR-dependent remodeling of synapse location in aggression circuits after ELS Recent work from our lab demonstrates that ELS alters multiple structural features of excitatory connectivity within the MeApv–VmHvl aggression circuit. Foot shock–based ELS increases dendritic spine density and the number of excitatory synapses formed directly onto dendritic shafts (Bartsch et al., 2023b). Thus, ELS appears to affect both synapse number and synapse location within this pathway. Because shaft synapses bypass spine compartmentalization, their formation could have important consequences for how excitatory input is integrated within the circuit. This structural remodeling is dependent on NMDAR activation during stress exposure, as pharmacological blockade with MK-801 prevents both spine and shaft synaptogenesis (Bartsch et al., 2023b). In addition to increasing synapse number, ELS appears to alter synapse localization by biasing spine synapses toward dendritic regions with greater predicted influence on neuronal output. Within the MeApv–VmHvl pathway, stress preferentially increases proximal dendritic spine density, while leaving distal spine populations relatively unchanged (Bartsch et al., 2023b). Because proximal inputs are less attenuated as they propagate toward the soma, and because shaft synapses directly contact the dendritic shaft rather than a spine compartment, these findings suggest that ELS redistributes excitatory input toward sites predicted to exert stronger control over neuronal firing (Fig. 2).Fig. 2Proposed role of location-specific NMDAR-dependent structural plasticity in the MeApv–VmHvl pathway on ELS-induced aggression. Location-specific synaptogenesis may contribute to the persistence and intensity of long-lasting aggression. Proposed role of location-specific NMDAR-dependent structural plasticity in the MeApv–VmHvl pathway on ELS-induced aggression. Location-specific synaptogenesis may contribute to the persistence and intensity of long-lasting aggression. These findings suggest that changes in synapse number alone may not fully capture the functional consequences of stress-induced plasticity. Future studies will be required to determine the relative contributions of increased synapse number, shaft synaptogenesis, and redistribution of excitatory inputs to stress-induced aggression. Functional implications: structural plasticity as a mechanism for persistent aggression after ELS Stress-induced changes in synapse type and location provide a structural framework for understanding how ELS alters aggression circuits. By redistributing synapses toward locations with a greater impact on neuronal output, ELS may shift these circuits toward a more persistently excitable and less regulated state. In the MeApv–VmHvl pathway, this redistribution is predicted to enhance neuronal responsiveness to social stimuli, thereby increasing the likelihood of aggressive responding (Fig. 2). Together, these findings support a model in which ELS promotes location-specific, NMDAR-dependent structural plasticity with the potential to alter the functional output of aggression circuits. By increasing excitatory connectivity at sites predicted to exert greater influence over neuronal firing, stress-induced remodeling may enhance responsiveness to social experiences and contribute to persistent changes in aggressive behavior. NMDAR antagonists and the modulation of stress-induced plasticity: clinical implications Because NMDAR-dependent structural plasticity may contribute to persistent remodeling of aggression circuits following ELS, pharmacological modulation of these receptors has potential relevance for treating stress-related aggression. Clinical evidence suggests that NMDAR antagonists such as ketamine and memantine can reduce symptoms linked to pathological aggression in some populations. Preclinical studies further suggest that these effects may involve modulation of stress-induced changes in synaptic structure and circuit function. Therapeutic potential of NMDAR antagonists in stress-related disorders PTSD is frequently associated with heightened irritability, anger, exaggerated defensive responding, and reactive aggression, particularly in response to perceived threat or provocation. These behaviors are thought to emerge from chronic hyperarousal, dysregulated threat processing, and impaired emotional regulation following traumatic stress exposure, with reactive and defensive forms of aggression being especially linked to hyperarousal symptom clusters including exaggerated startle, sleep disruption, irritability, and persistent physiological activation (Novaco and Chemtob, 2015; Orth and Wieland, 2006; Taft et al., 2017). Clinically, ketamine has emerged as a promising intervention for treating symptoms of PTSD, including hyperarousal, reactivity, and aggression (Albuquerque et al., 2022; Almeida et al., 2024; Sicignano et al., 2024). A landmark study by Feder and colleagues demonstrated that a single subanesthetic infusion of ketamine produced substantial reductions in PTSD symptoms within 24 h, including decreases in intrusion, avoidance, and hyperarousal/reactivity (Feder et al., 2014). Follow-up studies using repeated-dose regimens reported more sustained therapeutic effects, with symptom improvements comparable to established trauma-focused psychotherapies (Duek et al., 2023; Feder et al., 2021). Importantly, PTSD symptom improvement following long-term ketamine treatment has been associated with increased functional connectivity between the ventromedial PFC and amygdala (Norbury et al., 2021). Given the role of this pathway in top-down regulation of aggressive and threat responses, these findings suggest that ketamine may influence functional balance within aggression circuits. Although functional connectivity does not directly measure structural remodeling, the observed alterations are consistent with ketamine-induced synaptic remodeling observed in preclinical models. For example, ketamine has been shown to restore stress-induced synaptic deficits in the PFC and hippocampus and promote synaptic remodeling associated with behavioral recovery in mice (Li et al., 2011; Moda-Sava et al., 2019). Together, these findings raise the possibility that ketamine influences stress-sensitive circuits, in part, through modulation of synaptic and structural plasticity. Emerging clinical evidence in other patient populations further supports a role for ketamine in modulating aggressive behavior. Case reports and longitudinal clinical studies indicate that subanesthetic ketamine treatment can reduce severe rage episodes and aggression-related symptoms, particularly in pediatric populations (Berner, 2007; Papolos et al., 2013, 2018). Interpretation of these findings must account for the sedative and psychotropic effects that can occur at higher doses, as well as ketamine's non-selective pharmacology. In addition to NMDARs, ketamine has been reported to interact with serotonergic systems, including 5-HT2 receptors, which are implicated in the psychoactive effects of serotonergic psychedelics and have also been linked to ketamine's psychotropic actions (Kapur and Seeman, 2002; Vollenweider et al., 2000; Waelbers et al., 2013). However, the studies discussed here used subanesthetic dosing regimens that did not produce overt sedation or marked perceptual disturbances (Duek et al., 2023; Feder et al., 2014, 2021). These findings indicate that ketamine's therapeutic effects are not solely attributable to its acute sedative or psychoactive effects. Memantine, a lower-affinity NMDAR antagonist which has been effective in reducing aggression and agitation associated with Alzheimer's disease and dementia (Cummings et al., 2008; Wilcock et al., 2008), has also shown promise in the treatment of PTSD. Open-label studies report improvement in core PTSD symptoms, including hyperarousal, avoidance, and cognitive impairment (Hori et al., 2021; Ramaswamy et al., 2015). These clinical effects are consistent with preclinical evidence that memantine can modify stress-related behavioral outcomes and synaptic plasticity markers in hippocampal and prefrontal circuits (Amin et al., 2015; Quan et al., 2011). Translational implications of targeting NMDAR-dependent plasticity Unlike conventional anti-aggression treatments that primarily suppress behavioral output acutely, NMDAR-targeting interventions may alter the persistence of stress-induced behavioral phenotypes by modulating underlying circuit plasticity. Because NMDARs are major regulators of synaptic and structural plasticity, modulating NMDAR signaling during or after stress exposure could theoretically limit the formation of stress-induced circuit changes or promote adaptive remodeling after stress has occurred. Thus, interventions targeting NMDAR-dependent plasticity may ultimately provide a means of modifying the underlying circuit abnormalities that contribute to persistent aggression following ELS. Despite this therapeutic potential, NMDAR antagonists should not be treated as mechanistically interchangeable. Although ketamine and memantine both block NMDARs, they differ in pharmacological properties that may contribute to divergent behavioral effects following stress exposure (Lipton, 2006; Paoletti et al., 2013; Zanos and Gould, 2018). This pharmacological diversity may be especially relevant in stress-induced aggression, where treatment effects appear to depend strongly on timing and behavioral context. For example, in our traumatic stress model, ketamine administered during stress enhances later excessive and reactive aggression, whereas memantine prevents this increase (Bartsch et al., 2023a; Nordman et al., 2022), indicating that these drugs may differentially influence stress-induced circuit remodeling. One possible explanation is that ketamine's effects during stress are not mediated solely through NMDAR blockade. Ketamine can increase AMPA receptor-mediated excitatory signaling and engage BDNF/TrkB-, mTOR-, and synaptogenesis-related pathways (Autry et al., 2011; Li et al., 2010; Zanos and Gould, 2018). When administered during periods of heightened stress-related circuit activity, these plasticity-promoting effects could theoretically facilitate persistent remodeling of aggression-related circuits. By contrast, these same plasticity-promoting effects may produce very different outcomes when ketamine is administered after stress exposure. In this context, ketamine may instead support recovery or adaptive remodeling in regulatory circuits, including prefrontal networks. Consistent with this possibility, ketamine has been shown to reverse stress-induced spine loss and restore coordinated activity in prefrontal microcircuits (Li et al., 2011; Moda-Sava et al., 2019). These findings highlight an important translational consideration: the effects of NMDAR-targeting interventions may depend strongly on whether treatment occurs during or after stress exposure. Interventions administered during stress may drive aberrant forms of circuit remodeling that heighten behavioral pathology, whereas post-stress interventions may instead promote adaptive remodeling within regulatory networks. This distinction may be especially important for ketamine, whose plasticity-promoting effects could facilitate remodeling that drives aggression when stress- and aggression-related circuits are strongly active, but support recovery when administered after stress exposure. Future studies should therefore consider both the timing of intervention relative to stress exposure and the distinct pharmacological properties of individual NMDAR antagonists when evaluating their therapeutic potential. Defining how different NMDAR-targeting strategies influence structural plasticity within aggression circuits may ultimately help identify approaches for reducing persistent aggression following ELS. Conclusions and future directions Across human and rodent studies, ELS alters connectivity, dendritic architecture, and synapse organization in aggression-related circuits. Converging evidence identifies NMDAR-dependent structural plasticity as a key mechanism through which transient stress exposure becomes embedded in these circuits, potentially producing long-lasting increases in aggressive behavior. A key insight from recent work is that ELS not only alters the number of excitatory synapses, but also their spatial organization along dendrites. In the MeApv–VmHvl aggression circuit, stress-induced increases in excitatory shaft synapses and proximal dendritic spines suggest a redistribution of synaptic input toward locations that more effectively drive neuronal firing. This shift in synapse location provides a structural mechanism through which ELS may increase circuit excitability and lower the threshold for aggressive responding, highlighting synapse placement as a critical determinant of circuit function and behavioral output. These insights have important translational implications. Pharmacological targeting of NMDARs, particularly with agents such as ketamine and memantine, shows promise in modulating stress-related circuit remodeling. While these interventions may restore aspects of synaptic plasticity and circuit function, an important challenge for future research will be determining whether they can also normalize the spatial reorganization of synapses within affected circuits. Such an approach would be critical for reversing the stress-induced hyperexcitability in aggression circuits that may be responsible for the ensuing pathological aggression. Addressing these questions will require temporally precise pharmacological manipulations combined with longitudinal structural analyses and in vivo circuit recordings. Ultimately, defining when and how NMDAR-dependent structural plasticity can be modified will be essential for clarifying the mechanisms linking ELS to persistent aggression and may inform more precise strategies for mitigating long-term violence and aggression associated with ELS. Funding This work was supported by a 10.13039/100009633Eunice Kennedy Shriver National Institute of Child Health and Human Development NIH grant (R15HD110925) to J.C.N. CRediT authorship contribution statement Jessica T. Jacobs: Conceptualization, Formal analysis, Methodology, Supervision, Validation, Writing – original draft, Writing – review & editing. Jacob C. Nordman: Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Writing – original draft, Writing – review & editing. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.