1. The Neuroanatomy of Aggression
1.1. Frontal Cortical Structures
Where to start? Principally, where most neuroscientific research concerning behavior and decision-making starts—the cortex. Cortical structures are fundamentally associated with aggression; amongst those, structural and functional deviations in the prefrontal and medial temporal regions have been extensively investigated and meta-analyzed
[1][2]. For instance, the ability to control aggression was linked to the anterior cingulate cortex (ACC) in both rodents and human beings. Volumetric and neuronal activity alterations were found to correlate with social misconduct and callousness in adolescents
[3][4] as well as with aggression in mice
[5]. Additionally, van Heukelum and colleagues
[5] aimed to provide a link between ACC activation and the suppression of pathological aggression. The study applied a chemogenetic approach to activate the ACC and consequently repressed pathological aggression almost completely in mice.
Furthermore, sub-cortically both prefrontal and temporal abnormal white matter integrity was reported in 15 male convicted rapists compared to controls
[6]. In light of these findings, interventions targeting prefrontal areas seemed to be expedient. To date, four studies applied transcranial direct current stimulation (tDCS) to prefrontal areas attempting to modulate aggression and aiding a causality understanding. However, the results were mixed. Stimulating the right dorsolateral prefrontal cortex (dlPFC) reduced proactive aggression
[7], whereas stimulating the left dlPFC triggered the opposite effect if participants were already in a negative affective state
[8]. On the other hand, stimulating the inferior prefrontal cortex yielded no result at all
[9]. Still, bilaterally stimulating the dlPFC caused participants to view aggression and sexual violence as more morally despicable and decreased the intentions of committing aggression
[10]. Conversely, lesions to the dlPFC promoted physical aggressiveness, as demonstrated in a study with Vietnam War veterans who received penetrating brain injuries
[11].
A different mean of stimulation, continuous Theta-burst transcranial magnetic stimulation (cTBS), further cooperated an involvement of dlPFC activity in aggression. Interestingly, inhibiting the left dlPFC caused healthy volunteers to act reactively and proactively more aggressively, contradicting the findings described above
[12]. An activating cTBS protocol across both hemispherical dlPFCs, however, led to an increased accuracy in of identifying emotions and reduced likelihood of aggression in participants with an antisocial personality disorder
[13].
The work of Chester and colleagues
[14][15][16] implicated two additional parts of the prefrontal cortex (PFC) in the complex matter of aggression; the ventromedial (vm) and the ventrolateral (vl) PFC. The vmPFC showed volumetric and gray matter deficits in the context of heightened aggression, whereas augmented activation of the vlPFC during rejection could be linked to increased retaliating aggression. Additionally, an overall decreased gyrification of the frontal lobe, as well as prefrontal thinning, were related to intensified aggressive behavior in school children
[17]. The involvement of these PFC areas in circuits of escalated aggression has been highlighted in rodents as well
[18][19].
Finally, another subregion of the prefrontal cortex, the orbitofrontal cortex (OFC), seems to play a role as well, especially given the famous case of Phineas Gage
[20]. It does not surprise that across species, OFC abnormalities or chemogenetic inactivation are associated with increased aggression
[21][22][23].
1.2. Temporal Cortical Structures
An initial report in a small population of highly violent criminals was the first to implicate metabolic abnormalities of lower medial temporal lobe structures in aggression
[24]. Volumetric differences in adolescents with a conduct disorder diagnosis were then described by Kruesi and colleagues
[25]. Next, a review of 17 imaging studies further strengthened the involvement of temporal structures in aggressive behaviors
[26], with the recent work of Buadas-Rotger and colleagues
[27] linking angry face reactivity in the superior temporal gyrus (STG) to task-related aggression. On a network level, this hyperreactivity was related to activation in the amygdala. Hence it is of no surprise that the neuro-moral theory of antisocial behavior emphasizes the involvement of the superior temporal gyrus, the angular gyrus, and the tempoparietal junction as key brain areas
[28]. A recent meta-analysis of 2022 further confirmed this role of temporal lobe structures activities in aggression-prone individuals
[29].
1.3. The Striatum
The striatum is a forebrain structure divided into a ventral and dorsal part, which is in humans further separated into the caudate and the putamen. Whereas the first receives strong dopaminergic inputs from the ventral tegmental area (VTA), the latter one is mostly innervated by the substantia nigra (dorsal part). The ventral part of the striatum is thought to facilitate learning, motivated behavior, and reward evaluation
[30]. Unsurprisingly, imaging studies revealed functional abnormalities during increased retaliation-oriented aggression
[15]. Additionally, violent offenders responded to provocations with high striatal reactivity
[31]. Finally, increased volumes of the caudate, the putamen, and the nucleus accumbens (NAc; ventral striatum) were associated with augmented reactive aggression
[32], while increased dorso-striatal activity measured with functional magnetic resonance imaging (fMRI) revealed a higher motivation to punish unfairness
[33]. Interestingly, recently developed animal models demonstrated a “highly rewarding” aspect of aggression and violence to be mediated in dopaminoceptive D1-neurons in the NAc
[34]. This “winner effect” describes the phenomenon that victorious mice are more likely to maintain successful aggressive behaviors
[35], prefer the compartment associated with the victory
[36], and respond in an operant conditioning paradigm with lever-pressing to encounter an intruder mouse to fight it
[37].
1.4. The Limbic System
Hostility and anger combined in one outburst is called rage
[38]; such rage is known to originate in the septum across species
[39][40][41], a limbic structure in the epicenter of a vast neuronal aggression network connecting the amygdala, the VTA, and the hippocampus, and even extending to the PFC
[42][43]. Modern neuroscientific circuitry studies based on optogenetics in mice revealed glutamatergic inputs from the CA2 region of the hippocampus and dopaminergic innervations from VTA to the septum to promote aggression
[44][45]. On the other hand, overall inhibition and activation of the descending outputs from the lateral part of the septum (LS) to the hypothalamus were shown to modulate aggression bidirectionally. Inhibition of LS increased aggression through a GABAergic mechanism in rodents
[46], whereas activation of the LS-to-hypothalamus connectivity reduced aggressive behavior
[41], indicating an intricate sub-circuitry within the septum and its subdivisions.
Longstanding experimental evidence implicated hypothalamic structures in aggression and violence independent of species (for review, see
[47]), even coining the term “hypothalamic attack area”, located at the ventrolateral pole of the ventromedial hypothalamic nucleus (VMHvl;
[48]). Single-unit electrophysiology and population optical recording demonstrated an increase in activity in the VMHvl immediately before the initiation of an attack
[37]. Silencing this brain area abolished inter male aggression while activating it led to off-target attacks on inanimate objects and female mice
[49]. Additionally, optogenetic stimulation of the prefrontal-cortical innervations to the more mediobasal and lateral parts of the hypothalamus was also found to be crucial for initiating violent bites in intermale aggression in mice
[50].
In human beings, the “hypothalamic attack area” is suggested to be located in the more posteromedial part of the hypothalamus, also called the “Triangle of Sano”
[51]. However, even though initial surgical interventions suggested promising results
[52], deep brain stimulation of the posteromedial hypothalamus during pathological aggressiveness was of limited success
[53][54][55][56].
Another brain area implicated in aggressiveness—and where our memories are formed—is the hippocampus. A recent chemogenetic approach, for instance, demonstrated ventral hippocampal control of stress-induced aggression in mice
[57]. From a human perspective, a study with 67 participants suffering from intermittent explosive disorder (IED) revealed morphometric deformation in the hippocampus, which according to the authors, represented significant neuronal loss
[58]. Along this line of thinking, the volumetric ratio between the hippocampus and amygdala has been implied in many psychopathologies, including heightened aggression
[59]. On a functional level, Nikolic and colleagues
[29] reported increased activity in both the left hippocampus and amygdala in individuals with a known history of violence during anger-eliciting tasks. Increased connectivity between the right hippocampus and the cingulate cortex could be associated with aggressive behavior of patients with a history of mild concussions
[60].
The amygdala is a complex brain area that can be divided into a central and medial part on a microanatomical level. It plays a role in a multitude of cognitive functions like learning and memory. However, the amygdala is also implicated in several pathological conditions like anxiety disorder or addiction. Haller suggested in his review
[61] that predatory aggression is especially mediated through the central part of the amygdala. In line with this thought, chemogenetic experiments identified a neuronal subtype in the posterodorsal-medial part of the amygdala controlling aggressive behavior in both male and female mice
[62], while estrogen receptor-carrying neurons in the posterior amygdala signaling to the hypothalamus fulfilled the same role in inter-male aggression
[63]. An optogenetic approach identified the posterior ventral segment of the medial part of the amygdala to be essential in primed aggressive behavior
[64]. Here optogenetic high-frequency stimulation mimicked aggressive encounters and subsequently increased aggressive behavior in mice. Overall chemogenetic inhibition of the medial amygdala abolished lactating female-to-male mice aggression
[65]. In human beings, amygdala reactivity to fearful facial expressions was linked to impulsive aggression
[66], while children showing pronounced aggressive behavior had a decreased amygdala volume compared to controls
[17]. In line with this observation, another study reported significant structural deformations of the amygdala in IED patients
[58]. A longitudinal study with 503 participants running from first grade to the age of 26 provided further evidence of a link between decreased amygdala volume, violence, and even psychopathic traits
[67]. Nonetheless, a recent study failed to link amygdala volumes to scores in the revised Psychopathy Check List (PCL-R;
[68]). In contrast, a longitudinal observational study from Saxbe et al.
[69] linked aggression and externalization running through family lines to a volumetric increase in the right amygdala. Fundamental work from Raine and colleagues
[70] went even further and differentiated brain metabolic activity regarding proactive and impulsive aggression in convicted murderers. PET scans revealed that both groups had heightened metabolism in the right limbic system (i.e., amygdala, hippocampus, and midbrain), although being “excessively” augmented in predatory murderers. Accordingly, the work of da Cunha-Bang
[31] reported increased amygdala reactivity to provocations in violent offenders. In another set of studies, Heesink et al.
[71], Varkevisser et al.
[72], and Ibrahim et al.
[73] pointed out that the connectivity between the amygdala and PFC was decreased in war veterans as well as children with aggression issues. Conversely, Saxbe et al.
[74] found such a decrease in connectivity to be restricted to the amygdala and the posterior cingulate cortex. On the other hand, resting-state fMRI revealed an overall increase in amygdala connectivity but pronounced severity in aggression was associated with a global reduction in connectivity in the vmPFC and dorsal ACC
[75]. However, the trait “aggressiveness” did not relate to amygdala-OFC connectivity
[76].
Finally, the VTA in itself and especially its primary dopaminergic cell population projecting to the LS were recently demonstrated to powerfully modulate aggression in rodents
[77]. On a molecular level, dopamine 2 (D2)-receptors expressed on GABAergic neurons were identified as key players mediating such behavior
[45]. Interestingly, winning “mice territorial disputes” was shown to prime VTA-to-NAc innervations, making the occurrence of aggression more likely
[78]. Along with these types of observations, a recent study was able to link VTA hyper-reactivity to impulsive aggression and anti-sociality
[79]. It is also worthwhile to note that Been and colleagues
[80] underlined the VTA as a key node in the female rather than male circuitry of aggression.
1.5. The Parietal and Occipital Lobe
The precuneus, as part of the parietal lobe, is a brain area implied in a variety of complex functions and is part of the default mode network. Unsurprisingly, a recent meta-analysis of fMRI studies confirmed a significantly changed connectivity of the precuneus and frontal brain structures in aggression-prone individuals
[29], as well as in juvenile violent offenders
[81] and adolescents displaying disruptive behavior
[82]. Additionally, increased neuronal reactivity to socio-emotional stimuli in the precuneus was linked to aggression
[83][84].
The cuneus, a part of the occipital lobe, is a central part in the evaluation of threat. In line with this, work from Varkevisser and colleagues
[72] and Heesink and colleagues
[71] demonstrated increased connectivity between the ACC and the cuneus and its overall activation in impulsive-aggressive combat veterans. Interestingly, the recent work of Zhu and colleagues
[85] reported a link between the cortical thickness of both the parietal and occipital lobes and overall aggressiveness in 240 participants. Finally, Nikolic and others
[29] reported an overall reduced activity of the left occipital cortex in their recent meta-analysis of imaging studies in aggressive individuals.
2. The Neurotransmission and Genetics of Aggression
2.1. Serotonin, Dopamine and Their Degradation
Serotonin (5-hydroxytryptamine; 5-HT) plays an important role in higher cognitive functions as well as mood and is considered “the” most important neurotransmitter contributing to aggression
[86]. It can act through 15 overall receptors divided into 7 families: 5-HT
1 to 5-HT
7. Of those, are all but 5-HT
3, which is a ligand-gated sodium channel, are either excitatory or inhibitory G protein-coupled receptors
[87].
The so-called “Serotonin deficiency” hypothesis dated to work from Linnoila and others
[88], where 5-hydroxyindoleacetic acid (5-HIAA), a serotonin metabolite, turned out to be significantly lower in the cerebrospinal fluid of violent offenders. Yet, despite that this inverse relation was repeatedly described as “strong” or “well established”, a recent meta-analysis of studies including more than 6500 participants demonstrated not more than a small, marginally significant effect
[89].
Despite such a modest effect, modern neuroscientific tools keep pointing towards a serotonergic role in aggression. Since the discovery that the central nervous system has its own enzyme to synthesize serotonin (Tryptophan hydroxylase-2 (Tph-2);
[90]), it was shown that a partial or complete knockdown led to more aggressive behavior in rodents
[91][92][93][94] and could be linked to a diminished 5-HT
1A receptor sensitivity
[95]. On the other hand, highly aggressive wild-derived mice had higher mRNA levels of Tph-2 than other laboratory animals
[96]. In human beings, functional polymorphisms like TPH2 G-703/T (G/G instead of G/T and TT) could be linked to anger-related traits in Korean women
[97] and an anger-trait and negative OFC volume interaction
[98], as well as changed in OFC serotonin synthesis in aggressive subjects
[99]. Similarly, Laas and colleagues
[100] found in Estonian adolescents the T/T variant carriers of the TPH2 G-703/T polymorphism show the least aggressiveness.
In line with Tph-2 knockdown results, the pharmacological depletion of serotonin itself also increased the attack duration in mice
[101]. On a receptor level, altering serotonergic cell signaling either by genetic overexpression of the serotonin 1a receptor (5-HT1A receptor) in adult mice or pharmacological 5-HT1A receptor activation led to an immediate increase in aggressive behavior
[102]. Conversely, systemic administration of the 5-HT1A receptor antagonist 8-OH-DPAT inhibited social-isolation-enhanced aggression in mice
[103]. Specific 5-HT1A receptor activation in the hypothalamus, on the contrary, had sex-dependent behavioral effects in hamsters. While female hamsters reacted aggressively, male hamsters reduced their aggressiveness
[104]. Further complicating the matter, microinjections of the novel 5-HT1A receptor agonist into the OFC had anti-aggressive properties, thereby opposing most of the findings described above
[105]. Finally, a recent pharmacological study implicated the 5-HT2C receptor in aggressive behavior as well; 5-HT2C receptor antagonism with the compound SB243213 reduced social-isolation-induced aggression
[106].
In humans, multiple single nucleotide polymorphisms (SNPs) in the 5-HT2A receptor have been related to antisocial and aggressive behavior in young adults diagnosed with conduct disorder and in two traditional and one industrial Russian populations
[107][108]. Furthermore, post-mortem investigations of the OFC in nine individuals with a history of antisocial behavior and substance use disorder (SUD) revealed increased 5-HT2A receptor levels compared to a SUD-only group and healthy controls
[109]. However, frontal cortex 5-HT2A receptor binding density could not be linked to trait aggressiveness or impulsivity in healthy human subjects
[110]. Interestingly, the same research group did report a positive correlation between striatal 5-HT1B receptor binding and the trait anger in violent offenders
[111]. Somewhat surprising, given the finding of da Cunha-Bang and colleagues
[110], the 5-HT1B receptor SNP rs6296 could only be related to childhood aggressiveness but did not associate with aggression during adulthood in a Finnish population
[112]. In a third study, da Cunha-Bang and colleagues
[113] used the fact that 5-HT4 receptor binding could be used as a proxy for 5-HT levels. Here, total 5-HT4 receptor radiotracer binding correlated highly with self-reported aggression levels in men. In other words, the lower the global serotonin level, the higher the self-reported aggression score.
Dopamine is a neurotransmitter attributing salience to stimuli, generating perceptional valence, and thereby promoting motivation to act. It also helps animals and human beings to obtain rewards and avoid negative outcomes. Dopamine asserts control over neuronal activity via two groups of dopaminergic receptors—the d1-like and the d2—like family. The first group consists of d1- and d5-receptors, the latter one of d2-, d3-, and d4-receptors
[114].
Mahadevia and colleagues
[45] were able to pinpoint augmented aggression in mice to a d2-receptor-mediated mechanism in the LS, while aggression self-administration and aggressive behavior were controlled by NAc d1-receptor neurons
[34] through the transcription factor ΔfosB
[115]. On the other hand, d1 receptors were found to be significantly reduced in the frontal cortex
[116], with lower dopamine concentrations in the PFC as well as the hippocampus in highly aggressive rodents
[117]. These findings were supported by a large gene expression investigation revealing Slc-gene-related alterations linked to repeated victorious aggressive encounters in mice
[118]. In this research, Slc genes linked amongst others to the dopamine transporter were significantly downregulated in the VTA as well as the NAc, with upregulation patterns in the PFC. Somewhat consistently, methylphenidate, a dopamine transporter blocker, reduced aggression and promoted sociability in mice
[119]. Bridging these differences, the administration of a novel dopamine stabilizing compound (-)-OSU61612 had reportedly anti-aggression properties in socially isolated rats
[120].
The results outlined above appeared to be translatable since the work of Pape et al.
[121] concluded that methylphenidate led to a normalization of NAc connectivity to cortical regions linked to moral decision-making and attention in adolescents with disruptive behavior disorder. However, behavioral changes were not assessed during the exploration.
An array of investigations looking at variations in dopamine genes revealed a significant correlation with aggressive behaviors. In children, for instance, the seven-repeat allele variant of the d4 receptor gene served as a significant predictor for aggression and violence in a four-year follow-up study
[122]. Similar findings were described in young adults by Buchmann et al.
[123] and Wu and Barnes
[124], who were able to link the variant to both aggression and altered cortisol responses, as well as psychopathy. Furthermore, the seven-repeat allele variant was highly present in a cohort of Russian and Chechen felons convicted for severe violent crimes, while a functional variant of the dopamine transporter was more associated with habitual offenders
[125]. The same dopamine transporter gene variant was also found to be more prevalent in Pakistani inmates convicted of murder
[126].
The degradation process of serotonin and dopamine occurs mainly through two enzymes—monoamino-oxydase (MAOA) and catechol-O-Methyltransferase (COMT)
[127]. Both enzymes are known to have functional genetic variants linked to aggression and antisocial behavior, which have been studied extensively over the past decades
[128]. Several upstream variable-number tandem repeat variations (uVNTRs) have been described in the promotor region of the MAOA gene, which controls its enzymatic activity. Of those, the two-repeat variant (2R; MAOA-low) is associated with lower transcriptional efficiency than the more common three- or four-repeat variants (3R or 4R; MAOA-high), resulting in different metabolic efficacies
[127]. While MAOA occurs in both the degradation of serotonin and dopamine, COMT regulates exclusively extracellular dopamine levels by metabolizing dopamine into 3-Methoxytyramine (3-MT), which is in turn converted into homovanillic acid (HVA) by MAOA. The most common functional COMT polymorphism is Val108/158Met, where the amino acid methionine is replaced by valine with the Val/Val variant causing the enzyme to be two to four times more active
[129].
In their landmark study from 2002, Caspi and colleagues
[130] were the first to demonstrate an interaction between violence, adverse childhood events, and the MAOA-low variant. However, Ficks and Waldman
[131] called the overall effect size of MAOA uVNTRs on antisocial behavior, including aggression, “modest” in their meta-analysis of 31 follow-up studies. Hence, it seems intriguing that there are criminal proceedings allowing MAOA-low genotyping as evidence, even resulting in lesser charges or sentences
[132]. However, the most substantial evidence for the role of MAOA in aggression was demonstrated by MAOA knockout mice being more aggressive
[133] and an observation made by Labonté and colleagues
[134]. In the latter study, genetic modification of the expression of a long noncoding RNA regulating MAOA in the hippocampus led to a decrease in MAOA and an exacerbation of impulsive aggression in mice. Similar results were reported via pharmacological inhibition of MAOA during developmental phases in rodents
[135]. Since pharmacological blockade of serotonin reuptake reduced the augmented aggression in MAOA-deficient mice, aggression may very well be a presynaptic serotoninergic effect rather than dopamine-mediated
[136]. On the other hand, a recent study on 150 forensic psychiatric outpatients receiving aggression treatment cooperated the trauma and aggression interaction findings, but only for those patients with multiple traumas. Additionally, no effect on the treatment outcome was reported
[137]. Nonetheless, a study a novel EEG based investigation pointed to hyper-responsiveness to threatening voices in MAO-low carriers
[138] and increased aggressive reactivity to provocation
[139] as well as morphological changes in the amygdala
[140], while other studies in different ethnicities fully replicated Caspi’s findings
[141][142]. To complicate matters more, gender may play a role as well. The MAOA × childhood experiences × aggression interaction had opposing effects in male and female participants and could be linked to amygdala activity during a reactive aggression task
[143]. Finally, a study on male Italian prisoners reported dissuading results. Higher scores of aggression self-reports were observed in MAOA-low carriers, but only if they were exposed to physical neglect during growing-up; otherwise, the highest scores were those of MAOA-high carriers
[144]. Similarly, Zhang et al.
[145] reported the MAOA-high carrier to be the most aggressive in Chinese male adolescents, provided that a second polymorphism in the serotonin transporter gene was present.
The Valine/Methionine (Val/Met) COMT SNP research in human beings produced contradicting results regarding which variant (Val/Val or Met/Met) may influence more aggression or be linked to psychopathy
[129][146], though the Met/Met-induced aggressive behavior was shown to be phenotypically preserved across species
[147]. On the contrary, Hygen and colleagues
[148] suggested that the Val/Val variant makes adolescents more malleable, rather than vulnerable, to adverse childhood experiences, leading to more aggressive behaviors. In line with this, male adolescent carrier for the Val/Val variant and with attention deficit disorder displayed poorer impulse control and reduced fear empathy
[149]. Then again, Val/Val carriers in a Swedish population were associated with lower levels of physical aggression when exposed to violence compared to Met/Met carriers while still having a positive parent-child relationship
[150]. Finally, the work of Fritz and colleagues
[151] suggested a possible novel angle that may explain some of the dichotomous findings. In their study, a double SNP analysis focusing on the degradation pathway of dopamine was conducted in forensic inpatients. The adverse childhood x aggression interaction occurred only in participants carrying both the Val/Val highly active and the MAOA-low variants, suggesting that 3-MT may be responsible for the aggressive phenotype.
2.2. Glutamate
Glutamate is the most important excitatory neurotransmitter in the CNS and mediates its effect through two classes of receptors—ionotropic and metabotropic receptors. The most prominent ionotropic receptors are the α-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and Glutamate N-methyl-D-aspartate (NMDA) receptors. In the metabotropic receptor class, there are currently eight receptors in three functional groups described—mGluR
1 and mGluR
5 (group I), mGluR
2 and mGluR
3 (group II), and mGluR
4, mGluR
6, mGluR
7, and mGluR
8 (group III)
[152].
In 2004, work from Vekovischeva and colleagues
[153] first implicated a direct role of AMPA receptors in aggression, as mice with genetically deleted or functionally reduced AMPA receptors show significantly less inter-male aggressive behavior. Ten years later, a pharmacological approach with a systemic administration of an AMPA receptor antagonist corroborated those findings
[154]. Interestingly, higher AMPA receptor levels were reported in the amygdala and the PFC in highly aggressive mice
[155]. Finally, Zha et al.
[156] demonstrated that projections from the VMHvl to the posterior part of the amygdala mediated aggression in an AMPA receptor-dependent manner.
Contrarily, targeting NMDA receptors to alter aggressive behavior appears to yield a somewhat biphasic response. Systemic ketamine administration, an NMDA receptor antagonist, decreased aggression at high doses but augmented it at low doses in adolescent mice
[157]. Pharmacological blocking of NMDA receptors restricted to the PFC, however, had a diminishing effect of aggression in MAOA deficient mice solidly
[158][159]. Targeting NMDA receptor subunits like GluN1 further strengthened these pharmacological findings, as GluN1 knockout mice displayed reduced aggressive behavior
[160]. Furthermore, decreases in GluN2b subunit expression in the lateral amygdala was associated with aggression in socially isolated rodents
[161].
Amongst the three groups of metabotropic receptors, interventions targeting group I had the most substantial effect. Blocking mGluR
5 pharmacologically induced significant alterations in aggressiveness in both mice and hamsters
[162][163]. Finally, recent work in mice implicated a small subpopulation of VTA glutamatergic neurons receiving strong inputs from the habenula to turn escape behavior into reactive, defensive aggression
[164].
Somewhat surprising were the reported findings from Craig et al.
[165]. In their study, ACC glutamatergic concentrations in the ACC were distinguishable according to the prevalent trait in the participants. Whereas callous-unemotional individuals had increased glutamate levels in the ACC, those showing proactive aggression had decreased levels. In line with these findings, another study investigating aggressive behavior in individuals with antisocial personality disorder found a positive correlation between glutamine (a precursor for glutamate) levels in the dlPFC and aggression
[166]. However, Ende and colleagues
[167] did not observe augmented ACC glutamate levels in aggressive test subjects with borderline personality disorder, though a positive correlation was shown with measures of impulsivity.
2.3. γ-Aminobutyric Acid (GABA)
GABA is the inhibitory neurotransmitter in the central nervous system and binds to two classes of receptors, the GABA
A receptor, a ligand-gated ion channel, and the GABA
B-receptor, a metabotropic G protein-coupled receptor
[168][169]. Both received substantial scientific attention in the context of aggression, especially since it became apparent that alcohol mediates some of its CNS effects through GABAergic receptors (for a summary of earlier research, please see
[170]).
In mice, a recent systematic analysis combining microarray, quantitative polymerase-chain-reaction (PCR), and magnetic resonance imaging (MRI) based spectroscopy linked a 40% reduction of GABA in the ACC together with a 20-fold increase of the GABA-degrading enzyme
Abat to excessive aggression in mice
[171]. On the other hand, a pharmacology-based approach micro-infusing the GABA
A receptor agonist muscimol into the ACC demonstrated an increase in hyper-aggressiveness in mice
[172]. Similarly, muscimol infusion into the LS changed the behavior of defeated Syrian hamsters from submissive to aggressive
[173], an effect that was corroborated by Borland and colleagues
[174] independent of the social experience of defeat. In line with these findings, muscimol- driven GABA
A receptor activation of the OFC yielded the same behavioral outcome in rats, suggesting that GABA
A receptors may play an important role in the modulation of aggression
[23]. In a global genetic modification approach using point-mutated mice, Newman and colleagues
[175] further specified the GABA
A receptors containing an α2 subunit to be mediating aggression. Aside from some methodological oddities, novel optogenetic stimulation of neuronal pathways connecting through or originating in GABAergic neurons in the medial amygdala pointed collectively towards a central GABAergic role in promoting aggression
[62][64][176][177].
On a human level, the aforementioned preclinical findings appear to be translatable. In both patients with attention deficit disorder and borderline personality disorder, GABA levels in the ACC negatively correlated with self-rating aggression scores in an MRI-based spectroscopy approach
[167]. Blood GABA concentrations in violent offenders with antisocial personality disorder positively correlated with aggression
[178]. Furthermore, GABA
A receptor 2 gene polymorphisms rs279826 and rs279858 A-allele carriers were demonstrated to have an adverse stressful life event and aggression interaction
[179].
2.4. Inflammatory Markers
Cytokines and prostaglandins have been demonstrated to act both as mediators of inflammation and neurotransmitters
[180][181]. Hence, it is not surprising that several studies on cats were able to demonstrate a link between cytokine administration and augmented aggression
[182][183][184]. In line with this, genetic deletion of the two tumor-necrosis-factor alpha receptors caused the durations of aggression to decrease
[185]. Interestingly, a study based on a resident-intruder-challenge revealed that pharmacological or genetic alterations of interleukin-1 receptors in the dorsal raphe nucleus consistently increased aggressive behavior in mice
[186]. Finally, lipopolysaccharide challenges in rats selected for the trait “aggressiveness” caused an augmented upregulation of interleukin-1 in the frontal cortex but a decrease in the hippocampus compared to controls
[187].
In humans, studies from Coccaro and colleagues
[188][189] were able to link peripheral interleukin-6 to impulsive-aggressive subjects and increased cerebrospinal fluid levels of the soluble interleukin-1 receptor II to aggressive patients with personality disorders. Earlier work of Provencal and others
[190] found the anti-inflammatory cytokines interleukin-4 and interleukin-10 to be lower in children with aggression issues. However, a recent study failed to find an association between functional SNPs within interleukin-1beta, interleukin-2, and interleukin-6 in aggressive children
[191].
2.5. Opioid Receptors
The mu(µ)-opioid receptor 1 (OPRM1) is commonly regarded as the most relevant opioid receptor for aggression. Especially its functional single nucleotide polymorphism C77G (corresponding to A118G in human beings) was linked to augmented cortisol responses and heightened aggressive behavior in rhesus monkeys
[192]. Driscoll and colleagues
[193] further supported this interaction, linked it to serotonin, and expanded it as a predictor of alcohol-stimulated aggression. It is also worthwhile mentioning that heroin addicted rats produced more aggressive offspring than their control group, which likely suggests that µ-opioid receptor activation causes aggressive behaviors
[194]. In another study, Weidler et al.
[195] demonstrated a correlation between the OPRM1 SNP A118G and physical aggression in human beings. A very recent work of Cimino and colleagues
[196] was also able to associate the A118G SNP with disruptive behavior and limitation in mood regulation in children.
2.6. Orexin
Preclinical and clinical research also indicates a central role for orexin in the control of aggressive behavior. Recent work from Flanigan and colleagues
[197] implicated orexin signaling via the orexin 2 receptor (OxR2) on GABAergic cells in the lateral habenula to play a role in intermale aggression in mice. Previously, the same group has also suggested orexin to control the valence of aggression through a strong habenula-to-VTA neuronal connectivity in mice
[198].
In humans, the HCRTR1 rs2271933 A/A carrier genotype was associated with heightened aggressive behavior
[199]. Such genotype-phenotype interaction was argued to be based on an augmented reward association with violence
[200].
2.7. Oxytocin/Vasopressin
Oxytocin and vasopressin are fundamentally linked and guide the regulation of social behavior, such as parental nurturing and social cognition, but they also relate to affective states and emotional discrimination
[201][202][203]. Over the last decades, the hormone vasopressin has received a great amount of attention in the context of aggression (for a detailed review, see
[204]). For instance, vasopressin release patterns were demonstrated to differ in low and high-anxiety rat strains across brain areas in the context of intermale aggression
[205]. Furthermore, glutamate-vasopressin interaction in the lateral hypothalamus augmented steroid-induced aggression in Syrian hamsters
[206]. Conversely, genetic deletion of the vasopressin receptor 1b (V1b) reduced aggression in male mice
[207]. Somewhat counterintuitive, heightened oxytocin levels in the paraventricular nucleus (PVN) and the amygdala was also linked to high levels of maternal aggression in lactating rats
[208]. The interplay between oxytocin and vasopressin continued to be complex, as demonstrated in recent work in the ventral and dorsal LS
[209]. Using a mixed neuropharmacological, optogenetic, and chemogenetic approach, Oliveira and colleagues showed that an increase in oxytocin level in the ventral LS and a decrease in vasopressin level in the dorsal LS were associated with aggressive behavior in rats. Strikingly, systemic administration of both vasopressin and oxytocin suppressed isolation-enhanced aggressive behavior in male mice in the social interaction test
[210].
On a human level, intranasal administration of vasopressin caused an increase in “preemptive strikes” in a computer simulation in both male and female participants
[211]. In test subjects with high Machiavellian traits, a combined administration of testosterone and vasopressin augmented punishing behavior during a provocation paradigm
[212]. In a different study, even though increased activity in the right superior temporal sulcus was measured during a competitive reaction game, intranasal vasopressin administration failed to alter the behavior in the test group
[213].