There is growing scientific interest in understanding the biological mechanisms affecting and/or underlying violent behaviors, in order to develop effective treatment and prevention programs [1
]. In this regard, the relatively recent appearance of the field of neurocriminology represents an important advance in our understanding of these problems by applying the neuroscientific perspective to their study. In fact, neurocriminology aims to establish the neurobiological basis for crime and violence. Specifically, this neuroscientific subdiscipline has incorporated several tools and/or procedures, such as neuroimaging techniques, genetic markers, and hormonal measurements, among others, to predict these antisocial behaviors [3
Neuroimaging techniques are non-invasive, and make it possible to visualize brain structures and functional connectivity in brain networks, thanks to their good spatial and functional resolution. Indeed, functional magnetic resonance imaging (fMRI) has offered insight into the functional synchrony between brain structures, that is, how the activation of several brain structures is temporally coordinated [4
]. These techniques have usually relied on showing changes in the activation of different brain networks by analysing the blood oxygen level-dependent (BOLD) presented in these brain regions, which is especially sensitive to the increase in blood flow in the cerebral capillaries of the activated neuronal regions [6
Studies using fMRI to assess brain networks have demonstrated that altered functional connectivity across distant brain regions might make individuals prone to violence [8
]. In fact, most of the research in this field has highlighted that the alteration in the cerebral connectivity between the key nodes involved in emotional and cognitive behavioral regulation might explain this proneness to violence. Particularly, the inhibitory malfunction of the frontal lobe (i.e., prefrontal structures, frontal gyrus…) would lead to an overactivation of the limbic system (i.e., amygdala, hypothalamus, hippocampus…), which under certain stimuli might facilitate impulsive and/or reactive violence. In this regard, it has been suggested that the activation of frontal structures facilitates self-regulation and control over emotion-related behaviors by attenuating limbic responses to emotional stimuli and/or the context [8
]. By contrast, several authors demonstrated that individuals characterized by predatory and instrumental violence (proactive) might present normal prefrontal cortex (PFC) functioning, and an increase in dorsolateral PFC activation has even been described during emotion processing tasks in these individuals [16
]. Furthermore, studies have indicated that the reduction in amygdala activation during emotion processing might be characteristic of instrumental violence [19
]. In this case, unimpaired or even higher frontal activation is related to the ability to control impulses and/or emotional processing, but also to certain alterations in empathic abilities. Unfortunately, the majority of the previously mentioned studies analyzed the functional activation of these brain structures in response to certain tasks (i.e., exposure to a stimulus, emotional induction tasks…), and so little is known about whether these brain structures and their connections in the absence of external stimulation might be employed as reliable markers of proneness to violence.
In recent years, neuroscientific research has not only been interested in brain functioning during a task, but it has also focused on the intrinsic activity within the brain at rest without any external stimulation (resting-state functional connectivity; RSFC). In this direction, several studies have demonstrated the existence of resting intrinsic brain connectivity (networks) when an individual is awake and alert [20
]. Synchronicity across a group of distant brain regions during resting periods has been called the default mode network (DMN; precuneus/posterior cingulate cortex, medial PFC and medial, lateral, and inferior parietal cortex). Although it has been suggested that the DMN is relatively deactivated during a demanding task [23
], some authors have maintained that it might make an active contribution to cognitive processing [26
], or even be important for cognitive abilities [28
] or a correlate of certain personality traits [31
]. With regard to the aforementioned, it should be mentioned that the activation of individual brain structures that encompass the DMN does not necessarily mean that the DMN is activated. In fact, it is necessary to consider how these brain structures establish and maintain networks with other brain structures outside the DMN during the resting state. Since evidence about whether the DMN or other brain networks might be important in violence proneness is less clear, it would be important to summarize the main results of this field of research.
In light of the above, this review aims to highlight the association between the RSFC among specific brain structures and the predisposition to experiencing anger and/or responding to stressful and distressing situations with anger. Specifically, we will focus on the following concepts related to violence: reactive violence (characterized by impulsivity, emotional drive, and lack of self-control); proactive violence (characterized by planning, high awareness of the purpose of this conduct, and low emotionality); trait anger (frequency of experiencing feelings of anger); and anger expression (frequency of externally and/or internally manifesting angry feelings). The present study will first describe the main findings on the association between RSFC and self-reported anger in non-violent normative individuals and in others with mental disorders. Then, we will describe the association between RSFC and changes in anger levels after laboratory tasks in non-violent and violent populations. Next, with the main purpose of analyzing whether these connections could be employed as reliable markers of violence proneness, we will analyze the RSFC in highly violent populations (i.e., young offenders, inmates, and soldiers) in comparison with non-violent individuals. Finally, considering the existing data, we will discuss the implications for clinical practice and further research.
The results described here offer a better understanding of RSFC that might explain proneness to violence, particularly the neural pathways that underlie key variables of violence, such as reactive and proactive violence, trait anger, and anger expression and/or control. It should be noted that only one study failed to find a significant association between the RSFC of the bilateral anterior temporal lobe and mOFC and the anger trait; the other 16 manuscripts did find a significant association between these variables. Even though several studies highlighted that the diminished RSFC between the PFC and the amygdala increased proneness to reactive violence, we also summarized other brain networks that include additional cortical (i.e., insula, gyrus (angular, supramarginal, temporal, fusiform, superior and middle frontal), ACC, and PCC) and subcortical brain structures (i.e., hippocampus, cerebellum, ventral striatum, and nucleus centralis superior) that might facilitate the onset of this type of violence. Moreover, we also described the neural pathways that might explain proactive violence and feelings of revenge, which are focused on the RSFC between the OFC, ventral striatal, angular gyrus, mid-occipital cortex, and cerebellum (Table 2
Initially, we concluded that the diminished RSFC between frontal structures such as the PFC (OFC and vmPFC) and frontopolar cortex (Brodmann area 10; PFC), limbic structures such as the amygdala and the anterior insula, parietal cortex regions such as the supramarginal gyrus (Brodmann area 40; parietal lobe), the angular region (posterior to the supramarginal gyrus), and the nucleus centralis superior (median raphe nucleus; brainstem) was related to a high predisposition to experiencing anger and/or responding to stressful and distressing situations with anger (high anger expression-out and low control-out) in a normative population and in participants with mental disorders.
As previously stated, it appears that prefrontal structures tend to maintain inhibitory projections to other brain structures involved in emotional reactivity, such as the amygdala and/or the insula, and so the failure to regulate this reactivity might lead to high irritability or hostile reactions [8
]. Although these studies did not analyze the RSFC between these brain structures, all of the manuscripts that were included in this systematic review found a diminished RSFC connectivity between the PFC and amygdala, which has been associated with proneness to experiencing feelings of anger and difficulties in controlling anger expression [34
]. In fact, alterations in the dorsal and ventral PFC, amygdala, and angular gyrus have commonly been associated with rule-breaking behaviors, alterations in moral judgement and reasoning, and emotion regulation [8
]. In this regard, we reinforced the hypothesis that diminished RSFC between the frontal and limbic systems might be characteristic of violent populations, since one of the studies that was included in this review concluded that a group of violent and impulsive soldiers presented a lower RSFC between the left dorsolateral PFC and the basolateral amygdala (bilateral) than a non-violent group [47
]. Conversely, research conducted with normative young adults demonstrated that individuals with higher RSFC between the mOFC and the basolateral amygdala showed less aggressive strategies on an anger induction laboratory task [46
]. Moreover, another study stated that reactive violent offenders presented a diminished RSFC between the left mOFC and left amygdala, but an increase in paralimbic RSFC after an emotion induction task [47
Although it appears that the diminished RSFC between the PFC and amygdala might be employed as a useful marker for proneness to violence, other brain networks should be considered in order to offer a broader understanding of the complex phenomenon of violence. Even though other limbic structures and their projections are involved in proneness to violence, we can conclude that high RSFC between the amygdala and the inferior frontal gyrus and left superior temporal gyrus was associated with high anger traits in several populations (normative and with mental disorders). Furthermore, the amygdala and the ACC maintained high RSFC with the fusiform and lingual gyrus, cuneus and precuneus calcarine cortex, and superior occipital cortex in impulsive and aggressive individuals. Recently, a research group demonstrated that violent individuals with schizophrenia presented hyperactivation of the ACC when perceiving negative images [51
], especially in highly impulsive schizophrenic individuals [52
]. This result coincides with the hypothesis that a large percentage of violent offenders present an attentional bias toward negative stimulus and/or a hostile attribution bias [53
]. Moreover, other brain networks composed of the inferior frontal and temporal gyrus, ACC, and anterior insula are involved in voluntarily and actively sharing the emotional experience of other individuals (through intentional empathic processes), which is extremely important in behavioral regulation [8
]. Thus, we sustain that this specific brain network facilitates the onset of violence due to attributing negative and hostile intentions to others, which is congruent with the scientific literature in this field.
Regarding other limbic structures, the hippocampus and the anterior insula maintain anger expression facilitation projections with the supramarginal and angular gyrus, the superior and middle frontal gyrus, the ventral striatum, and the ACC. In this regard, the middle frontal gyrus plays an important role in hostile cognitive bias and angry rumination (or how often an individual tends to re-experience negative feelings) [54
]. We also hypothesized that the heightened connectivity between the left/right orbitofrontal cortex (OFC) in violent inmates [49
] may play a role in angry rumination. Furthermore, it was recently demonstrated that brainstem alterations (except its volume) seem to partly explain high irritability and proneness to react violently in males with autism [12
]. Moreover, in two groups of violent inmates, the authors concluded that the PFC (mPFC and OFC) also maintained lower connectivity with adjacent areas and the cerebellar vermis, compared to non-violent males [41
], with the cerebellum joined to the basal ganglia, and the supplementary motor area being important in impulse control [9
]. Therefore, we might summarize that these RSFC patterns in a cognitive model where high anger traits that were linked to empathic alterations (i.e., perspective-taking disruptions, emotion-decoding deficits…), hostile cognitive biases, and heightened anger rumination might lower the threshold for reacting with violence in ambiguous contexts, physically facilitating reactive violence.
With regard to proactive violence, one study that was included in our review used self-reports to analyze the RSFC underlying this kind of violence. However, the authors did not compare reactive and proactive violence, and so it may be difficult to establish in this review whether there are differentiated neural pathways to each kind of violence [56
]. Kolla, Dunlop, Meyer, and Downar [38
] concluded that high RSFC between the ventral striatal and the angular gyrus was associated with high proactive violence strategies. Curiously, psychopaths (who usually but not always employ proactive violence) have been found to present structural abnormalities in the striatum—particularly increased volume—compared to non-psychopaths [57
]. Moreover, it has been suggested that the angular gyrus belongs to the brain networks underlying moral reasoning [8
]. Thus, these results offer information about which neural pathways underlie this kind of violence, which is defined as predatory, and is often characterized by premeditation and being cold-blooded [58
]. Extending these findings, another study [42
] revealed that revenge feelings were positively correlated with the RSFC between the OFC and the cerebellum, angular gyrus, and midoccipital cortex. Furthermore, recent research highlighted that psychopathic traits, as well as the risk of reoffending, maintained a positive relationship with the grey matter volume of the cerebellum [9
]. If we try to explain why these neural pathways explain revenge feelings, we can assume that these neural pathways are associated with several processes that are important for revenge, such as the anticipation of consequences of violent behavior (good or bad) and reimagining different contexts to consume the desire for revenge [58
Finally, one way to prevent these antisocial behaviors is to promote empathic abilities in violent populations through various psychological interventions. Thus, it has been hypothesized that there is an overlap between these brain regions that underlie violence and empathy, but the pattern of functional connectivity underlying each of them is inverse [57
]. Regarding RSFC, a recent study demonstrated that individuals (men and women) with higher empathic abilities tend to present greater RSFC between the mPFC and the dorsal ACC, the precuneus, and the left superior temporal sulcus [58
]. Conversely, low levels of empathy have been related to lower RSFC between the mPFC and ACC [30
]. Moreover, it has been stated that high affective empathy tends to be related to stronger RSFC between the ventral anterior insula, OFC, amygdala, and perigenual anterior cingulated, and that high cognitive empathy is related to greater RSFC between the brainstem, superior temporal sulcus, and ventral anterior insula [59
]. This statement is congruent with the assertion that there is a certain overlap between the brain structures that underlie violence and empathy, but they present inverse RSFC patterns. Thus, we might conclude that the brain RSFC in empathic abilities is more harmonious (i.e., positive RSFC between the PFC and the limbic system), than the pattern for violence (i.e., negative RSFC between PFC and the limbic system), which describes an imbalance in the brain functioning of individuals who tend to present proneness to violence. Based on the summarized results of our review and the brain RSFC underlying empathy, we proposed the facilitating and inhibiting brain networks for reactive violence described in Figure 2
Three important limitations of the studies carried out are the lack of a homogeneous population and their limited sample size. In fact, there is great diversity in the relevant variables, and not all of the studies reported or controlled the potential confounding effects of variables such as handedness, drug use, educational level, economic level, ethnicity, and psychopathology, among others. Furthermore, only a few studies included women, and so most of the research was conducted with males. Although small sample sizes kept us from properly estimating the differences between groups, most of the studies applied Bonferroni corrections for multiple comparisons. Moreover, they employed different anger trait questionnaires that measure different facets of anger (trait, physical, or verbal tendency to express anger, anger expression in general, proactive violence…). Lastly, several of the articles that were included in our review based their conclusions on seed-based findings that are necessarily limited to a priori regions. In this regard, the absence of a finding in another region is very different from a study that may have used ICA. Hence, it is difficult to obtain unanimous conclusions about the association between RSFC and anger.
In summary, the present review study confirmed that reduced frontal and limbic connectivity is a good correlate for several variables, such as trait anger and anger expression or control, which are important in violence proneness. Nevertheless, in order to offer a broader understanding of violence proneness, we need to contemplate other resting-state brain networks, including cortical regions (i.e., gyrus, parietal, temporal, ACC…) and subcortical structures (i.e., hippocampus, insula, brain stem, cerebellum…), in addition to the PFC and amygdala. Therefore, based on the studies summarized in this manuscript, along with those that analyzed the RSFC underlying empathic processes, we proposed a model to explain anger proneness to reactive violence (Figure 2
). In sum, our review offered more insight into the importance of studying the brain RSFC underlying several important processes for violence proneness. It also reinforced the need to carry out further studies that analyze the importance of the RSFC using larger sample sizes, contemplating several populations, and employing a common anger assessment. Moreover, it would be necessary to check whether these RSFC could be considered temporally stable (as a ‘trait’), or whether they might change after interventions focused on promoting behavioral regulation and cognitive improvements. In fact, several studies demonstrated that after a focused intervention to promote cognitive and empathic changes in groups of intimate partner violence (IPV) perpetrators, these individuals improved specific cognitive and empathic abilities [60
]. Thus, this research allows us to detect whether these changes correspond to specific brain network changes. Therefore, caution should be used in interpreting these results, in order to develop effective intervention programs to reduce proneness to violence.