The production of testosterone increases in the male fetus early in the third month of development through birth. It mediates effects on the male genitalia and hypothalamus during this time. About 2 months after birth there is another surge in testosterone production which last several months although its significance is not understood. The final surge in testosterone production occurs during puberty where it contributes to growth, development of the primary sex organs and secondary sexual characteristics, and influences sexuality, personality, violence, and other cognitive aspects (Ramirez, 2003).

One of the difficulties in studying the effects of testosterone is to distinguish between correlation and causality. There are studies which have concluded that males who commit violent acts (including domestic violence) have higher testosterone levels and some males with high testosterone levels display high irritability (George, 2001). Men with higher testosterone levels have been shown to be less likely to marry, more likely to divorce, more likely to leave home after marriage, more likely to engage in extramarital affairs, and more likely to abuse their spouses than men with lower levels of testosterone. Married men are less likely to commit homicide; a crime whose frequency is similar in divorced and single men (Booth, 1993; Mazur, 1998). In surveys of prison inmates, testosterone levels are higher in those who had committed violent crimes and in those that had more prison rule violations. Men with higher levels of testosterone have been observed to engage in behaviors such as delinquency, drug use, alcohol abuse, military AWOL, and promiscuity (Dabbs, 2000). A study done on prison inmates and testosterone levels concluded that among the perpetrators of homicide, testosterone levels were highest among those who knew their victims and planned the murder in advance (Dabbs, 2001). In women prison inmates, testosterone levels are associated with higher levels of criminal violence and aggressive dominance (Dabbs, 1997). Testosterone levels in veterans were correlated with the amount of combat they experienced. Trial lawyers had higher testosterone levels than non-trial lawyers. Blue collar workers were higher in testosterone than white collar workers (Dabbs, 2000). Testosterone levels increase after winning a sporting competition (in participants and fans alike) and decrease after losing (Dabbs, 2000).

Testosterone in males is highest in the morning and is higher in late summer and early autumn than at other times of the year. Levels of testosterone fall over the course of a lifetime and can decrease temporarily in illness (Dabbs, 2000).

Anabolic steroids increase sex drive and enhance mental acuity and these effects have led to increased abuse of these substances in non-athletes (Pinna, 2005). In humans, high doses of anabolic steroids has been linked to indiscriminate violence sometimes referred to as "'roid rage". Studies have shown that anabolic steroids increase the incidence of verbal aggression, fighting, homicidal behavior and violence towards women. Studies with rats also indicate that anabolic steroids increase aggression (McGinnis, 2002). The known effects of anabolic steroids on mood in both males and females include mania, depression, irritability, and aggression (which has led to violent acts and even attempted homicide). The increased aggression caused by anabolic steroids seem to be mediated through a decrease in the brain neurosteroid allopregnanolone (Allo). Allo, in turn, increases the signal transduction of GABA receptors (Pinna, 2005). The decrease in allopreganlone levels is a factor in PMS (Brizendine, 2006).

Since the early history of the vertebrates, the hormone testosterone has mediated aggressive behavior. In humans and nonprimate humans, increased testosterone levels are associated with aggression and dominance behavior. In human males it is associated with violent crime, problems in the workplace, drug use, and other antisocial behaviors. Studies have shown higher than normal levels of adrenal androgens in aggressive children (Giammanco, 2005). Testosterone levels are higher in adolescent boys perceived to be socially dominant. In other species, aggressive acts are often initiated by males of lower social status in an attempt to be more socially dominant. Therefore, the lack of strong evidence for a direct correlation between testosterone levels and aggression may be influenced by the aggression committed by adolescents of a lesser perceived social status (with lower levels of testosterone) who are trying to become more dominant (Ramirez, 2003). One study found that perpetrators of domestic violence who also had alcohol dependence had higher testosterone levels than normal but that this was not true of perpetrators who did not abuse alcohol (George, 2001).

There are a number of hormones which can affect aggression and violence in addition to testosterone. Increased aggressive behavior may also be linked to lower levels of DHEAS, a lower ratio of testosterone to estrogen, lower levels of FSH, lower levels of prolactin, and higher levels of LH, androstenedione, and DHEA. The adrenal gland secretes several hormones which are called androgens because of the similarities of their actions to those of testosterone, androstenedione, DHEA (dehyroepiandrostenedione), and DHEAS (dehyroepiandrostenedione sulfate). Their effects are more easily observed in females since in males their effects are overshadowed by the testosterone secreted by the testes. There is some evidence suggesting that higher levels of adrenal androgens are factors in increased aggression and antisocial behavior (Ramirez, 2003).


Nitric oxide serves three main functions in the body. Outside of the nervous system, macrophages can synthesize it to destroy tumor cells and endothelial cells can secrete it to relax blood vessels. In the brain, nitric oxide can affect the activity of neurons and the enzyme which synthesizes it, nitric oxide synthase (NOS), is concentrated in regions of the brain which determine emotions. In animal studies, mutations in NOS greatly increase male aggression, but not female aggression. Testosterone is required for this increase; no such increase is observed in castrated males. In fact, normal female mice do display maternal aggressiveness toward intruders which is reduced in NOS mutant females. Thus it appears that the activity of NOS affects aggression but has opposite effects in males and females. Androgens tend to inhibit NOS and estrogens tend to increase its activity (Chiavegatto, 2003). In mice, absence of the gene nitric oxide synthase affects violence and sexual behavior, in part through effects on the serotonergic pathways and serotonin receptors (Reif, 2003).


There are two genes for monoamine oxidase, MAO-A and MAO-B. Deletions of MAO-A in humans result in mental retardation, autistic behavior, and other abnormalities (Reif, 2003). Mutations in MAO-A and 5TTT affect the organization of the cerebral cortex in mice (Reif, 2003).The absence of MAO-A expression in mice results in increased levels of some neurotransmitters (dopamine, serotonin, and NE), higher aggression, and inappropriate sexual activity in males. In humans, a mutation in MAO-A causes Brunner syndrome in which males suffer from mild retardation and display a variety of aggressive and hypersexual behaviors (in addition to other behaviors ranging from arson to suicidal behavior). This is the only example known which fulfills the OGOD (one gene, one disease) model for behavioral disorders. Variations in the promoter region are known to affect panic and depression in females and aggression in males (Reif, 2003).

The general model is that dopamine facilitates aggression while serotonin inhibits it. While drugs which are serotonin reuptake inhibitors decrease aggression in most individuals, there are some in which treatment actually increases aggression. The same is true of benzodiazepines which facilitate GABA receptors (Nelson, 2006).
In mice, aggression increases with tryptophan-free diets, lesions in serotonin-making neurons, or drugs which inhibit serotonin while aggression decreases with diets with precursors of serotonin and serotonin reuptake inhibitors (Nelson, 2006).Alleles of the dopamine receptor D2 and dopamine transporter genes are associated with increased aggression in humans (Chen, 2005). A certain allele of the serotonin 2A receptor gene is more likely to be found in criminals and psychiatric patients than in control groups. This variation has no other observable affect on personality scores (Berggard, 2003). Decreasing the activity of serotonin pathways is known to increase aggression in both humans and other animals. Increased aggression occurs when brain the levels of serotonin precursors decrease, under the influence of drugs which decrease serotonin synthesis, in response to tryptophan-free diets, and when neurons which produce serotonin are destroyed by lesions. Increasing serotonin levels has been observed to decrease aggression (Chiavegatto, 2003).

Low levels of cortisol are associated with increased aggression. This reduction in cortisol secretion can be observed in children and may thus be a predictor of aggressive behavior. Interestingly, there seems to be an interaction between culture and cortisol levels. For example, differences in the rise of cortisol levels after experiencing insult have been observed between adolescents raised in the Northern U.S. compared to the Southern U.S. (Ramirez, 2003). In animals, chronic stress and elevated corticosteroid levels decrease aggression while abnormally low function of glucocorticoids increased aggression (Nelson, 2006; Ramirez, 2003).

Pituitary hormones may also be factors in aggression. Studies have linked increased aggression to lower levels of FSH and prolactin and higher levels of LH and DHEA (Ramirez, 2003). In primates, lower cholesterol levels are associated with increased aggression. In humans, a number of studies have shown lower cholesterol levels are linked to increased levels of violence and death by violence (Golomb, 2000). Depression can increase levels of aggression (Nelson, 2006).


Given rapid changes in levels of violence in human cultures, there must be societal and environmental cues for violence in addition to biological signals. During the 1990s, there was a correlation with homicide rate and unemployment rate while in the 1980s, homicide rates remained constant despite drops in unemployment (Caulkins, 2007). Between 1960 and 1980, homicide frequency in the United States doubled (from 5.1 to 10.2 per 100,000 population) but by the year 2000 had returned to a value near the 1960 value (5.5 per 100,000) (Caulkins, 2007). Some have hypothesized that homicide rates are influenced by the rate of children who experience violence or the availability of abortion (Caulkins, 2007).

Aggression can also be increased through environmental variables. Mice become more aggressive when isolated and isolation is required for increased aggression in nitric oxide synthase mutants to become manifest (Chiavegatto, 2003).


Many of the social explanations for crime rates cannot explain why crime rates dropped during the 1990s. One model does account for this change: crime as facilitated by childhood exposure to lead. Lead is a neurotoxin and there is association between areas with high lead exposure and crime rate, after controlling for other variables. Since the elimination of lead based paint and leaded gasoline, lead exposure in U.S. children has dropped (Nevin, 2007).


There in interest in whether dietary factors can increase aggression. Although it has been observed that low blood sugar increases irritability and aggressive individuals may have low blood sugar, this is a correlation and may not represent a causal relationship. For example, another variable, such as serotonin levels, might increase aggression and lower blood sugar. The inclusion of omega-3 fatty acids in diet has been shown to decrease aggression and some studies have linked aggression to food intolerance (Benton, 2007).


There is conflicting evidence on whether cartoon violence increases aggression in children. One factor in the contradictory findings of studies may stem from the variations between the studies (such as whether the cartoon was comedic or whether aggression was measured as acts against peers or against a toy). Some have concluded that cartoon violence can disinhibit aggressive behaviors by making children more likely to display behaviors they have already learned (Kirsh, 2006).
In general, violence in comedic cartoons has not been shown to increase violence towards other children although violence in non-comedic cartoons is associated with increased aggression directed at peers and toys. Some studies have shown an increase in aggressive thoughts (through questionnaires) although no changes in aggressive behavior were observed (Kirsh, 2006).

Studies have concluded that violent video games increase aggression, anger, aggressive thoughts, and desensitization to violence. About 85% of games involve some degree of violence and about half involve intense violence (Carnagey, 2007). Exposure to video game violence decreased empathy and increased aggressive attitudes (Funk, 2004). Violence in film has been shown to result in desensititzation to violence (Carnagey, 2007). Individuals who rate high on aggressiveness respond more to violence in film, video games, than nonviolent men (Giumetti, 2007).


In some religious and ethnic groups, 'honor killings' take the lives of women who have acted differently than the norms set by the family or community. The UN estimates that 5000 women are killed a year (Meetoo, 2007).


Biologically, humans are most closely related to apes and descriptions of ape violence can suggest parallels to human violence. An estimated 10% of gorilla infants are killed by an adult male gorilla who displaces another male as leader of a harem of females. The majority of males commit infanticide at least once during their lives and most females will have an infant killed by an adult male gorilla (Bagemihl, 1999; Wrangham, 1996).

Chimps can attack each other in lethal gang attacks. High ranking female chimps have been observed to kill low ranking females and to commit infanticide of the offspring of low-ranking females. Aggression is often displayed towards immigrant females. High-ranking females mature faster than low ranking females, perhaps because they are better nourished (De Waal, 2001).Male chimpanzees are known to kill the infants of chimpanzees who belong to other groups. A few cases are known in which males killed infants in their own group if the female was a new immigrant to the group. Infanticide was not committed on the offspring of immigrant females once they had mated with the males of the group (De Waal, 2001).

Chimpanzees and humans are unusual among animals in that they are prone to kill individuals of their own species which belong to different social groups (De Waal, 2001). In 1974, the first case in which non-humans deliberately sought for an individual of their own species in order to attack and kill them was recorded. A group of common chimps silently passed into the territory of a neighboring group, encountered a solitary male, and beat him to death. In 1977, the members of one group of common chimps fatally assaulted the males of a neighboring group one by one over time and abducted the females until the neighboring group no longer existed (Wrangham, 1996).

Compared to common chimpanzees, Bonobo males are less aggressive and form closer bonds with females which are not limited to the period of estrus. Unlike common chimp males, Bonobo males do not form lethal raiding parties or commit infanticide (Savage-Rumbaugh, 1998).

In both orangutans and common chimpanzees, heterosexual intercourse can involve forceful coercion of the female. In orangutans, female coercion may be a factor in as many as 1/3 of matings. Once, an orangutan male raped a human woman. In common chimpanzees, females can be assaulted by males of the group which apparently results in a female being more likely to submit to the male (Wrangham, 1996).