M.L. Walsh

Aggression in humans: what is its biological foundation?

Although human aggression is frequently inferred to parallel aggression based on testosterone in nonprimate mammals, there is little concrete support for this position. High- and low-aggression individuals do not consistently differ in serum testosterone. Aggression does not change at puberty when testosterone levels increase. Aggression does not increase in hypogonadal males (or females) when exogenous testosterone is administered to support sexual activity. Similarly, there are no reports that aggression increases in hirsute females even though testosterone levels may rise to 200% above normal. Conversely, castration or antiandrogen administration to human males is not associated with a consistent decrease in aggression. Finally, changes in human aggression associated with neuropathology are not consistent with current knowledge of the neural basis of testosterone-dependent aggression. In contrast, human aggression does have a substantial number of features in common with defensive aggression seen in nonprimate mammals. It is present at all age levels, is displayed by both males and females, is directed at both males and females, and is not dependent on seasonal changes in hormone levels or experiential events such as sexual activity. As would be expected from current knowledge of the neural system controlling defensive aggression, aggression in humans increases with tumors in the medial hypothalamus and septal region, and with seizure activity in the amygdala. It decreases with lesions in the amygdala. The inference that human aggression has its roots in the defensive aggression of nonprimate mammals is in general agreement with evidence on the consistency of human aggressiveness over age, with similarities in male and female aggressiveness in laboratory studies, and with observations that some neurological disturbances contribute to criminal violence. This evidence suggests that human aggression has its biological roots in the defensive aggression of nonprimate mammals and not in hormone-dependent aggression based on testosterone.

Hormone-dependent aggression in male rats is proportional to serum testosterone concentration but sexual behavior is not

Male hooded rats were castrated and implanted with Silastic capsules (1.57 mm i.d.; 3.18 mm o.d.) having a testosterone-filled space 0, 7, 22, 60, or 90 mm long. All animals were returned to their original group cages for a three-week period to allow hormone concentrations and behavioral tendencies to stabilize. Each male was then housed with an intact female in a large cage. Aggression by the male toward an unfamiliar male was tested at weekly intervals for three weeks. Sexual behavior with an estrogen/progesterone-primed ovariectomized female was tested on each of the subsequent two weeks. Serum testosterone was measured during the following week. The frequency of aggression was correlated with serum testosterone concentration up to the normal level and did not increase with higher serum testosterone concentrations. In contrast, sexual behavior was virtually absent in animals with no testosterone replacement and normal in all other groups. These results demonstrate a clear dissociation in the dependence of hormone-dependent aggression and sexual behavior on serum testosterone concentration. In a male cohabiting with a female, sexual experience activates hormone-dependent aggression toward an unfamiliar male but the level of aggression that develops depends on the serum testosterone concentration in the resident male.

Serum estradiol concentration required to maintain body weight, attractivity, proceptivity, and receptivity in the ovariectomized female rat

Female hooded rats (230 to 260 g) were ovariectomized and given a subcutaneous implant of an estradiol-filled Silastic tube. The length of the tube was varied in order to produce a variety of serum estradiol levels. In the first experiment, animals were weighed over a 6-week period following surgery and then tested for sexual responsiveness to a male. The results demonstrated that ovariectomized females with an implant maintaining a serum estradiol concentration at about 15 pg/ml maintained body weight at the same level as that of intact females. A smaller implant gave rise to a higher weight gain and a larger implant to a lower weight gain. All implants resulted in a continuous state of receptivity. In a second experiment, ovariectomized females were implanted with smaller estradiol-filled implants in order to determine the threshold for maintaining proceptivity and receptivity. The results indicated that with a serum estradiol concentration below 15 pg/ml, the frequency of lordosis and of ear wiggling and darting decreased. Progesterone injections facilitated both proceptive and receptive behavior. In addition, following progesterone injections, the time required for a male to mount a female 10 times was decreased in females with low or no estradiol replacement. These results indicate that a constant concentration of estradiol at about the mean level present throughout the estrous cycle will result in normal body weight regulation and will maintain sexual behaviors that normally occur only during estrus. These results emphasize that Silastic implants of estradiol do not mimic normal endocrine function since, even at low levels, estradiol implants produce continuous receptivity.

Cohabitation with a female activates testosterone-dependent social aggression in male rats independently of changes in serum testosterone concentration ☆

Male hooded rats (350 to 450 g) were sham-castrated, castrated and implanted with testosterone-filled, or castrated and implanted with empty Silastic tubes. Twenty-four hours postoperatively the animals in each group were housed with a female or a male similar in size to the female. Beginning one week following surgery and continuing for three weeks thereafter, the female or male cagemate was removed once each week while a 15-min test of aggression toward an unfamiliar male intruder was conducted. During the aggression tests, lateral attacks, lunge attacks, bites, on-top, and piloerection were recorded. At the first aggression test, males housed with females were significantly more aggressive than their counterparts housed with males. In contrast, different testosterone regimes did not consistently influence the initial activation of intermale social aggression. At the second and third tests, males with testicular testosterone or a replacement were significantly more aggressive than their castrated controls on most measures but males housed with females continued to be more aggressive than the comparable group housed with males. These results suggest that normal fluctuations in serum testosterone concentration associated with sexual interaction are not necessary for the initial activation of intermale social aggression. Both repeated exposure to unfamiliar males as well as cohabitation with a female are effective stimuli for activation of testosterone-dependent social aggression.

Interaction of estradiol, testosterone, and progesterone in the modulation of hormone-dependent aggression in the female rat.

Female rats that had become aggressive as a result of cohabiting with a sterile male were ovariectomized and implanted with Silastic tubes of estradiol, testosterone, and progesterone, estradiol and testosterone alone, or with empty tubes. The implants were designed to model serum concentrations present during the last week of pregnancy (estradiol, 0.06 ng/ml; testosterone, 2.6 ng/ml; progesterone, 70 ng/ml). Following a test of aggression 1 week postoperatively, estradiol and testosterone implants were replaced with ones designed to maintain the lower hormone levels present following parturition (0.02 ng/ml; 0.6 ng/ml, respectively). Progesterone was not replaced. At the first aggression test, females with estradiol and testosterone alone displayed significantly more aggression than females with these hormones plus progesterone. Both groups were more aggressive than females without hormone replacement. Following the exchange of large implants for small ones, females that previously had progesterone increased in aggression while females that previously had only estradiol and testosterone decreased in aggression. Both groups continued to be more aggressive than the group without hormone replacement. High serum progesterone present near the end of pregnancy appears to moderate the expression of aggression supported by estradiol and testosterone. Conversely, progesterones decline at parturition appears to produce a rebound facilitation of aggression even though serum estradiol and testosterone simultaneously decline.

Activation of aggression in female rats by normal males and by castrated males with testosterone implants

Female hooded rats were continuously housed with an intact male, a castrated male with subcutaneous testosterone implants, or two other females. At weekly intervals over a 10-week period, the cagemate(s) and pups were removed and aggression by the female toward an unfamiliar female intruder was observed over a 15-min period. On the 11th week each female was subjected to this intruder test in an unfamiliar cage. On the 12th week, a final test was conducted in each females living cage with a male rather than a female as the intruder. The aggressive behaviors recorded were attacks, bites, on-top, and piloerection. Females housed with normal males displayed a significant increase in aggression prior to parturition. Their aggressiveness persisted through the 10th test with peaks at parturition and the start of lactation. Females housed with castrated males also displayed significant increases in aggression but without the peaks associated with parturition and lactation. Their aggressiveness also persisted throughout the test period. Females housed with other females showed a small increase in aggression over weeks. All groups showed virtually no aggression in the unfamiliar cage. All females displayed some aggression toward a male intruder but the level of aggression was highest in maternal females. The results demonstrate that aggression qualitatively similar to that displayed following parturition and during lactation can be elicited in nulliparous females.

Hormone-dependent aggression in the female rat: Testosterone plus estradiol implants prevent the decline in aggression following ovariectomy

Female rats were individually housed with a sterile male for the duration of the experiment. Beginning 7 to 10 weeks after the start of cohabitation, each female was tested for aggression toward an unfamiliar female at weekly intervals for 3 weeks. Females that displayed consistent and substantial aggression were given one of the following treatments: ovariectomy followed by both testosterone and estradiol implants, ovariectomy followed by 2 empty implants, or sham ovariectomy followed by 2 empty implants. The implants were subcutaneously placed hormone-filled Silastic capsules. They were expected to produce a serum testosterone concentration of 0.5 ng/ml and an estradiol concentration of 15 pg/ml. Postoperatively, the aggression of each female continued to be assessed on a weekly basis for 3 weeks. Ovariectomized females with hormone implants displayed a level of aggression postoperatively similar to that of sham-ovariectomized females and significantly greater than that of ovariectomized females with empty implants. These results, together with others, suggest that estradiol and testosterone act together to form the hormonal foundation of hormone-dependent aggression by females cohabiting with a sterile male.

Competitive experience activates testosterone-dependent social aggression toward unfamiliar males

Male hooded rats (350 to 450 g) were castrated and given subcutaneous implants of testosterone-filled or empty Silastic tubes. Four weeks later, half of the animals with testosterone implants were housed with an animal with an empty implant and left for 6 weeks. The other animals were adapted to a food-deprivation schedule, housed in testosterone-implant/sham-implant pairs and given a series of food-competition tests. Following the competition tests, all animals were observed individually in their living cage for aggression toward an unfamiliar intruder. Within the competitive situation, animals with testosterone implants were more aggressive and more successful at maintaining access to food than their cagemates with sham implants. In the unfamiliar intruder test, animals with testosterone implants that had been subjected to food competition were more aggressive toward an unfamiliar intruder than were animals with testosterone implants that had not been given competitive experience. Animals with testosterone implants given competitive experience were more aggressive than their castrated cagemates, but animals with testosterone implants not given competitive experience were not more aggressive than their cagemates. These results demonstrate that testosterone-dependent social aggression fostered by a competitive situation is elicited by an unfamiliar male intruder. They also confirm other evidence that activation of social aggression does not appear to require increased testicular testosterone secretion.

Testosterone supports hormone-dependent aggression in female rats☆

Female hooded rats were ovariectomized and implanted with a single testosterone-filled Silastic tube or an empty tube. The tube size was one which allowed a release of testosterone at the high end of the mean normal serum testosterone concentration for intact females. Following a 7-day recovery period, all rats were placed on a 23-hr food-deprivation schedule and adapted to a highly palatable liquid food over a 5-day period. Each animal with a testosterone implant was then housed with an animal of similar weight but an empty implant. The pairs were subjected to a series of 3 restricted-access competition tests (1/day) followed 4 days later by a series of 3 free-access competition tests. The animals were then separated, adapted to a bland liquid food, and paired with new partners. They were then subjected to the restricted- and free-access food-competition tests but with bland food as the incentive. During the first 6 competition tests there were no significant differences between groups in aggression or in time spent licking at the food spout. During the second series of tests, females with testosterone implants were more aggressive and more successful at maintaining access to the food than were their competitors with empty implants. The difference between groups occurred during the free- as well as the restricted-access tests. The effectiveness of physiological levels of testosterone in supporting aggression is attributed to the use of a test situation that activates as well as elicits hormone-dependent aggression. These results suggest that testosterone may be the hormonal substrate for hormone-dependent aggression in female rats.

Medial accumbens lesions attenuate testosterone-dependent aggression in male rats

Male hooded rats were castrated and implanted with testosterone-filled Silastic tubes appropriate for maintaining a normal average serum testosterone concentration. They were then given lesions of the medial accumbens nucleus or sham lesions. Twenty-four hours postoperatively each male was housed with a female. Beginning 7 days following pairing and continuing once each week for 4 weeks, each lesioned or sham-lesioned male was observed for aggression toward an unfamiliar male intruder. On the day following each test of aggression toward an unfamiliar male, each lesioned and sham-lesioned male was assessed for defensiveness toward an experimenter. Rats with medial accumbens lesions displayed significantly less aggression toward an unfamiliar male intruder during each of the weekly tests than did sham-lesioned animals. The attenuation was most pronounced in animals with lesions damaging the posterior part of the medial accumbens nucleus (also designated as anterior portion of the bed nucleus of the stria terminalis) in the region of the crossover of the anterior commissure. Although medial accumbens lesions are known to make individually housed rats hyperdefensive toward an experimenter, lesion-induced hyperdefensiveness was not observed in the pair-housed animals in the present experiment. It is argued that the medial accumbens/bed nucleus of the stria terminalis area is an important region in the anterior forebrain for the modulation of hormone-dependent aggression.