Inhibition of GABA transaminase, mostly by valproic acid, decreased attack behavior by resident male mice toward intruders (DaVanzo and Sydow, 1979; Puglisi-Allegra and Mandel, 1980; Haug et al., 1980; Poshivalov, 1981; Sulcova et al., 1981; Puglisi-Allegra et al., 1979, 1981; Simler et al., 1983), concurrent with an increase in GABA levels in several brain areas, particularly in olfactory bulbs and striatum. Also, GABA or valproate, given either systematically or directly to the olfactory bulbs, blocked mouse-killing behavior (Mack et al., 1975; Mandel et al., 1979; Molina et al., 1986), but a GABA agonist can induce killing behavior (Depaulis and Vergnes, 1983).

Defensive behaviors by rats are enhanced by GABA receptor antagonists such as picrotoxin administered either systemically, intracerebroventricularly, or directly into the periaqueductal gray area; inhibition of GABA degradation does not increase defensive behavior (Rodgers and Depaulis, 1982; Depaulis and Vergnes, 1985, 1986).

Given the ubiquity of GABA in many brain areas and its role in various physiological and behavioral functions, particularly in motor and convulsive disorders, it is not altogether surprising that the interpretation of the evidence on GABA's role in different kinds of aggressive behavior ranges from inhibition to facilitation. The classic view attributes an inhibitory role to GABA, particularly with regard to mouse killing in rats, attack behavior by isolated mice, and defensive reactions in rats (e.g., Mandel et al., 1979, 1981). It is now evident that agonists at benzodiazepine receptors lead to increased GABA transmission due to activation of GABA_A receptors, which in turn increases chloride flux causing hyperpolarization. The data on benzodiazepine-type anxiolytics (see below) increasing and decreasing aggressive behavior in the contexts of maternal defense, dominance, and territorial fighting in animals and in several clinical cases of violent outbursts point to a modulatory role of GABA_A receptors.

Early evidence implicated brain acetylcholine in the induction of killing behavior in laboratory rats and in "rage" and defensive reactions in cats and rodents (see reviews by Romaniuk, 1974; Allikmets, 1974). However, evidence on brain acetylcholine and human aggression and violence is chiefly limited to the effects of nicotine (e.g., Bell et al., 1985, Eichelman, 1986, 1987).

Cholinergic agonists at muscarinic receptors such as arecoline,

carbachol, muscarine, or acetylcholine when given directly into certain forebrain structures evoke rage-like and defensive responses in male and female cats (e.g., Hernandez-Peon et al., 1963; Grossman, 1963; Baxter, 1968a; Beleslin and Samardzic, 1979; Brudzynski, 1981a,b); these responses are blocked by muscarinic antagonists such as atropine or scopolamine (see Table 5). A second set of findings links activation of muscarinic receptors to killing behavior in laboratory rats and domestic cats (Table 5). Systemic administration or intrahypothalamic injection of muscarinic agonists or acetylcholinesterase (AChE) inhibitors induces animals to kill (e.g., Bandler, 1969, 1970; Smith et al., 1970; Berntson and Leibowitz, 1973). Killing behavior, whether induced by cholinergic agonists or part of the animals' repertoire, is blocked by antimuscarinic drugs (see Table 5). A third set of studies reports on increased pain-induced defensive responses in rats that have been given carbachol or the AChE inhibitor physostigmine in the lateral hypothalamus or basolateral amygdala (Rodgers et al., 1976; Bell and Brown, 1980).

It is not too surprising that comparable data on heightened aggressive or violent behavior in humans after muscarinic receptor activation have not been forthcoming. These substances are too toxic and have potent effects on many autonomic functions that are severely compromising.

Nicotine has been found to exert relatively specific antiaggressive effects in several animal species. A subcutaneous (s.c.) "smoking dose" of nicotine (25/µ/kg) specifically reduced aggressive acts and postures in rats, and pain-induced biting and postures in squirrel monkeys and rats (Driscoll and Baettig, 1981; Emley and Hutchinson, 1983; Waldbillig, 1980). Nicotine also decreased killing behavior by rats and cats (Bernston et al., 1976; Waldbillig, 1980), although predatory killing by ferrets remained unaffected (Meierl and Schmidt, 1982).

Experienced aggressive-type smokers titrate their nicotine intake in part to reduce their anger, and withdrawal from smoking is associated with increased hostility and irritability (e.g., Bell et al., 1985). Nicotine cigarettes decrease experimental measures of aggression in a competitive task for human subjects (Cherek, 1981, 1984). The therapeutic potential of nicotine in the control and management of aggressive and violent behavior has not been dissociated from the enormous health risks that are associated with tobacco smoking.