cited in Miczek and Donat, 1989). A series of recent studies in squirrel monkeys and rhesus monkeys with chronic administration of the hallucinogenic "designer" amphetamine derivatives "ecstasy" and ''eve," which have profound cytotoxic actions on brain 5-HT and markedly lower CSF 5-HIAA, has revealed no evidence for increased violent activity (e.g., Ricaurte et al., 1985, 1989; Molliver 1987).

To interpret the significance of CSF 5-HIAA data, the relative contribution of anatomically distinctive pools of 5-HT neurons in different brain regions to CSF metabolite levels needs to be determined. Seasonal and circadian rhythmicity, activity levels, and nutritional status, in addition to the propensity to engage in aggressive behavior, are among the prominent determinants of synthesis and metabolism of 5-HT and 5-HIAA. These variables are not appropriately reflected in single measurements at a single time.

During the late 1970s, Brown, Goodwin, and their associates reported that two samples of institutionalized Navy men showed an inverse relationship (r =-.77, -.78) between ratings of a life history of events that were thought to reflect "aggression" and CSF 5-HIAA (Table 7; Brown et al., 1979, 1982). These aggression ratings were also related to the number of suicide attempts, which—in an earlier influential Scandinavian study—were found to be inversely related to the concentration of 5-HIAA in CSF of patients with unipolar depression (Asberg et al., 1976). Low CSF values for 5-HIAA were subsequently measured in certain samples of alcoholic violent offenders with "impulsive" personality, homicidal men (Linnoila et al., 1983; Lidberg et al., 1985), and impulsive arsonists (Virkkunen et al., 1989a). These latter studies led to the proposal that low CSF 5-HIAA concentration reflects a disorder of poor "impulse control rather than aggressiveness or violence as such" (Virkkunen et al., 1989a; Linnoila et al., 1989).

So far, low 5-HIAA in CSF has been inversely correlated not only with the original subpopulation of violent suicide attempters (Asberg et al., 1976), but also with clinician-rated or self-reported ratings of a life history of aggression (Brown et al., 1979, 1982; Linnoila et al., 1983; Lidberg et al., 1985); Rorschach ratings of hostility and anxiety (Rydin et al., 1982); outwardly directed hostility (Van Praag, 1982; Roy and Linnoila, 1988); criminal behavior (Linnoila et al., 1983; Lidberg et al., 1985; Van Praag, 1982; Virkkunen et al., 1989a,b); self-reported behavior problems during

childhood (Kruesi et al., 1990); and preoccupation with violent thoughts (Leckman et al., 1990). The correlation coefficients usually range between r =-.46 and r =-.78 (Table 7). There is considerable controversy in interpreting the significance of 5-HIAA measurements from CSF (e.g., Eriksson and Humble, 1990). It is difficult to account for the rostrocaudal gradient in 5-HIAA concentration: the highest concentration is in the lateral cerebral ventricles, but most measurements stem from lumbar regions of the spinal cord. How exactly CSF 5-HIAA concentrations relate to the anatomically differentiated 5-HT neuronal pathways and how this measure reflects 5-HT turnover or monoamine oxidase activity remain to be specified.

CSF 5-HIAA and MAO in blood platelets show a weak correlation—in some samples a positive correlation and in others a negative one (e.g., Asberg et al., 1987). It is unclear whether and how platelet MAO-B on the one hand, and A- and B-type MAO in brain are related. Ellis (1991) has summarized the correlative studies suggesting that low platelet MAO levels may serve as markers for suicidal behavior, increased sensation seeking, impulsiveness, childhood hyperactivity, alcoholism, and criminality. However, many failures to replicate these reports and the lack of relationship between blood and brain measures need to be resolved before low platelet MAO and low CSF 5-HIAA can be accepted as biological markers for violence (Asberg et al., 1987).

Alterations in 5-HT activity have also been related to aggressive behaviors by resorting to measurements of blood chemistry (Table 3, section C). These peripheral indices include whole blood levels, plasma tryptophan/neutral amino acid ratio, and binding of tritiated imipramine to blood platelets. One study reported an inverse relationship between plasma 5-hydroxyindole and hyperactivity/aggression in 24 mentally retarded patients (Greenberg and Coleman, 1976). A further study reports that 11 out of 15 male outpatients seeking treatment for frequent bouts of verbal and physical aggression had slightly lower 5-HT uptake into blood platelets than matched controls (Kent et al., 1988; Brown et al., 1989). Although one study found a positive correlation between plasma 5-HT and conduct ratings in adolescent males (Pliszka et al., 1988), another study reported no difference in whole blood 5-HT between violent offenders and normal controls (Virkkunen and Narvanen, 1987). A modest inverse correlation between imipramine binding to 5-HT uptake sites on blood platelets and parent-rated aggressiveness in children with conduct disorder was also reported (Stoff et al., 1987). The interpretation of these peripheral