||Comments by Burr Eichelman
William S Middleton Memorial Veteran Hospital, University of Wisconsin, Madison, Wisc. 53705
Brain mechanisms of aggression: Dilemmas of perspective. Adams has made a significant contribution to the aggression literature with his attempt to catalogue various experiments in terms of their modulation of offense, defense, and submission. The attention to behavior as offense, defense, or submission has been overly long in coming to the aggression literature. This categorization must be taken into account by current researchers and integrated with Moyer's (1968) groupings and Reis's (1974) separation ot affective and predatory aggression. It is important for defensive behaviors to be included within the rubric of aggression, since many defensive attacks inflict severe physical damage.
The attempt to categorize aggressive behavior as offensive, defensive, or submissive has limitations. The behavioral repertoire for a given animal in an agonistic encounter may encompass all three behavioral groupings. Intermale aggression in previously isolated mice illustrates this point. Both mice may initially tail-rattle, labeled by Adams as defensive behavior, but they both also move in and out ot offensive sideways and upright postures and the bite-and-kick attack. Only over time. as a dominance is established, does the distribution of the behaviors change, Similarly, in the model of shock-induced fighting in the rat, it becomes perhaps a semantic argument to determine whether an increased frequency of boxing (noted as a component of both defensive and submissive behavior) in this paradigm is an increase in defensive behavior (lacking the lunge-and-bite attack) or an increase in submissive behavior (eg, in Rodgers and Brown 1976). And there are further subdivisions of these groupings which need to be made. Beleslin and Samardzic (1977) describe behaviors induced in the cat with muscarin and carbachol, which they divide into fear and irritable aggressive behaviors, yet both of these seem to be represented by Adams's defense category.
In relation to neurophysiology and the integration of multiple experiments, Adams himself notes inherent limitations in his paper. Some of these limitations should be underscored, though they need not detract from the utility of the paper in cataloging data and generating new experiments. First, the brain mechanisms proposed are essentially "one-way" mechanisms of either inhibitory or excitatory neuronal systems. Reciprocal inhibitory systems and cross-talk between various brain regions is acknowledged but finds little expression in the conceptualization. Not only are neuronal systems bidirectional between various regions, but they also include multiple neurotransmitter systems (e.g. the nigro-striatal system with reciprocal interaction involving dopamine, acetylcholine, and GABA). Secondly, the model deals with aggressive behaviors induced with the onset of brain activity. There is a whole class of behaviors in which both the environmental and neural antecedents are the cessation ot a stimulus. Aggression (defense) can be induced by morphine withdrawal (Boshka, Weisman, and Thor 1966); or by extinction in an operant paradigm (Azrin, Hutchinson, and Hake 1966); or by the cessation of reinforcing brain stimulation (Hutchinson and Renfrew 1978). This class of behaviors needs amplification within Adams's conceptual framework.
The lesion and stimulation studies reviewed must also be subjected to some cautionary interpretation. Most of the stimulation studies relating to defensive behavior can be interpreted on the basis of inducing pain with the concomitant induction ot boxing or other defensive postures, much the same as with footshock-induced fighting. Perhaps certain operant methods might be used to assure that defensive aggressive behavior induced by central grey stimulation was not the animal's response to centrally induced pain. Conversely, lesion effects must also be cautiously interpreted when large lesions in the brain stem are required to abolish defensive behavior. Careful attention should be given to the specificity of these lesions in relation to other behaviors, particularly in terms of activity level and general motor coordination. Some of the studies'reported by Adams have made such attempts, while others have only superficially described the general state of the lesioned animals. Finally, within the cataloging of experiments as carried out by the author, there is some selectivity with respect to data contrary to the thesis expressed. One such example is noted regarding the cingulate cortex. Adams suggests that lesions of the cingulate should increase defense, and he cites such an example in the cat, However, lesions of the cingulate cortex in the rat decrease the frequency of shock-induced fighting composed of defensive postures (Blanchard and Blanchard 1968; Eichelman 1971).
The neural mechanisms of longitudinal neural circuits as set forth by Adams do not encompass another alternate (and equally biased) view of the central nervous system often utilized in pharmacology. Such a view deals with central "tone" or levels of excitability and irritability. The widespread distribution of noradrenergic neurones originating from the locus coeruleus can aid in illustrating this view. Stimulation of the locus and firing of its cells must have effects that involve many brain nuclei and regions. Further, such "tonal" modulation may even involve nonneural elements. Glial cells have been reported to contain binding sites for putative neurotransmitters and pharmacologic agents such as diazepam (Henn and Henke 1976) These elements may significantly influence whether any aggressive behavior occurs, or they may provide an alternative "consociate modulator" system much more diffuse than the postulated VMH entity of Adams's thesis.
Lastly, the brain mechanisms postulated within Adams's paper, and the experiments that they suggest, remain preponderantly within the area of preclinical, nonhuman research. For the clinician it is very difficult to conceptually transfer data relating to vibrissae and defensive behavior to patients with a diagnosis of Explosive Personality Disorder. Considerably more conceptualization must be applied to human aggressive behavior to meaningfully integrate this murine and feline model with clinical experience. Much more ethological work must be carried out to categorize human agonistic behavior as offensive or defensive. Clinicians have instead tended to describe dystonic human aggressive behavior in terms of impulsivity (O5M III, 1978) or an inability to inhibit aggression (perhaps a deficit in passive avoidance), rather than in terms of offensive or defensive behavior patterns. Human behavior is also much less accessible to modulation by brain stimulation or lesioning. Thus the transfer of knowledge from the rodent or cat to man may be severely limited; this limitation may even extend to the asking of similar questions in the human. One is hard pressed to draw parallels with limbic function when septal lesions or stimulation in the rodent induce marked behavioral change in contrast to minimal effects in the primate or human. It may be that more productive modeling for clinical research will develop from the analysis of environmental antecedents to aggression and perceptual differences between eggressive and nonaggressive individuals (e.g. Kinzel 1970), or through pharmacological manipulations, both as tools for understanding the central nervous system and as beneficial modulators of dystonic behaviors.