Brain Mechanisms

Offense and defense vs. rage and fear: A matter of semantics? The manner in which events are classified determines the manner in which they are studied. Adams's conceptualization that agonistic behaviors are controlled by distinct offense, defense, and submission motivation systems is capable of organizing a great deal of neuropsychological data and accordingly provides a neurotaxonomy that begins to combine distal evolutionary and proximal neurophysiological causes into a viable model of aggressive behavior. The categorization of agonistic behaviors with reference to the deep structure of brain circuits is probably more useful for understanding aggression than taxonomies that focus merely on eliciting conditions or on specific kinds of aggressive acts. Different eliciting conditions (releasing stimuli) probably converge upon common control systems, and these intervening "command systems" can probably initiate a number of related behavior sequences. which are then filtered and molded by their outcomes.

In conceptualizing discrete offense, defense, and submission systems, Adams is concerned that the postulation of separate defense and submission systems may be gratuitous; but in the end he remains convinced ot the utility and reality of the distinction. I am not. Might not offense and defense suffice as primary systems, with submission being merely a conditional, and hence a diminutive, form of defense, in the same way that anxiety may be a learned diminutive form of fear? In Adams's system, submission becomes a separate category, because the defense system (as used by Adams) can instigate certain offensive gestures (lunge and bite attack), and hence a separate category is required for the tendency of animals to remain totally defensive. However, if one assumes that defensive threat and aggression - which can be evoked from cornered animals or induced by brain stimulation (so-called affective attack) - take the form they do because both defense (fear?) and offense (rage?) systems are concurrently activated, then the need for a separate "submission motivation" system disappears.

I believe that the available data concerning aggressive behavior can be handled as readily by two primary motivational (or command) systems as by three. For instance, the argument that medial hypothalamic lesions increase defense and reduce submission, while amygdala lesions decrease defense and increase submission, may be semantically recast to conclude that medial hypothalamic lesions increase both offense and defense (leading to a high incidence of affective attack or defensive threat), while amygdala lesions reduce offense (hence all forms of aggression), leaving defense (and hence submission) processes without opposition. Of course, it must be emphasized that, within such a conceptualization, defense has a somewhat different connotation (namely fear) than might be agreed to by Adams.

Although I take issue with the distinction between defense and submission, I strongly agree with Adams's postulation of a "consociate modulator." All specific social acts are expressed in broad affective contexts - of social comfort or social unease - and through some kind of neural representation of such processes, whether they be called "consociate modulators" or simply "social affect" (Panksepp et al, 1978), the probability of offensive or defensive emotions is heightened or diminished. Although Adams's system is useful for analyzing aggressive behavior, on a more general level, the proper taxonomy for the basic neural systems that subserve instinctual emotive behavior patterns remains an open issue. Thus, one potential problem with Adams's taxonomy is its exclusive focus on emotive behaviors analyzed in the context of aggression. Potential behavioral controls exerted by these same systems in nonsocial contexts are ignored. Although it is possible that these systems act solely or primarily in the context of aggression, my opinion is that offense and defense systems are really generalized emotive systems that can control behavior in a variety of situations. Accordingly, I have recently used the label "rage sensory-motor command system" and "fear sensory-motor command system" (Panksepp 1980) to describe what t think Adams is referring to in his "offense" and "defense motivational systems" [see also Kupfermann & Weiss: "The Command Neuron Concept" BBS 1(1) 1978].

My use of everyday emotional terminology is specifically intended to convey the possible generalized roles of hard-wired emotive systems in the brain. Thus, I would assume that the neural representations of "frustration"(i.e. the failure of expectancies to be met) would "enrage" animals and thus activate the same primitive command circuits that are active when a male rat intrudes on the homeground of another male. This kind of seemingly minor semantic distinction can have marked experimental repercussions. For instance, rather than simply looking at the activity of ventrolateral hypothalamic cells during aggression, one might also be led to look at their activity during extinction of appetitive tasks, with the prediction that certain cell populations that fire when rats attack intruders would also be active during extinction. Simply viewing such cells as part of an "offense" circuit may not lead as readily to such a broader analysis.

To take another example I would predict that threat to the bodily integrity of an animal, whether arising from a laboratory shock-generator or the blood-lust of a predator, would trigger common limbic command systems. A "fear" designation for such a system could, I think, more readily coordinate experimental results derived from different threatening situations. I think such distinctions are important, and under a different verbal disguise they have led to the very real controversy as to whether the hypothalamus contains "specific" control systems that orchestrate behavioral acts or .'nonspecific " emotive command systems for biassing classes of behavioral tendencies (see Valenstein 1973) Although this controversy still remains unresolved, my evaluation of the evidence is that the limbic "command" systems that can sustain "stimulus-bound" behaviors are generalized emotive circuits (Panksepp 1971; 1980). Since Adams analysis is implicitly based on the existence of such "stimulus-bound" behaviors, it might be useful if he clarified his stance on the issue.

In the final accounting, what matters in the selection of terms is the success of the research that various conceptualizations generate. If the underlying brain processes are nonspecific, then it would be useful if the labels for the systems reflected the nature of the generalized processes rather than a more limited class of behaviors that can be provoked. Of course, these issues can only be resolved empirically, but I think the nonspecificity notion can accommodate more of the existing data then approaches that seek a more articulated phrenology in the executive systems for emotional behaviors (Panksepp 1980). In any case, our own working hypothesis is that the limbic "command" systems that sustain "stimulus-bound" behaviors are normally active during broad classes of environmental events that have required similar types of adaptive responses in the evolutionary history of the species. Thus, emotive "command" systems may activate sets of related behaviors in a variety of situations, and contiguous reinforcement processes may provide the selective force for determining which acts become habits in the behavioral repertoire of animals. Thus, I see submission as a habit that arises from repeated activation of defense systems, as well as perhaps from repeated nonreinforcement or punishment of offensive gestures. Rather than being a primitive behavior control system in its own right, submission may be the final behavioral outcome of the consequences of offensive and defensive acts.

In any case, I think Adams's contribution is a substantial step in the right direction. Brain systems are provided with conceptual tags that can guide further work into the deep functional structure of neural circuits controlling emotive behavior. However, I would be tempted to take an even larger step in the same direction - to see whether we can make sense of behavior by categorizing the command systems that appear to run through the limbic system in terms of certain old emotional constructs that have been avoided in behavioral research for the better part of this century. Although the surplus subjective connotations of terms such as rage and fear can easily lead to anthropomorphic excesses, I suspect that such a set of terms can help us come close to the reality of limbic organization than more conservative concepts. Indeed, by thus identifying and labeling neural systems in terms of distal evolutionary causes rather than ongoing neurophysiological processes, we may begin to make better sense of the myriad facts that the fabric of the brain continues to yield in abundance.

ACKNOWLEDGMENT: Supported by NIMH Research Scientist Development Award MH-00086.