Allan F. Mirsky and Allan Siegel

Over the past four decades, there has been an increasing body of data in the human literature on neuropsychiatric disorders that raises the question about a possible relationship between the abnormal function of specific regions of the brain and the occurrence of violent and aggressive behavior. The view that violence and aggression are human behaviors symptomatic of an underlying brain disorder, rather than simply acts to be punished under law, is relatively new. It is true that the distinction "between the harmful act that was traceable to fault and that which occurred without fault" extends back to ancient Hebrew law (American Bar Association, 1983:7–271). However, the scientific facts that have been offered as the basis for what Monroe has referred to as "the episodic dyscontrol syndrome" (Monroe, 1970) or other involuntary acts, are of relatively recent origin. The region of the brain most often linked with this form of behavioral dysfunction is referred to as the ''limbic system." Research on the limbic system (Figure 1) identified an apparently unitary cerebral region or limbic lobe (Broca, 1878) at the juncture of the forebrain and brain stem, which Papez (1937) and MacLean (1952) later identified as

FIGURE 1 Diagrammatic representation of the principal subcortical connections of the limbic system viewed from the mesial surface of one hemisphere. Important connections to the brain stem reticular formation have been omitted, and others are only approximately represented. Some abbreviations: A, anterior nucleus of thalamus; Ant., anterior; AMYG, amydala; Int., intermediate; CM, center median; Sens., sensory; Occip., occipital; G., gyrus; Stria Ter., stria terminalis; St. Med., stria medullaris; MD, medial dorsal nucleus of thalamus; O.B., olfactory bulb; Ling., lingual; Temp., temporal; Hypoth., hypothalamus; MM, mammillary bodies; Parolf., parolfactory; H., habenular nucleus. SOURCE: Penfield and Jasper (1954).

the cerebral substrate of emotional behavior. The work of Klüver and Bucy (1939) and of Rosvold et al. (1954) is also relevant here. These researchers demonstrated that monkeys surgically deprived of portions of their limbic system had major changes in their social and affective behavior. These studies had a significant impact on the thinking about the relation between cerebral structures and/or systems and abnormal behavior, including a number of neuropsychiatric disorders. In terms of its overall organization, the limbic system includes the hippocampal formation, amygdala, septal area, cingulate gyrus, and prefrontal cortex (according to some authors). Several other brain structures have been considered

part of this system because of their neural associations with limbic structures. These include the hypothalamus and midbrain periaqueductal gray matter (PAG). Collectively, these regions comprise a functional unit that is sometimes referred to as the "limbic-hypothalamic-midbrain axis." The importance of this research was that there was a brain system that could be implicated in neuropsychiatric disorders, including those in which violence or aggressiveness was a major symptom. This system could be a focus of research, both clinical and laboratory based, and could provide the target or basis for new treatment possibilities. Some of this research is reviewed in the section of this paper on human studies of aggression and violence.

In 1974, one of the authors reviewed the literature on the relationship between aggressive behavior and brain disease, and concluded that the available evidence did not support the view that aggressive behavior in humans could be attributed to brain disease (Mirsky and Harman, 1974). The question arises as to whether, in the ensuing 16 years since that paper was written, sufficient additional data have been gathered to alter that conclusion. Recent studies related to that question are also reviewed below.

In view of the putative relationship between brain dysfunction and aggressive behavior, it is the purpose of this paper to summarize briefly the neural mechanisms of aggressive behavior as discovered from animal models, and to review human studies on the relationships among brain dysfunction, neuropsychiatric disorders (including abnormal development), cognitive processes, and the symptoms of violence and aggression. The section that deals with animal models focuses on two behaviors that can be readily elicited in the cat: quiet biting "predatory" attack and affective defense. It is our belief that the neural substrates and mechanisms underlying these distinctive forms of aggressive behavior in the cat may also regulate aggressive reactions at the human level or, possibly, provide a framework for understanding human violence and aggression.

In the thesis advocated here, an analogy can be made between the relationship of the limbic system to the hypothalamus and midbrain PAG and that of the motor cortex and reticulospinal fibers to the spinal cord concerning the modulation of "emotional" and "motor" systems, respectively. With respect to motor systems, the final common output pathways for the expression of motor responses such as fine movements of the extremities or walking movements are governed by the activity of cells located