Results of the present study, along with previous findings on the midbrain central gray and its role in affective defense in the cat, indicate that the central gray is a nodal point in the neural organization of pain-induced defensive boxing. It is situated at the confluence of sensory and motor systems critical for the behavior, and it appears capable of acting independently of the forebrain. These data suggest that it may function as a modal command system in the sense of Kilmer and McCulloch  or as a patterning mechanism in the sense of Flynn  in the generation of pain-induced boxing.
The central gray receives projections of the sensory systems involved in defensive boxing. Extensive terminations of the paleo-spinothalamic tract, which carries pain information from the spinal cord, enter the gray from its lateral margins at the level of the superior colliculus . Many cells in the central gray respond primarily to painful peripheral or nerve stimulation, presumably as a result of input from this tract . Tactile information, particularly from the face and presumably the vibrissae, is strongly represented in the central gray , although the pathways have not been delineated. It is also possible that visual information reaches the central gray through the adjacent superior colliculus.
The major output fibers from the central gray exit via its lateral margins, either in directly radiating fibers (Weisschadel's radiations) or in conjunction with commissural fibers of the superior colliculus [7, 13], and descend diffusely into the reticular formation where they have not been traced further. It is likely that the interruption of these lateral output pathways along with the lateral input pathways of the midbrain central gray by the lesions of Kesner and Keiser  produced the deficits they observed in the defensive boxing of rats.
Although the central gray also has extensive connections with the forebrain, which could theoretically supercede the central gray in the neural organization of pain-induced defensive boxing, data from the present study indicate that these connections are probably not necessary for the behavior. Lesions anterior and ventral to the central gray which completely severed the dorsal longitudinal fasciculus and the fibers from the central gray to the ventral tegmental area of Tsai, respectively, did not abolish boxing. According to anatomical studies, these are the two primary ascending pathways from the central gray at the level of the superior colliculus [7, 13]. Similarly, these lesions interrupted the major descending pathways from the forebrain to the central gray. Lesions anterior to the central gray interrupted the descending components of the dorsal longitudinal fasciculus which is the major pathway from the forebrain to the central gray . Lesions ventral to the central gray interrupted fibers from the fornix which bypass the mammillary bodies and distribute to the anterior central gray, fibers from the stria medullaris and habenula which travel with the habenulo-interpeduncular tract and then leave it to travel through the midbrain and distribute to the caudal central gray, and fibers of the medial component of the midbrain extension of the medial forebrain bundle . The only remaining pathways from the forebrain to the central gray in these animals were the lateral component of the midbrain extension of the medial forebrain bundle and possible indirect connections via the superior colliculus or diffuse pathways in the reticular formation.
In the cat it has been shown that the forebrain is not necessary for affective defense behavior. Although the behavior was undirected and had a rather high threshold for elicitation, affective defense has been obtained in chronic decerebrate preparations [3, 14]. Comparable studies have not yet been attempted in the rat.
(Continued on next page)