Updated Review-6

The previously proposed separation of defense and submission motivational systems has not been supported by subsequent research. Hence, as mentioned in the previous paragraph, the present model recognizes only one defense motivational mechanism with two differentiated pools of neurons mediating anti-predator and consociate defense respectively. This corresponds to the distinction by Blanchard et al. (1998) between anti-predator and conspecific defense, except for the term "conspecific." I prefer "consociate" on the grounds that the results of cross-fostering or co-habiting animals with other species can produce similar results, suggesting that the distinction is at least partly determined by experience. As a result of this change, the new model is considerably simpler and more elegant.

Anti-predator defense, as distinguished by Blanchard et al. (1998), includes "flight (followed by avoidance), freezing, and defensive threat (including sonic vocalizations) and attack. Specific postures (e.g. the upright posture) may occur but are less systematic than are those seen to conspecific [consociate] attack." They also suggest "a stronger role for learning in the defensiveness of laboratory rodents to attacking conspecifics than in reactions to a predator" (Blanchard et al., 1998). Although they do not make the comparison, one may suppose that the distinction between anti-predator and consociate defense corresponds to the difference between wild rat defense and laboratory rat defense, respectively, which we observed in our laboratory and which is described in detail by Blanchard et al. (1986): "One important difference ... was the magnitude of defensive behavior. Although both wild and laboratory rats reacted defensively when attacked, it was notable that the wild rats placed as intruders into established colonies showed consistent defense before being attacked. The defensive reactions of wild rats include baring of the teeth, shrieking vocalization, and jump-attack, a leaping bite directed toward the head of an opponent ... wild rat defense was so superior to that of laboratory animals that wild rats placed as intruders into laboratory rat colonies were seldomly successfully bitten despite many attempts by the attackers."

The distinction between anti-predator and consociate defense applies as well to carnivores and primates as well as muroid rodents. In our laboratory study of macaques (Adams, 1981) we could observe two levels of defense, one more submissive [consociate] and another more assertive and potentially damaging [anti-predator]. Among carnivores, submissive and anti-predator defense are particularly easy to distinguish in the behavior of dogs and wolves.

Anti-predator defense in muroid rodents shares most of the motor patterns of consociate defense, but is more intense, especially with regard to the damaging motor pattern of lunge-and-bite attack. Anti-predator defense may include the full submissive posture (Monassi et al., 1999) and 22 kHz ultrasound cries (Blanchard et al., 1991). In the 1979 review (Adams, 1979a) the latter was considered to be unique to submission, i.e. consociate defense. The inputs of anti-predator and consociate defense are apparently similar for the most part, but there must be at least one input that can modulate the level of defense, making it less damaging, especially the lunge-and-bite attack. As in the 1979 review this function is attributed to a "consociate modulator."

The motivational mechanism for defense, both anti-predator and consociate, consists of a differentiated pool of neurons in the dorsal midbrain central gray and adjoining dorsal midbrain. Data provided to support this assertion were reviewed previously (Adams, 1979a), and subsequent research has tended to confirm the analysis. The fundamental data come from neuronal recording during defense defense against a second attacking cat (Adams, 1968a, 1969b) and during shock-elicited fighting in the rat (Pond et al, 1977). In most respects the results were similar. In the cat almost half of the cells in the dorsal and dorsolateral central gray and region immediately lateral responded maximally during affective defense. In the rat slightly more than half of the cells in the same region responded maximally during shock-induced fighting. There was one difference in the results: in the cat, 4 of the 8 neurons in the defense region that fired maximally during affective defense were almost silent otherwise, (one fired once when the partition was opened between the animals but no fighting occurred and another fired when the cat was pinched and attacked the experimenter). This was not found in the rat study. This could reflect a difference in the neural mechanisms between rats and cats, but it seems more likely to have been a function of the different testing situations employed which may have produced different background levels of defense.

Single neuron responses of the midbrain central gray also have been recorded in rats in response to electrical stimulation of the brain, but since the testing was done under anesthesia, the results could not be related to naturally-occurring defense behavior (Sandner et al., 1979, 1982, 1986).

While less direct than microelectrode recording, many studies have measured c-Fos immuno-reactivity of cells of the midbrain central gray as a function of defense behavior, as well as differentiating functions of dorsal from ventral central gray by lesions and identifying neurotransmitters influencing cells in this region. In c-Fos studies with behavioral controls, Kollack-Walker and colleagues compared and found rates of activity in the dorsal periaqueductal gray of hamsters higher during defense than during mating behavior (Kollack-Walker and Newman, 1995) and higher during defense than offense in one study (Kollack-Walker et al., 1997), although this was not found in their other study (Kollack-Walker and Newman, 1995). Although lacking control procedures using other active social behaviors, Martinez et al (1998) also found increased c-Fos activity in the dorsal periaqueductal gray during defense in the rat. In response to a predator (a cat), increased c-Fos activation of midbrain central gray neurons has been found in the rat (Canteras et al., 1997, Comoli et al., 2003, Dielenberg et al., 2001).

(continued on next page)