Brain Mechanisms

The advantages of simple systems in neuroethology. The central issue in the Adams's paper is best described as the localization of neural structures associated with defensive end offensive fixed-action patterns. In general, lesion and stimulation experiments can only suggest that neural activity at certain loci may be necessary or sufficient to elicit behavior. The validity of these interpretations depends upon the additional constraints that (a) the neural deficit due to lesions be confined to the target structure, and (b) that selective electrical stimulation ot a neural locus reasonably mimics physiological activity at the same site when the behavioral activity is elicited naturally. Were these tests performed, the results would indicate the localization of the above mentioned mechanisms. To extend our understanding beyond a gross structural association, it is necessary to establish the degree of functional heterogeneity of the neurons at a stimulated locus, the character of physiological activity in the relevant neurons, the patterns of functional connectivity within and between loci, and the dynamics of synaptic and ensemble interactions that release and control the behavior These are the issues that address the neural "mechanisms."

The circuitry that Adams postulates for the release of defense behavior could be constructed with about twenty neurons This is several orders of magnitude fewer than the number of participating nerve cells indicated by lesion studies. These considerations imply that the methodology or the conceptual approach may be seriously wanting.

I believe that. in our present state of ignorance, the invertebrates and lower vertebrates provide more appropriate subjects for the study of the neural basis of fixed-action patterns. Behavioral and neurophysiological analyses indicate that these systems exhibit a significant economy in the number of neurons controlling a behavior and substantially less variability in the behavioral output. Individual neurons at all levels of escape and defense pathways can be uniquely identified (Zucker 1972; Auerbach and Bennett 1969; Getting 1977) and connectivity patterns and synaptic actions are amenable to direct cellular analysis. Although there are still some obstacles to a completely safisfactory description of these systems (Kupferman and Weiss 1978; Davis 1977; Glantz 1977), the limitations do not appear to be essential aspects of either the methodology or the conceptual approach. Experiments can be performed to determine whether the physiological activity of a single neuron is necessary or sufficient to release a given behavior (Glantz 1977; Taghert and Willows 1978). Furthermore, the interpretation of these experiments does not require questionable assumptions, such as the homogeneity of a large population of synchronously activated cells, nor the ambiguity inherent in the unphysiological nature of such a manipulation .

A few of the more salient, emerging generalizations are discussed in several recent symposia (Stein et al 1973; Fentress 1966b; Galun et al 1976, Hoyle 1977). Briefly, they are as follows (i) The decision process for defense or escape is invested in a small number of high-threshold, rapidly conducting, multisegmental interneurons. (ii) The interneurons project to all of the segments participating in the action pattern. (iii) The intersegmental interneurons either synapse directly with the relevant motoneurons or with local premotor interneurons (iv) Speed and synchronization are optimized by the presence of electrotonic connections between interneurons and motoneurons, between synergist interneurons, and between synergist motoneurons (v) The sensory filters are composed of a small number of primary afferents or sensory interneurons, whose input requirements are precisely tuned to the requirements of the behavior (vi) Adaptive variations of the behavior pattern arise from the activation of different subsets of sensory neurons (vii) Lability (e.g. habituation) arises in the afferent limb of the neural pathway. (viii) Concurrent with the onset of behavior, the afferent input may be inhibited by the decision-making intersegmental interneuron.

All of these generalizations apply to the escape systems of the crayfish and hatchet fish, and instances of all of these phenomena have been observed in escape and defense systems of other arthropods, molluscs, annelids, and fish These generalizations provide an important foundation for further research into the neuronal mechanisms of defensive fixed-action patterns. At present, however, we lack an appropriate methodology to address these issues in mammalian systems