THE ROLE OF SUBCORTICAL STRUCTURES IN CONDITIONED REFLEXES

Anokhin has made two major points of great importance for theories of higher nervous activity. First he has emphasized that in complex, learned behaviors the cerebral cortex is not acting alone, but is in close association with subcortical systems. With the rich interchange known to exist between these systems it is appropriate to seek the mechanism of conditioned reflex formation and performance in the cortical-subcortical relationships rather than predominantly in the cortex itself. While pointing out this importance of subcortical structures, Anokhin has also clearly demonstrated that they are intricately organized and cannot be regarded simply as an indiscriminate “energizing influence” on the cortex. His experiments repeatedly and convincingly documented this second point. Together with many other facts they should convince us that the medial brain stem systems are capable of more precise action than is indicated in their diffuse or “nonspecific” effect upon the EEG. I am certainly in agreement with Anokhin on both of these points. Thus, rather than discuss details of his paper or simply restate his case, I hope it will be more valuable to offer a summary of data, for brevity and convenience taken chiefly from our laboratory, which give general support to his view and are relevant to his consideration of the role of subcortical structures in conditioned reflexes.* Using a defensive conditioned reflex (CR) in cats, we found during the course of training that EEG arousal reactions to an auditory conditional stimulus (CS) always attained a high level of consistency before the first respiratory or flexion CR appeared.’ Prima facie this indicates participation of subcortical mechanisms, although obviously the effect might be initiated from the cortex. Graphs of the per cent occurrence of these EEG arousal reactions had the form of typical “learning curves” for habituation, conditioning, and extinction. Surprisingly, however, under the conditions of these experiments “habituation” of the arousal reaction to the CS seemed to take place during the early training, even though a shock to the leg was being paired with the CS. Another unexpected finding was that the intensity of the arousal reaction diminished after the CRs became consistent, and with continuation of training a state was reached at which leg flexion CRs could be made with no apparent alteration in the electrical activity of postcruciate, marginal, and middle ectosylvian gyri. Pursuing this matter independently, Edward Beck and his colleagues a t the University of Utah2 have found with sleep-deprived cats that when such flexion CKs are made in the absence of low-voltage fast activity in the neocortex, a change in activity is still always found in certain subcortical areas. The * The original work described in this discussion was performed in the laboratories of the University of Michigan under Research Grant B-1068 from the National Institute of Neurological Diseases and Blindness, Public Health Service, Bethesda, Md., and by a research grant from Foundations’ Fund for Research in Psychiatry, New Haven, Corn.