GENERAL DISCUSSION 649 



that self-Stimulation, at least in learning, must be a jamming of the normal activi- 

 ties. Now I would think that at the present time we are not certain what action on 

 nerve units our svstem of self-stimulation has. One thing is to define the stimulation 

 from the point of view of behaviour because it results in a behaviour, but I think it 

 is important for us to realize that this behaviour might be the result of jamming 

 one mechanism and releasing another. And it is very clear to us, and we can give 

 several examples in which stimulation sometimes elicits that kind of behaviour, 

 sometimes it prevents this kind of behaviour. And I think we can frequently get oft 

 the track bv assuming immediately that selt-stimulation, or any other sort of 

 stimulation, is necessarily excitation. It might be excitation in behavioural terms 

 but in neural terms it might be inhibition or jamming. 



Olds. We are surprised how often cortical units that we try to reinforce appear 

 to be inhibited by stimulation of the very hypothalamic areas which cause self- 

 stimulation. We are sure from proximal recording m the hypothalamus that 

 hypothalamic units, have, of course, their tiring rate augmented. That is, we can 

 elicit proximal hvpothalamic units, which is not surprising, nevertheless we are 

 probablv disrupting patterns. My own conception oi the disruption experiments is 

 precisely that while the stimulation of these hypothalamic mechanisms to excessive 

 discharge causes repetition of the preceding behaviour and I think leads to organiza- 

 tion of positive reinforcement mechanisms, at the same time it causes disruption 

 of the spontaneous patterns which are there, and I think thereby confusion ensues 

 whenever acuity in a learning situation is required. This is of course theoretical. 



Myers: I should like to respond to Dr Morrell's earlier comments regarding 

 oscillation in learned behaviour. Dr Morrell described a kind of instability or 

 variation in the level of sustained performance in some ot his animals. We have 

 found both stability and lack of stability in our own learning experiments with cats 

 and I think we can specify to a certain extent the situations in which each may occur. 

 When a learned response seems 'easy' for an animal — and 'easy' may be defined 

 either in terms of total number of trials required to learn or in terms ot the objective 

 degree of difference between two patterns to be discriminated — then remarkable 

 stability of performance usually obtains. This stability manitests itself not only in 

 constant high level performance from day to day but also, as Dr Konorski has 

 described, in high level performance on resetting alter long, practice-free intervals 

 of up to a year or more. Instability of performance occurs, on the other hand, when 

 the dirtkulty of the response approaches the limit of the animal's capacity to 

 attend to the differences in the stimuli. Brain-operated animals may exhibit insta- 

 bility in performance on 'easv' discriminations implying that an 'easy' discrimination 

 may become 'difficult' tor them. 



Chow. The most fundamental problem is the first point that Dr Thorpe made. 

 This problem of self-stimulation, learning and behavioural oscillation is to me an 

 irrelevancy. The most fundamental point is how a single nerve cell in its environ- 

 ment which is probably relatively stable gets this oscillation. At the cellular or 

 molecular level vou must think ot a mechanism to account tor it. A build-up and 



