284 BRAIN MECHANISMS AND LEARNING 



Stimuli could signal cessation of unconditioned excitation (Barlow, 1956); 

 in the present experiments, a cat came to react to a formerly ineffectual 

 tone after the latter was followed, in a regular fashion by substraction of 

 prolonged painful stimuli. Accordingly, 'signal reactions' may indeed be 

 created as suggested by Konorski when absolute 'stimuli' are provided by 

 interruption of biologically significant excitations — feeding, pain (Konor- 

 ski, 1948);^ this is not altogether surprising since suppression of pain 

 (feeding) constitutes a biologically important reward (punishment), 

 probably associated with central nervous readjustments as great as those 

 accompanying their application. 



Behavioural effects of SS cessation (when not preceded by T) and of T 

 initiation (during application of SS) were variable. Nevertheless, affniity 

 between consequences of both phenomena could be established, both as a 

 group and individually. In the tirst place, variability though wide was not 

 unlimited and lists of single movements and of combined responses 

 provoked by SS substraction and by T presentation were coincident. In 

 the second place, observation of cats answering to SS cessation with one 

 predominant behaviour pattern indicated that this same performance 

 constituted their most frequent response to T. Therefore, when T was 

 apphed cats behaved in an anticipatory fashion as if SS had been inter- 

 rupted (even though excitation were sustained). 



Tones used in this training routine also acquired the ability to modify 

 potentials evoked in sensory cortex by SS excitation; effects consisting 

 essentially in masking or cancelling of peaks N3, P3, N2 and, commonly, 

 P2 and Ni; analysis of electrographic changes may serve to clarify 

 problems posed by this type of learning. Similar blocking influences 

 act upon ccutical sensory evoked potentials as a result of: (1) physio- 

 logical stimuli (e.g. blowing air into nostrils); (ii) central nervous stimuli 

 (e.g. of MRF); (iii) anti-cholmesterase administration (Bremer and 

 Stoupel, 1959; Desmedt and La Grutta, 1957; Gauthier, Parma and 

 Zanchetti, 1958; etc.). Agents (1), (ii) and (iii) exhibit the second common 

 feature, of provoking EEG 'dcsynchronization'. It is possible that T 

 applied during SS also exerted an EEG 'desynchronizing' influence. This 

 contingency was supported by the following observations: {a) such an 

 effect was observed frequently as a generalized or localized (to somatic 

 sensory cortices) phenomenon when T was presented alone;- (/') on T 



' One cannot discard the possibility that new hnks could be established if indifferent tones 

 were associated with cessation of relatively 'non-significant' stimuli (e.g. repetitive flash). 



- When T was applied during SS, 'dcsynchronization' on sensory cortex was, strictly 

 speaking, difficult to appreciate. Subsequent tracings added little because cessation of SS 

 itself (preceded or not by T) was followed always by low-voltage, fast activity. 



