J. C. ECCLES 337 



because ot the increased efficacy of the synaptic actions exerted by all those 

 neurones activated initially. There would thus be a further reinforcement 

 of the synapses responsible for the unique pattern of activation and 

 response, with consequently a more effective channelling; and so on, 

 cumulatively, tor each successive application of that sensory input. 

 Necessarily, the postulated changes in synaptic efficacy must be of very 

 long duration — days or weeks. There is no way in which relatively brief 

 durations of synaptic change for each synapse of a serial arrangement can 

 sum to give a more prolonged change. 



In designing experiments to test for this postulated effect of use in 

 causing a prolonged increase in synaptic efficacy, it was initially much 

 simpler to test for the opposite effect — namely regression of synaptic 

 function with disuse. The very recent investigations on excess use will be 

 considered subsequently. Furthermore, the postulated effects of use in 

 increasing synaptic efficacy and of disuse in depressing it can be investi- 

 gated most rigorously with monosynaptic pathways, i.e. where there is 

 only one synapse interpolated between the known input and the observed 

 output. 



The extreme complexity of neuronal connections in the higher levels of 

 the central nervous system has, hitherto, prevented critical testing of the 

 various hypotheses proposed for the neural mechanism of learning. As a 

 consequence there has been an unrestricted growth of speculation, as may 

 be seen by reference to Gomulicki's comprehensive review (1953). It is an 

 initial postulate of the present investigation that synpases at spinal level 

 share in some measure the synaptic property that is responsible for the 

 phenomena of learning at higher levels of the nervous system. 



The experiments have been restricted to the monosynaptic activation of 

 motoneuroncs by impulses in the large afferent fibres from the annulo- 

 spiral endings of muscle spindles. The advantage of this system is that the 

 input is under precise control and the output can be measured as the size of 

 the reflex spike response. This size is directly related to the number of 

 motoneurones discharging, and hence to the efficacy of their mono- 

 synaptic activation. As a routine procedure before all the monosynaptic 

 testing described below, the spinal cord was transected in the upper 

 lumbar or lower thoracic region so as to eliminate the possibility of reflex 

 inequality caused by a tonic asymmetric discharge of suprasegmental 

 origin. In addition to the investigation of the responses evoked by single 

 afferent volleys, there has regularly been an investigation into the potentia- 

 tion of these responses that follows a high frequency tetanus of the afferent 

 pathway usually at about 400 sec, the so-called post-tetanic potentiation. 



