IDENTIFICATION AND ANALYSIS OF SINGLE UNIT ACTIVITY IN CENTRAL NERVOUS SYSTEM 



might Ije required for the inhibitory interneurons 

 proposed by Eccles et al. (20). c) The unit in question 

 should follow trains of stimuli up to a rate of at least 

 500 impulses per sec. This figure may eliminate some 

 small afferent fibers and may let in some synapses of 

 very high safety factor but probably separates the 

 vast majority of postsynaptic from presynaptic ele- 

 ments. (T) If the unit responds to stimulation of a 

 dorsal root or peripheral nerve which is cut distal to 

 the point of stimulation, then there should be only one 

 response for each stimulus. This last criterion elim- 

 inates afferent fibers carrying dorsal root reflex re- 

 spon.ses (to, 25, 56), but these have never been ob- 

 served at shorter latencies than 2.5 msec, and cannot 

 follow at high frequencies (26). i) Primary afferent 

 fibers not separated from their sensory receptor cells 

 can often be made to fire repetitively by natural 

 stimuli such as touch, pinch or stretch. These trains 

 of impulses are characterized by their extremely regu- 

 lar rhythms under sustained excitation and can 

 readilv be distinguished from most postsynaptic ele- 

 ments in this regarded. Again this type of criterion 

 tends to prevent the discovery of regularly firing post- 

 synaptic elements should these exist. While none of 

 the above criteria is definitive alone, together they 

 form a rather satisfactory method for identifying pri- 

 mary afferents penetrated by microelectrodes in the 

 central nervous system. Using these criteria, primary 

 afferents have been identified as deep as 4.5 mm below 

 the dorsal surface of the cat's spinal cord (26). 



Motoneurovs 



Responses from motoneurons may be identified by 

 their correlation with antidromic stimulation with 

 the adoption of only very reasonable assumptions. 

 Microelectrodes in ventral root fibers of the spinal 

 cord give responses to ventral root shocks which are 

 similar to those recorded from dorsal root fibers. 

 When the electrode is in the spinal cord, similar short- 

 latency spikes are recorded following ventral root 

 stimulation (fig. loi?). These are presumabK- the axons 

 of motoneurons. 



However, another type of short-latency response 

 may be recorded following ventral root stimuli as 

 shown in figure lo.-l. Brock et al. (ii), Frank & 

 Fuortes (26) and Woodbury & Patton (60) showed 

 that these responses are of longer duration than axon 

 spikes (figs. lo.-l and E), are followed by a long-lasting 

 hyperpolarization (fig. 15) and when elicited within 

 a critical interval following a previous spike, break at 

 the inflection point on the rising pha.se as shown in 



U 



N>m 



FIG. 10. Antidromic conduction block in cat's motoneurons. 

 Electrodes inserted in the cord may pick up two types of short 

 latency responses to pairs of ventral root shocks. In unit of 

 column A the response to the second of a pair of shocks suddenly 

 drops to 30 to 40 per cent when the shock interval is reduced 

 below a critical stimulus interval (about i o msec, here, but 

 often much shorter or longer). Conduction block must occur 

 near microelectrode since blocked impulse is visible there. Unit 

 of column B shows instead a smooth decrease in height of 

 second response as stimulus interval is decreased. Calibration: 

 50 mv. Time in both columns : i msec. (Note diflferent sweep 

 speeds in A and B.) Only units like that in B are found in ventral 

 roots. [From Frank & Fuortes (26).] 



figure lo.-l. The short latency of these responses identi- 

 fies them as from motoneurons, and the inflection in 

 the rising phase has been interpreted as due to a loss 

 in safety factor for conduction for axon hillock to 

 soma (i I, 26). If one accepts that the block in conduc- 

 tion at the critical stimulus interval occurs at the axon 

 hillock then, since only elements with long-duration 

 spikes show evidence of conduction block, it may be 

 concluded that long spikes originate upstreain froin 

 the axon hillock, i.e. in the cell bodies or dendrites, 

 and short spikes originate in axons. 



The role of motoneuron dendrites in the generation 

 of potentials following antidroinic stimulation has 

 not yet been settled. Fatt (22) has recorded poten- 



