Forward and backward classical conditioning of the flexion reflex in the spinal cat

Abstract

Effects of forward and backward conditioned-unconditioned stimulus (CS-US) intervals on classical conditioning of the flexion reflex were examined in a cat spinal preparation. Interstimulus intervals (ISIs) ranging from +3.0 to -3.0 sec were employed in 9 experimental groups and the results compared with those of an explicitly unpaired control group. Forward conditioning produced an asymmetrical, inverted U-shaped gradient relating magnitude of conditioning to ISI for both acquisition and extinction. The optimum ISI was 1.0 sec. Backward ISIs also produced excitatory conditioning, with optimal conditioning at -0.25 sec. Unlike forward conditioning, backward conditioning produced little sign of retention during extinction trials. The results, which parallel in several ways those of ISI effects in studies of intact animals, support the hypothesis that backward and forward conditioning may be fundamentally different phenomena, under the control of different neural processes.

Evidence that the neural pathways involved in backward conditioning are different from those involved in forward conditioning

“…BC was demonstrated here to be elicited through spinal reflex pathways activated only by the large myelinated afferent fibers of the CS, while the same CS was unsuccessful in producing significant FC. In addition, the results are similar to previous observations that suggested that BC appears to be uncommonly influenced by extinction procedures compared to FC groups (Durkovic and Damianopoulos, 1986). This discussion centers around these findings and their implications for current theories of associative learning and for spinal conditioning as a simplified neural model of mammalian learning.…”

“…7). The similarity of these IS1 function curves suggests that most, or possibly even all, of the BC effects observed in the Durkovic and Damianopoulos (1986) study were brought about by spinal reflex pathways activated by A-alpha cutaneous fibers of the saphenous nerve. Any differences between the BC results of the 2 studies (e.g., BC acquisition effects were statistically significant in the -0.5 and -1 .O set IS1 groups in the earlier study) might be explained by (1) the fact that not all A-alpha cutaneous fibers of the saphenous nerve were activated by the CS in the present study (see Materials and Methods), in contrast to the earlier work, and/or (2) a contribution from A-delta fiber inputs to the BC results of the earlier work.…”

“…To make observations parallel to those conducted in the previous study (Durkovic and Damianopoulos, 1986), an analysis of variance was carried out on changes over trials, using the results of the -0.25 set backward, the + 1.0 set forward, and the unpaired control treatment groups. Analysis ofthe data from the 6 test trials during acquisition indicated significant treatment (F(2,21) = 3.47, p < 0.025) and trial (F(5,105) = 4.20, p < O.OOS), but no significant treatment x trial interaction (F(10,105) = 0.320, p > 0.05) effects.…”

“…The results are best approached in the context of a comparison with a parallel study on this preparation in which a higherintensity CS (activation of large and small myelinated cutaneous afferent fibers) was employed (Durkovic and Damianopoulos, 1986). To help with this comparison, we constructed "net conditioning" (Fig.…”

“…In a recent study it was observed that the interval between the onset of trains of electrical stimuli delivered to 2 different cutaneous nerves was a determining factor in the development of classically conditioned flexion reflex facilitation in the spinal cat (Durkovic and Damianopoulos, 1986). Thus, excitatory conditioning was observed for certain forward interstimulus intervals (ISIS) [conditioned stimulus (CS) onset preceded unconditioned stimulus (US) onset] and certain backward ISIS (US onset preceded CS onset), but not for others.…”