Unblocking in a runway discrimination problem produced by a surprising reduction in S− reward magnitude at the beginning of compound stimulus training

Bulletin of the Psychonomic Society 1984 ,22 (1),63-66 Unblocking in a runway discrimination problem produced by a surprising reduction in S- reward magnitude at the beginning of compound stimulus training STEVEN J. HAGGBLOOM ArkansasState University, State University, Arkansas The effects of a surprising reduction in S- reward magnitude on the blocking of the acquisition of stimulus control by brightness cues following prior acquisition of discriminative control by internal cues were investigated here in an instrumental discrimination problem in the runway. Blocking groups received large reward (L) in S+ and nonreward (N) in S- during Phase 1 and failed to learn about brightness when brightness cues were added as redundant, relevant discriminanda in Phase 2 and the reinforcement conditions remained the sarne as in Phase 1. Control groups received training with internal cues and brightness cues simultaneously and equally relevant either in Phases 1 and 2 or only in Phase 2. Brightness cues acquired substantial discriminative control in both control groups. Two unblocking groups received small reward (S) in S- in Phase 1, but otherwise were trained as the blocking groups were. The shift from S in S- to N in S- at the beginning of compound stimulus training attenuated the blocking effect in both unblocking groups. Blocking refers to the observation that a group that first learns to respond to some stimulus, A, and then experiences another stimulus, X, only as a redundant source of information, that is, as part of an AX compound, learns less about X than does a control group given training only with the AX compound. In Pavlovian conditioning experiments on blocking, a surprising increase in unconditioned stimulus (US) intensity coincident with the beginning of AX training results in learning about X that otherwise would have been blocked (e.g., Kamin, 1969 ; Mackintosh & Turner, 1971), a phenomenon referred to as unblocking. Unblocking in Pavlovian conditioning also has been produced by the surprising addition of a second US or the omission of an expected, second US (Dickinson, Hall, & Mackintosh , 1976), by a change in US locus (Stickney & Donahoe, 1983) , and by a memory reactivation treatment (Balaz, Gutsin , Cacheiro, & Miller, 1982 ; Schactman, Gee, Kasprow, & Miller, 1983) . The very considerable interest in treatments that produce unblocking is largely due to the fact that theoretical explanations of the blocking effect differ most with regard to predictions about conditions under which blocking will fail to occur. Thus, Wagner, Mazur, Donnegan , and Pfautz (1980) reported that a surprising decrease in US intensity at the beginning of AX training failed to produce unblocking , a result consistent with This experiment was supported by an Arkansas State University faculty research grant to the author. The author's mailing address is: Department of Counselor Education and Psychology, Arkansas State University, P.O. Box 2127 , State University, Arkansas 72467. 63 the Rescorla-Wagner model (e.g., Rescorla & Wagner, 1972), but at variance with theoretical formulations proposed by Logan (1977), Mackintosh (e.g., 1975), and Pearce and Hall (1980). The blocking effect has been demonstrated in instrumental conditioning in both the operant chamber (e.g., Miles, 1970 ; Neely & Wagner, 1974; Seraganian & vom Saal, 1969) and the runway (Capaldi, Verry, & Nawrocki, 1982 ; Haddad , Walkenbach, Preston, & Strong , 1981; Haggbloom, 1981), but there is no unequivocal evidence for unblocking in instrumental conditioning. Feldman (1971) and Neely and Wagner (1974) reported the only previous investigations of unblocking in an instrumental task. Neely and Wagner provided evidence that the unblocking effects obtained by Feldman and in two of their own experiments were due to generalization decrement brought about by a change in internal , reward-produced stimuli (Capaldi, 1967) accompanying the reward shifts. Haggbloom (1981) reported that rats trained on a runway discrimination problem in which internal , reward-produced, and intertrial-interval(ITI)-related cues were relevant discriminanda and brightness cues were irrelevant discriminanda in Phase 1 failed to learn about brightness when brightness cues were added as redundant, relevant discriminanda in Phase 2. The experiment reported here investigated whether a surprising decrease in S- reward magnitude at the beginning of Phase 2 would produce unblocking in the task employed by Haggbloom (1981). The operation of reducing S- reward in a discrimination problem is viewed here as analogous to increasing US intensity in Pavlovian conCopyright 1984 Psychonomic Society, Inc. 64 HAGGBLOOM ditioning. The present study does not distinguish among the several alternative theoretical explanations of blocking and unblocking effects, but it does bear on the generality of the unblocking effect and on the applicability of those theories , developed largely in the context of Pavlovian conditioning, to instrumental conditioning situations . There were six groups in the present experiment. Two blocking groups and a control group were trained as comparable groups employed by Haggbloom (1981) had been. A second control group received training in both Phases 1 and 2, with internal cues and brightness cues being relevant in both phases. Finally , two unblocking groups experienced a surprising reduction in S- reward magnitude at the beginning of compound stimulus training. METHOD Subjects The 60 male rats used as subjects were bred in the laboratory from Holtzman stock . Apparatus The apparatus consisted of two black and white runways fully described elsewhere (Haggbloom, 1981). The alleys were divided into three sections over which running time s were record ed (respectively, star t, run , and goal time). Start times were recorded from the opening of the startbox door , which triggered a .Ot-sec clock, to a point 30 cm into the alley. Run and goal times were recorded over the next 40 and 30 em, respectively. The offset of the first clock and the operation of the remaining clocks were controlled by photoelectric circuitry . Procedure The rats were housed in individual cages and had free access to water throughout the experiment. A food-deprivation schedule consisting of 12 g of Wayne Lab Blox per day was begun 14 days prior to the beginning of training. The deprivation schedule continued throughout the experiment ; the amount of food consumed in the runway was subtracted from the daily ration. On each of Days 12-14 of deprivation, the rats were handled in squads of three for 3 min per squad . After being handled on each day, the rats received 10 45-mg Noyes food pellets in a small dish placed in their home cages. Phase 1 discrimination training began on Day 15 of deprivation . A trial was initiated when the startbox door was opened approximately 3 sec after the rat had been placed in the startbox and regardless of the rat's orientation. A maximum ti (...truncated)