Response rate and sensitivity to the molar feedback function relating response and reinforcement rate on VI+ schedules of reinforcement.

In 5 experiments, the author examined rats' sensitivity to the molar feedback function relating response rate to reinforcement rate on schedules of reinforcement. These studies demonstrated that, at lower rates of responding, rats' performance on variable ratio (VR), variable interval (VI), and variable interval with linear feedback loop (VI+) schedules was determined largely by reinforcement of interresponse times; response rates were faster on VR than on both VI and VI+ schedules. In contrast, when procedures were adopted to maintain high rates of response, rats showed sensitivity to the molar characteristics of the schedules; they responded as fast on a VI+ schedule as on a VR schedule and faster on both of these schedules than on a yoked VI schedule. When the variance of response rate was manipulated, this factor was noted as an important element in determining sensitivity to the molar characteristics of the schedule. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Primacy of interresponse-time reinforcement in accounting for rate differences under variable-ratio and variable-interval schedules.

Pigeons' keypeck rates under VR schedules are typically higher than under VI schedules when between-schedule reinforcement rates are equated. Exp I, with 4 male White Carneaux pigeons, reproduced this between-schedule rate difference in a multiple VR-VI schedule. However, when the short interresponse times (IRTs) typically reinforced under VRs were required for VI reinforcement, between-component rate differences diminished. Exp II (4 Ss) replicated Exp I except that long IRTs reinforced under VI schedules were required for VR reinforcement. This manipulation eliminated between-schedule rate differences. In Exp III (4 Ss), VR/VI and VI/VR tandems had the same correlation between response and reinforcement rates. In a simulation, C. P. Shimp's (1969) IRT response rule was used as an algorithm to mimic between-schedule rate difference. Findings show that IRT reinforcement (a molecular factor), rather than the feedback function between response rate and reinforcement rate (a molar factor), accounts for the higher response rates under VR schedules. (25 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Free-operant performance on variable interval schedules with a linear feedback loop: No evidence for molar sensitivities in rats.

Four experiments examined rats' sensitivity to molar and molecular factors on instrumental schedules of reinforcement. rats were exposed to a variable interval schedule with a positive feedback loop (VI+), such that faster responding led to a shorter interreinforcement interval. In Experiments 1 and 2, rats responded faster on a variable response (VR) schedule than on either a VI schedule matched for reinforcement rate or a VI + schedule matched for the feedback function. In Experiment 3, rats responded no differently on a VI schedule than they did on a VI+ schedule with equated rates of reinforcement. In Experiment 4, rats responded faster on a VI+ schedule with an interresponse time requirement yoked to that experienced on a VR schedule, than on a VI+ schedule with the same feedback function as the VR schedule. Taken together these results suggest that rats are more sensitive to the molecular than the molar properties of the schedules. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Interresponse-time reinforcement and behavior under aperiodic reinforcement schedules: A case study using computer modeling.

Computer modeling was used to investigate the extent to which response rates under variable-interval and variable-ratio schedules are compatible with a simple process of interresponse time (IRT) reinforcement, as argued by Peele, Casey, and Silberberg (1984). Their computer model was duplicated, as well as its principal result of a large response rate difference between the interval and ratio schedules. After their model was run under a variety of interval and ratio schedules, it was found that the response rates produced did not exhibit patterns of sensitivity to schedule parameter variation found experimentally. Furthermore, the model predicted a large response rate difference between a variable-ratio and a "linear feedback" variable-interval schedule, contrary to the results of McDowell and Wixted (1986). We concluded that simple IRT reinforcement was probably not adequate as an explanation of schedule effects under aperiodic interval and ratio schedules, although a modification of the Peele et al. model incorporating behaviors that were not measured operants could exhibit schedule sensitivity. This suggested that realistic molecular models of schedule phenomena must involve more than simple IRT reinforcement. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A test of the melioration theory of matching.

Melioration theory entails that matching in concurrent schedules occurs because the subjects equalize the local reinforcement rates (reinforcers received for each alternative divided by the time allocated to each alternative). The role of local reinforcement rates was tested by using multiple schedules in which one component involved an alternative with a high absolute rate of reinforcement and a low local reinforcement rate while the second component involved an alternative with a low absolute rate and a high local rate. These alternatives were then presented simultaneously in probe trials to determine preference between them. Contrary to melioration, the absolute rate of reinforcement, not the local rate, was the controlling variable. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Effort as a determinant of choice in rats.

Studied the effect of response effort on rate of responding and choice behavior in a 2-bar situation in 2 experiments with 17 male Sprague-Dawley rats. Exp I used a limited test session and tested at a single FR ratio. Exp II employed a free behavior situation, in which total daily water intake was obtained by barpressing. Ratio requirements beginning with continuous reinforcement and followed by sessions of FR 5, 10, 20, 40, 80, 120, and 160 were tested. Relations between response rate and preference behavior were schedule dependent. On low-ratio schedules the highest response rates and preference were for the light bar. As the ratios were increased, however, response rates became higher and preference was switched to the heavy bar. (19 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Human choice in concurrent ratio-interval schedules of reinforcement

Seven undergraduates participated in a concurrent-choice experiment with monetary reinforcers. Response-independent analogues of variable-interval and variable-ratio schedules were used to assess whether subjects would maximize reinforcement rate. The optimal pattern of behavior, in terms of maximizing reinforcement rate, involved a large bias toward the ratio alternative, with only occasional sampling of the interval schedule. Most experiments with pigeons, however, demonstrate matching of response rates to reinforcement rates, with only slight biases for the ratio schedule. Although subjects in the present experiment allocated more time to the ratio alternative than required by matching, the magnitude of the bias did not approximate that predicted by a maximizing account. After exposure to clock stimuli correlated with the operation of each schedule, 1 subject's behavior did show a substantial level of bias, increasing the total number of reinforcers obtained, and lay at a point between the predictions of matching and maximizing. The other subjects, however, continued to respond less optimally. The present results can be accounted for by a view of matching that incorporates the effects of delayed reinforcement.     

Acquisition of preference in concurrent chains: Comparing linear-operator and memory-representational models.

Two experiments were conducted to test predictions of 2 models for acquisition of preference in concurrent chains. Pigeons were trained with pairs of fixed-interval and variable-interval terminal-link schedules in a successive-reversal design. Results showed that acquisition rate was faster when fixed-interval schedules preceded and when variable-interval schedules followed the reversal and was approximately constant when the same pair of schedules was used. These results were predicted by both linear-operator and memory-representational models. However, only the linear-operator model predicted effects of schedule duration, Overall, these results demonstrated that a simple linear-operator model can account for the major features of preference acquisition in concurrent chains and suggest that choice in transition and steady state may provide convergent validation of a single delay-discounting function. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Molar function depends on molecular structure of behavior.

Pigeons chose between green side keys, then waited a shorter or longer time before pecking a center key, and finally chose between red side keys. Two successive left choices (to green and then red) with a shorter wait intervening, or 2 successive right choices with a longer wait intervening, were intermittently reinforced with food. The 2 required waiting times and the relative frequency of reinforcement for the shorter reinforced pattern were varied. Molar preference, defined in terms of coherent responses that satisfied the molecular reinforcement contingency, conformed to the highly adaptive matching level, but molar preference, defined in terms of incoherent responses, did not. The molar matching result therefore generalizes to responses with complex molecular structures provided that analyses distinguish between coherent and incoherent responses. The results are compatible with the idea that awareness can facilitate adaptation. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Exploring and developing a connectionist model of animal timing: Peak procedure and fixed-interval simulations.

The researchers examined the performance of a version of R. M. Church and H. Broadbent's (see PA, Vol 78:9087; see also 1992) oscillator-based connectionist model of animal timing. The model simulated a peak procedure task and showed appropriate within-trial behavior and superposition. Exploration of the model included (a) transitions from 1 reinforcement time to another, (b) changes in the threshold for response generation, and (c) changes in the oscillator set used to represent time. Performance under fixed-interval (FI) reinforcement schedules could be simulated by adding a variety of response-generating rules to the model. The connectionist model appropriately simulated averaged response rate vs. elapsed time functions on most FI schedules; it could also mimic cumulative records. The researchers concluded that the connectionist model formed a good basis for further development as a general model of temporal regulation of behavior in animals. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Regulation during challenge: A general model of learned performance under schedule constraint.

Develops a general behavior-regulation model of learned performance related to the equilibrium approach of W. Timberlake (1980) and W. Timberlake and J. Allison (1974). The model is based on 4 assumptions: (1) Both the instrumental and contingent responses are regulated with respect to their own set points; (2) these set points can be measured in a free baseline when both responses are relatively unconstrained and simultaneously available; (3) a reinforcement schedule can be seen as a constraint function that cross–couples the environmental effects of regulatory systems underlying the instrumental and contingent responses, thereby challenging their set points; and (4) molar behavior change under a schedule represents a compromise between the deviations from set points forced by the constraint function. These assumptions are translated into a set of coupled differential equations describing 2 regulatory systems related by a schedule. After providing an exact solution for this model, 2 special cases are derived as current alternative models of learned performance, those of J. Allison (see record 1976-23479-001) and J. R. Staddon (see record 1980-27259-001). The model is consistent in form with data from a variety of simple schedules. (53 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Two- versus three-alternative concurrent-chain schedules: A test of three models.

Two experiments with pigeons used concurrent-chain procedures with variable-interval schedules as initial links and different delays to food as terminal links. Two schedules were present in all sessions, but a 3rd schedule was alternately present and absent in successive sessions. When the 3rd schedule delivered food with no terminal-link delay, the presence of this schedule led to an increase in preference for the schedule with the shorter terminal link of the 2 unchanged schedules. When the terminal-link delay for the 3rd schedule was 30 s, the presence of this schedule led to a decrease in preference for the schedule with the shorter terminal link of the 2 unchanged schedules. These results are inconsistent with the predictions of R. Grace's (1994) contextual-choice model, but they are consistent with 2 other mathematical models- delay-reduction theory and the hyperbolic value-added model. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Operant regulation of feeding: A static analysis.

Conducted 3 experiments using cyclic-ratio schedules, a rapid method for studying the operant regulation of feeding rate. The cyclic method produces results comparable to traditional but time-consuming parametric methods. Ss in Exp I were 8 female albino rats, 4 of which were also used in Exps II and III. Performance on cyclic-ratio schedules is well described by a linear regulatory model that embodies 3 quantitative feedback assumptions: (a) that rate of feeding is regulated by the rate of operant behavior, (b) that taste factors have an additive effect on the rate of the operant response, and (c) that regulatory "gain" is inversely related to body weight. This model accurately describes poorer regulatory performance at high body weights and following amphetamine administration, and the effects of altered diet palatability on preferred feeding rates. (28 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Translating behavioral research into practical terms for alcoholism.

Practical aspects of behavioral research are considered at three levels. At the theoretical level, behavior analysis emphasizes transactional interpersonal determinants of behavior in contrast to traditional intrapsychic determinants. At the actual level, rules and procedures for behavior analysis and modification have been developed. At the potential level behavioral research on "experimental alcoholism," schedules of reinforcement and the discriminative properties of punishment suggest new ways of looking at the problem of alcoholism. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Parametric manipulation of interresponse-time contingency independent of reinforcement rate.

The pecking of 4 White Carneaux pigeons was reinforced under a modified interval-percentile procedure that allowed independent manipulation of overall reinforcement (RFM) rate and the degree to which RFM depended on interresponse-time (IRT) duration. Results show that increasing the contingency, as measured by the phi coefficient, between RFM and long IRTs while controlling the overall rate of RFM systematically increased the frequency of those IRTs and decreased response rate under both of the RFM rates studied. Increasing RFM rate also generally increased response rate, particularly under weaker IRT contingencies. Random-interval schedules with comparable RFM rates generated response rates and IRT distributions similar to those obtained with moderate-to-high IRT RFM contingencies. It is suggested that IRT RFM contingencies inherent in random-interval and constant-probability VI schedules exercise substantial control over responding independent of overall RFM rate effects. The IRT RFM contingencies inherent in these schedules may actually mask the effects of overall RFM rate; thus differences in response rate as a function of RFM rate when IRT RFM is eliminated may be underestimated. (15 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Within subjects partial reinforcement extinction effects for a bar pressing task.

Trained 24 male albino Sprague-Dawley rats to barpress for water during both continuous and partial reinforcement of their responses. In extinction, Ss trained to press 1 bar during continuous reinforcement and a 2nd bar during partial reinforcement made more correct responses in the partial than in the continuous reinforcement condition. No difference in response rates was observed for Ss who learned the same response during both reinforcement schedules. Pairing different lighting conditions with the different schedules of reinforcement had no effect on these results, but did affect other response measures. (French summary) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Effect on choice of absolute and relative values of reinforcer delay, amount, and frequency.

Three experiments used pigeons to examine the assumption of the matching law that relative and not absolute values of reinforcement control preference. Experiment 1 examined this assumption for reinforcer delay, Experiment 2 for reinforcer amount, and Experiment 3 for reinforcer frequency. In each experiment reinforcers were programmed according to nonindependent concurrent variable-interval variable-interval schedules. In all three experiments, there was an orderly change in preference as the absolute value of the aspect of reinforcement being investigated was varied, even though relative value was held constant. Preference for the shorter delay decreased as the absolute values of delay were decreased; preference for the larger amount decreased as the absolute values of amount were increased; and preference for the more frequent reinforcer decreased as the absolute values of reinforcer frequency were decreased. Several alternative models were considered, but none of these models could describe all of the data. The implications of these findings for modeling of choice behavior are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)