Effects of varying stimulus disparity and the reinforcer ratio in concurrent-schedule and signal-detection procedures.

The present study measured the effects of stimulus and reinforcer variations on pigeons' behavior in two different choice procedures. Two intensities of white light were presented as the stimuli on the main key in a switching-key concurrent schedule and as the sample stimuli in a signal-detection procedure. Under both procedures, the scheduled rate of reinforcement was varied across conditions to produce various ratios of obtained reinforcement. These ratios were obtained for seven pairs of light intensities. In the concurrent schedules, the effects of reinforcer-ratio variations were positively correlated with the physical disparity between the two light intensities. In the signal-detection procedure, changes in the reinforcer ratio produced greater effects on performance when stimulus disparity was very low or very high compared to those found at intermediate levels of stimulus disparity. This discrepancy creates a dilemma for existing behavioral models of signal-detection performance.     

Differential reinforcement and signal detection.

Reinforcement was introduced for responses normally treated as errors in signal-detection procedures. The first experiment used a standard two-response discrete-trial procedure with no reinforcement for errors. Results showed that rats altered their response biases but maintained constant sensitivity to visual signals when reinforcement probabilities varied, and that their sensitivity depended on the physical difference between signals, in accordance with the predictions of signal-detection theory. Experiment II, with rats, and Experiment III, with pigeons, demonstrated that sensitivity decreased in this procedure when reinforcement was scheduled for errors with the signals held constant, despite independence of overall number of reinforcers and sensitivity. Experiment IV, with rats, replicated the decrease in sensitivity in a continuous procedure employing only one response. The decrements in sensitivity were similar across Experiments II, III, and IV, and accorded well with earlier research. Thus, contrary to a fundamental assumption of signal-detection theory, estimates of sensitivity are not always invariant with respect to the outcomes of responding, but depend on relative reinforcement of correct responses.     

Stimulus control and response bias in an analogue prey-detection procedure

The present study compared the performance of 6 pigeons trained to detect luminance differences in two different signal-detection procedures. Exposed to a three-key array, the pigeons were trained to peck the left key when the brighter of two light intensities had been presented on the center key and to peck the right key when the dimmer of two light intensities had been presented on the center key. Procedure A was a standard signal-detection procedure in which left/bright and right/dim responses produced food reinforcement and left/dim and right/bright responses produced periods of timeout. Procedure B was designed to simulate some of the contingencies operating in a prey-detection situation. Left-key responses produced reinforcement following the brighter center-key stimulus and a period of timeout following the dimmer center-key stimulus. Right-key responses always produced a short period of timeout irrespective of the stimulus. Within each procedure, the duration of timeout arranged for false alarms (left/dim responses) was varied between 3 s and 120 s. Measures of accuracy and response bias were compared between the two procedures. The timeout manipulation produced systematic, but relatively small, changes in these measures when right/dim responses (i.e., correct rejections) produced reinforcement (Procedure A). Arranging timeout for right/dim responses in Procedure B produced greater variability in accuracy and response bias than did arranging reinforcement, but this variability was not related to timeout duration. Overall, discrimination accuracy was considerably higher when right/dim responses produced timeout than when they resulted in reinforcement, and accuracy was accompanied by a large bias toward the response associated with reinforcement. These results are consistent with a recently proposed model of signal detection.     

Discriminating famous from fictional names based on lifetime experience: evidence in support of a signal-detection model based on finite mixture distributions

It is widely accepted that signal-detection mechanisms contribute to item-recognition memory decisions that involve discriminations between targets and lures based on a controlled laboratory study episode. Here, the authors employed mathematical modeling of receiver operating characteristics (ROC) to determine whether and how a signal-detection mechanism contributes to discriminations between moderately famous and fictional names based on lifetime experience. Unique to fame judgments is a lack of control over participants' previous exposure to the stimuli deemed "targets" by the experimenter; specifically, if they pertain to moderately famous individuals, participants may have had no prior exposure to a substantial proportion of the famous names presented. The authors adopted established models from the recognition-memory literature to examine the quantitative fit that could be obtained through the inclusion of signal-detection and threshold mechanisms for two data sets. They first established that a signal-detection process operating on graded evidence is critical to account for the fame judgment data they collected. They then determined whether the graded memory evidence for famous names would best be described with one distribution with greater variance than that for the fictional names, or with two finite mixture distributions for famous names that correspond to items with or without prior exposure, respectively. Analyses revealed that a model that included a d' parameter, as well as a mixture parameter, provided the best compromise between number of parameters and quantitative fit. Additional comparisons between this equal-variance signal-detection mixture model and a dual-process model, which included a high-threshold process in addition to a signal-detection process, also favored the former model. In support of the conjecture that the mixture parameter captures participants' prior experience, the authors found that it was increased when the analysis was restricted to names in occupational categories for which participants indicated high exposure.     

Moving beyond pure signal-detection models: comment on Wixted (2007)

The dual-process signal-detection (DPSD) model assumes that recognition memory is based on recollection of qualitative information or on a signal-detection-based familiarity process. The model has proven useful for understanding results from a wide range of memory research, including behavioral, neuropsychological, electrophysiological, and neuroimaging studies. However, a number of concerns have been raised about the model over the years, and it has been suggested that an unequal-variance signal-detection (UVSD) model that incorporates separate recollection and familiarity processes (J. T. Wixted, 2007) may provide an equally good, or even better, account of the data. In this article, the authors show that the results of studies that differentiate these models support the predictions of the DPSD model and indicate that recognition does not reflect the summing of 2 signal-detection processes, as the new UVSD model assumes. In addition, the assumptions of the DPSD model are clarified in order to address some of the common misconceptions about the model. Although important challenges remain, hybrid models such as this provide a more useful framework within which to understand human memory than do pure signal-detection models.     

Reinforcement contingencies and signal detection.

Pigeons were trained to discriminate temporal stimuli in a discrete-trial signal-detection procedure. Pecks to one side key were reinforced intermittently after exposure to one duration, and pecks to the other side key were reinforced intermittently after exposure to a different duration. In Experiment I, the allocation of reinforcers was varied systematically for correct responses and for errors, using a procedure that controlled the obtained numbers of reinforcers. When reinforcers were allocated symmetrically, the level of discrimination decreased as the proportion of reinforcers for errors increased. When reinforcers were allocated asymmetrically, the decrease in discrimination was less systematic. Bias toward one or the other side key roughly matched the ratio of reinforcers obtained by pecks at those keys, independent of the level of discrimination. In Experiment II, the overall rate of reinforcement for correct responses was varied both within and between experimental conditions. The level of discrimination was positively related to the overall rate of reinforcement. The discrimination data of both experiments were interpreted in relation to the contingencies of reinforcement and nonreinforcement, characterized by the average difference in reinforcement probability for correct responses and errors.     

Signal probability, reinforcement and signal detection.

Five pigeons were trained to detect differences in light intensity. Two stimuli, S1 and S2, differing in intensity, were arranged on the center key of a three-key chamber according to set probabilities. A peck on the center key turned on the two side keys. When S1 was presented on the center key, a peck on the left key was "correct" and when S2 was presented, a peck on the right key was "correct." Correct responses produced reinforcement and incorrect responses produced 3-second blackout. Detection performance was measured under three procedures. The first was a standard signal-detection design in which the probability of S1 was varied and the number of reinforcements obtained for correct responses to S1 was allowed to covary. In the second procedure, the probability of S1 was again varied but the distribution of reinforcements between the two choices was kept equal. In the third procedure, probability of S1 was held constant while the distribution of reinforcements was varied between the two choices. Changes in response bias were a function of variations in the relative reinforcement ratio for the choice responses and not a function of variations in the probability of stimulus presentation. Discriminability remained constant across the three procedures.     

Bias of phencyclidine discrimination by the schedule of reinforcement.

Pigeons, trained to discriminate phencyclidine from saline under a procedure requiring the bird to track the location of a color, received cumulative doses of phencyclidine, pentobarbital, or d-amphetamine with a variety of schedules of reinforcement in effect (across phases). When the same second-order schedules were used to reinforce responding after either saline or phencyclidine administration, stimulus control by phencyclidine did not depend on the schedule parameter. When different second-order schedules were used that biased responding toward the phencyclidine-correlated key color, pigeons responded on the phencyclidine-correlated key at lower doses of phencyclidine and pentobarbital than when the second-order schedule biased responding toward the saline key color. A similar but less marked effect was obtained with d-amphetamine. Attempts to produce bias by changing reinforcement magnitude (duration of food availability) were less successful. A signal-detection analysis of dose-effect curves for phencyclidine under two of the second-order schedules employed suggested that at low doses of phencyclidine, response bias is a major determinant of responding. As doses were increased, position preferences occurred and response bias decreased; at higher doses both response bias and position preference decreased and discriminability increased. With low doses of pentobarbital, responding again was biased but increasing doses produced position preference with only small increases in discriminability. At low doses of d-amphetamine responding also was biased, but bias did not decrease consistently with dose nor did discriminability increase. These experiments suggest that the schedule of reinforcement can be used to bias responding toward or away from making the drug-correlated response in drug discrimination experiments, and that signal-detection analysis and analysis of responding at a position can be used to separate the discriminability of the drug state from other effects of the drug on responding.     

Human Signal-detection Performance: Effects Of Signal Presentation Probabilities And Reinforcer Distributions

University students participated in one of four standard two-choice signal-detection experiments in which signal presentation probability was varied and the reinforcement distribution was held constant and equal. In Experiments 1, 3 and 4, subjects' performance showed a systematic response bias for reporting the stimulus presented least often. Experiments 1 and 4 showed that this effect was reliable with extended training and monetary, rather than point, reinforcement. In Experiment 2, all correct responses were signaled in some way, and this produced the opposite relationship between signal presentation probability and response bias. Experiments 1 and 3 found that explicitly deducting money (intended as punishment) for equal numbers of incorrect responses on each alternative, or varying the obtained overall rate of reinforcement, produced no clear change in response bias. The bias, shown by humans, for reporting the stimulus presented least often remains a challenge for theories of stimulus detection.     

A theory of attending and reinforcement in conditional discriminations

A model of conditional discrimination performance (Davison & Nevin, 1999) is combined with the notion that unmeasured attending to the sample and comparison stimuli, in the steady state and during disruption, depends on reinforcement in the same way as predicted for overt free-operant responding by behavioral momentum theory (Nevin & Grace, 2000). The rate of observing behavior, a measurable accompaniment of attending, is well described by an equation for steady-state responding derived from momentum theory, and the resistance to change of observing conforms to predictions of momentum theory, supporting a key assumption of the model. When probabilities of attending are less than 1.0, the model accounts for some aspects of conditional-discrimination performance that posed problems for the Davison-Nevin model: (a) the effects of differential reinforcement on the allocation of responses to the comparison stimuli and on accuracy in several matching-to-sample and signal-detection tasks where the differences between the stimuli or responses were varied across conditions, (b) the effects of overall reinforcer rate on the asymptotic level and resistance to change of both response rate and accuracy of matching to sample in multiple schedules, and (c) the effects of fixed-ratio reinforcement on accuracy. Some tests and extensions of the model are suggested, and the role of unmeasured events in behavior theory is considered.     

Stimulus discriminability in free-operant and discrete-trial detection procedures.

Six pigeons were trained to discriminate different light intensities in four experimental procedures. Experiment 1 compared stimulus discriminability in a yes-no signal-detection task with discriminability measures obtained from two free-operant procedures. Discriminability estimates were significantly lower in the detection procedure. Experiment 2 showed this lowered discriminability to be a function of the delay between stimulus presentation and the availability of the choice-response keys in the standard detection task. In addition, reinforcement sensitivity was lowest when correct choice responses were intermittently, rather than continuously, reinforced.     

Reinforcer control and human signal-detection performance

Eight humans participated in a two-choice signal-detection task in which stimulus disparity was varied over four levels. Two procedures arranged asymmetrical numbers of reinforcers received for correct left- and right-key responses (the reinforcer ratio). The controlled procedure ensured that the obtained reinforcer ratio remained constant over changes in stimulus disparity, irrespective of subjects' performances. In the uncontrolled procedure, the asymmetrical reinforcer ratio could covary with subjects' performances. The receiver operating characteristic (ROC) patterns obtained from the controlled procedure approximated isobias functions predicted by criterion location measures of bias. The uncontrolled procedure produced variable ROC patterns that were somewhat like the isobias predictions made by likelihood ratio measures of bias; however, the obtained reinforcer ratio became more extreme as discriminability decreased. The obtained pattern of bias was directly related to the obtained reinforcer ratio. This research indicates that criterion location measures seem to be preferable indices of response bias.     

Discriminability of fixed-ratio schedules for pigeons: effects of payoff values

Three pigeons, previously trained to discriminate different numbers of responses (fixed ratios), were tested under different reinforcement contingencies (payoff matrices) at two levels of sensitivity. For one subject, relative reinforcement magnitude was varied—at first, across sessions and then, at midsession by reversing values—without exteroceptive cues. For another, relative reinforcement magnitude and/or probability was varied every 50 trials with cues by correlating different payoff matrices with different key colors. For the third subject, relative reinforcement probability was varied more frequently with cues—in the limit, at random—to demonstrate stimulus control of response bias on a trial-by-trial basis. A signal-detection analysis showed that bias changed with payoffs, for as many as seven different matrices, while sensitivity remained unchanged. The obtained functions (receiver operating characteristics) were similar under different payoff conditions, which suggests that a single mechanism controls bias. However, they differed enough in slope to require a relatively complex account (e.g., the general Gaussian model of detection theory).     

Receiver operating characteristics from nonhuman animals: Some implications and directions for research with humans

Reviews of signal detection have largely overlooked the research involving nonhuman animal subjects. Some of this research is presented and reanalyzed here. Plots of receiver operating characteristics show that human and nonhuman signal-detection performance is very similar. The studies emphasize the series of discriminations that comprise signal-detection tasks and illustrate the systematic effects of different methods of arranging payoffs or feedback, of the consistency of that feedback, and of the physical disparity between response alternatives. The data provide some support for recent theoretical accounts that favor a criterion location measure of isobias over the likelihood ratio, but they also suggest that more systematic work is required in this area. Overall, this research supports many contemporary views concerning signal detection, and it provides an alternative way of looking at some recurrent issues. It also suggests that extensions of signal detection to analyze data from other research paradigms require some caution, and it offers directions for complementary research with human subjects.     

Response bias and the discrimination of stimulus duration

Pigeons discriminated stimulus duration in a psychophysical choice situation. Following presentation of any duration from a set of short duration (11 to 15 sec), responses on a red key were reinforced intermittently. Following presentation of any duration from a set of long durations (16 to 22 sec), responses on a green key were reinforced intermittently. Relative reinforcement rates were manipulated for choice responses across conditions. As relative reinforcement rates were varied, psychometric functions showed shifts in green-key responses at all durations. A signal-detection analysis showed that sensitivity remained roughly constant across conditions while response bias changed as a function of changes in relative reinforcement rate. Relative error rates tended to match relative reinforcement rates.     

Analyzing behavior implied by EWA learning: An emphasis on distinguishing reinforcement from belief learning

An important issue in the field of learning is to what extent one can distinguish between behavior resulting from either belief or reinforcement learning. Previous research suggests that it is difficult or even impossible to distinguish belief from reinforcement learning: belief and reinforcement models often fit the empirical data equally well. However, previous research has been confined to specific games in specific settings. In the present study we derive predictions for behavior in games using the EWA learning model (e.g., Camerer & Ho, 1999), a model that includes belief learning and a specific type of reinforcement learning as special cases. We conclude that belief and reinforcement learning can be distinguished, even in 2×2 games. Maximum differentiation in behavior resulting from either belief or reinforcement learning is obtained in games with pure Nash equilibria with negative payoffs and at least one other strategy combination with only positive payoffs. Our results help researchers to identify games in which belief and reinforcement learning can be discerned easily.     

Negative reinforcement and choice in humans

Participants chose between reinforcement schedules differing in delay and/or duration of noise offset. In Experiment 1 it was found that (1) immediate reinforcement was preferred to delayed reinforcement when amounts (durations) of reinforcement were equal; (2) a relatively large reinforcer was preferred to a smaller one when both reinforcers were obtained immediately; and (3) preference for an immediate, small reinforcer versus a delayed, large reinforcer increased as the delay preceding the large reinforcer increased, a sign of “impulsivity”. In Experiment 2, the schedules differed in amount or delay and equal intervals were added either to the constant parameter or the varied parameter. A shift from virtually exclusive preference to indifference occurred in the latter case but not the former, a result supporting a model of self-control that assumes that the value of a schedule depends on the ratio of amount and delay, and that choice between schedules depends on the ratio of these values.     

Dual-process theory and signal-detection theory of recognition memory

Two influential models of recognition memory, the unequal-variance signal-detection model and a dual-process threshold/detection model, accurately describe the receiver operating characteristic, but only the latter model can provide estimates of recollection and familiarity. Such estimates often accord with those provided by the remember-know procedure, and both methods are now widely used in the neuroscience literature to identify the brain correlates of recollection and familiarity. However, in recent years, a substantial literature has accumulated directly contrasting the signal-detection model against the threshold/detection model, and that literature is almost unanimous in its endorsement of signal-detection theory. A dual-process version of signal-detection theory implies that individual recognition decisions are not process pure, and it suggests new ways to investigate the brain correlates of recognition memory.     

Tolerance to effects of cocaine on behavior under a response-initiated fixed-interval schedule

Tolerance to effects of cocaine can be modulated by schedules of reinforcement. With multiple ratio schedules, research has shown an inverse relationship between ratio requirement and amount of tolerance that resulted from daily administration of the drug. In contrast, tolerance to the effects of cocaine on behavior under multiple interval schedules generally has developed regardless of interval value. Under interval schedules reinforcement depends on the animal making one response following a time interval. Thus, as time to respond increases, the time to reinforcement decreases. On the other hand, fixed ratio schedules require a specified number of responses to be made prior to reinforcement. Therefore, delaying the initiation of responding does not coincide with a significant decrease in the time to reinforcement. In the current experiment, 6 pigeons were trained to respond under a three-component multiple schedule, with a different tandem fixed-ratio 1 fixed-interval schedule in each component. The multiple schedule required one response, which was followed by one of three fixed-interval values (5, 15, or 60 s). Thus, the multiple schedule was interval-like because after the fixed-ratio 1, only one more response was required for reinforcement, but it was also ratio-like because the length of the pause at the beginning of each interreinforcer interval affected the time until the next reinforcer. Acute administration of cocaine generally resulted in dose-dependent decreases in responding. Chronic (i.e., daily) administration of a rate-decreasing dose resulted in tolerance patterns similar to those usually obtained with multiple ratio schedules. That is, the magnitude of tolerance was related inversely to schedule size. These results suggest that delay to reinforcement from the initial response may play a role in the development of schedule-parameter-related tolerance.