Effects of schedules of reinforcement on the variability of location
A handful of studies have examined the relationship between schedules of reinforcement and behavioural variability, with mixed results. The present study compares a greater range of schedules, examining the effects on location variability. Hens worked in an operant chamber containing five, horizontally arranged and active response keys. Experiment 1 compared eight schedules: FR 40, FR 10, FI (yoked FR 40), FI (yoked FR 10), VR 40, VR 10, VI (yoked VR 40) and VI (yoked VI 10). Experiment 2 compared a series of DRL schedules: 0.5 s to DRL 19.2 s, incrementing in step sizes of 20%. Location variability was measured as the percentage of switching across keys from within trials, between trials (switching from the reinforced response location to the first response location of the following trial), and the number of keys used. Results from Experiment 1 suggested no relation between schedules and location variability. However a relation between response rate and location variability was found; faster response rates resulted in less variability. Experiment 2, in attempt to control for response rate, found no relation between response rate and location variability. Switches within trials occurred more frequently in Experiment 2 than in Experiment 1, an aspect that needs further examination.
Behavioral Variability in Humans: Do Instructions Make a difference?
Previous research shows that reinforcement of variable responding will facilitate sequence learning in rats but may interfere with sequence learning in humans. The present study aimed to explore the notion of sequence difficulty in humans by manipulating both sequence length (6-12 items) and task instruction. Participants were randomly allocated to one of two groups: Control and Variable. In the control group sequences were only reinforced if they were the target sequence, in the variability groups sequences were concurrently reinforced on a Variable Interval 60-s schedule if the just entered sequence met a variability criterion and for every production of the target sequence. With the instructions - to earn as many points as possible by figuring out the correct sequence - the six-item sequence (122121) Control group were most likely to produce the target sequence, while for the twelve digit sequence (221112211121) there was no difference between the two groups. The Variable group were most likely to produce the target sequence for the intermediate nine-digit sequence (112212121). Preliminary findings on the same task without instructions suggest comparable findings. The use of sequence length as a definition of sequence difficulty in both the current and previous studies and the potential role of instructions are discussed.
Brushtail possums ('Trichosurus vulpecula') counting using response sequences under ratio reinforcement
The aim of this study was to replicate the 'Mechner Counting Procedure' and conduct a variety of conditions to determine whether possums have the cognitive ability to count. Previous research has shown that animals are able to count the amount of responses they make and the number of light flashes that occur. Six brushtail possums participated in three conditions where FR responses were required to gain access to food reinforcement. Reinforcement was delivered either upon the completion of an FR response requirement on lever A, or on the completion of the FR on lever A, followed by an additional response on lever B. The results have shown that the possums do have the cognitive ability to count the number of responses they make. The mean response on lever B typically occurred slightly above the FR response requirement across conditions.
Behavioural variability
The added reinforcement of variable responding will facilitate the learning of a difficult target sequence in rats, however, the added requirement of variability has been shown to impede difficult sequence learning in humans. The present study aimed to explore the notion of sequence difficulty in humans by manipulating sequence length (6-12 items). Eighty participants were randomly allocated to one of two groups: Control and Variable. In the control conditions sequences were only reinforced if they were the target sequence, in the variability conditions sequences were concurrently reinforced on a Variable Interval 60-s schedule if the just entered sequence met a variability criterion. For the six-item sequence (122121) the Control group were most likely to produce the target sequence, while for the twelve digit sequence (221112211121) there was no difference between the two groups. The Variable group were most likely to produce the target sequence for the intermediate nine-digit sequence (112212121).The use of sequence length as a definition of sequence difficulty in both the current and previous studies are discussed.
Weber's Law and the Scalar Property of Timing: A Test of Canine Timing
Domestic dogs completed a temporal bisection procedure that required a response to one lever following a light stimulus of short duration and to another lever following a light stimulus of a longer duration. The short and long durations across the four conditions were (0.5-2.0 s, 1.0-4.0 s, 2.0-8.0 s, and 4.0-16.0 s). Durations that were intermediate, the training durations, and the training durations, were presented during generalization tests. The dogs bisected the intervals near the geometric mean of the short and long-stimulus pair. Weber fractions were not constant when plotted as a function of time: A U-shaped function described them. These results replicate the findings of previous research reporting points of subjective equality falling close to the geometric mean and also confirm recent reports of systematic departures from Weber's law.
Demonstration of the Scalar Property of Timing with Possums ('Trichosurus vulpecula').
Many diverse species have demonstrated interval timing, the ability to respond appropriately to time in the range of seconds to minutes, suggesting that an ability to time is adaptive. The peak procedure is a common method of studying interval time perception. In the peak procedure, animals experience a mix of fixed-interval (FI) and extinction (EXT) trials. On EXT trials, responding typically increases to a peak at the time the FI schedule would normally deliver reinforcers before decreasing. Responding on different FI schedules within the peak procedure has been found to conform to Weber's law, whereby response variability is proportional to the length of the fixed interval. We conducted possibly the first investigation of the timing abilities of a marsupial common to Australia and New Zealand, the brushtail possum ('Trichosurus vulpecula'), using FI 15-, 30-, and 60-s schedules of reinforcement in the peak procedure. Response rates on EXT trials peaked at the time of usual reinforcer delivery, decreasing at longer time intervals, and were well fit by 3-parameter Gaussian curves, demonstrating the ability of possums to respond to time-based stimuli. Coefficients of variation suggested that the ability of possums to time was less accurate than that of mammals, but similar to that of birds, invertebrates, and reptiles. Coefficients of variation did not differ consistently over increasing FI intervals, showing that timing responses of possums likely conforms to the scalar property of timing also shown by other species.
Examining the effect of reinforcement on behaviour variability over multiple dimensions in humans
The independence of dimensions of operant responses by humans was investigated in two experiments using a computerized rectangle drawing task from Ross and Neuringer (2002). Variability on the dimensions of area, shape and location was required for reinforcement for one group (VAR); and variability was not required for the other (YOKE). For all three dimensions, U-values, a measure of variability, were higher for the VAR group than for YOKE group; and the number of trials that met the criteria for reinforcement was higher for the VAR group than for the YOKE group. In Experiment 2, reinforcement was contingent on variability on two dimensions regardless of variability on the third. Participants were divided into three groups; each group had one dimension that was not required to vary. U-values were higher for dimensions when reinforcement was contingent on varying shape and location, or area and location. However, U-values did not differ significantly across dimensions when reinforcement was contingent on varying just area and shape. The results of Experiment 1 and 2 are broadly consistent with those of Ross and Neuringer (2002). The importance of orthogonality of dimensions on this task will be discussed.
Motivating operations and animal welfare
Manipulation of unconditioned Motivating Operations (MOs) such as pain, food and water deprivation and the lowering of body weight can be potentially harmful and distressing to animals. Past research that has manipulated such MOs in order to motivate animals to respond for reinforcers is reviewed and directions for future research are discussed. The focus is on the extent that potentially distressing MO's need to be altered in order to change motivation for reinforcers in animals. In addition research directions for investigating alternative MOs that can motivate animals to respond for reinforcers are discussed.
The ability of two internal clock models to predict performance on a temporal bisection task
This research tested the ability of two competing models of animal timing, Learning to Time (LET) and Scalar Expectancy Theory (SET), to predict hens' performance on a temporal bisection task, in a replication of an experiment by Machado and Keen (1999). Hens were trained in two temporal discriminations; in Type 1 trials they learned to choose a red key after a 1-s signal and a green key after a 4-s signal and in Type 2 trials they learned to choose a green key after a 4-s signal, and a yellow key after a 16-s signal. After they learnt these discriminations, intermediate durations were presented. The resulting psychometric function did not superpose, violating the scalar property of timing. When novel key and duration combinations were presented and performance on subsequent generalisation tests closely matched LETS predictions. Overall, the results support the findings of Machado and Keen (1999) and supported LET's rather than SET's predictions.
An Analysis of U-Value as a Measure of Variability
The variability in behavior has frequently been assessed using a measure known as the U-value. Of concern in this article were the limits and constraints on U-value as a measure of variability. The relation between the U-value and aspects of variability was examined using three sets of simulated data. Our analysis demonstrates that the U-value as a measure of variability on its own fails to capture repetitive patterns in the sequence of responding. The U-value was shown to reflect the evenness of the distributions of responses across the categories/options used; however, when the number of categories actually used by the participant differed from the total number available, the relation between U-values and the number of categories allocated with responses was shown to be nonlinear. It was also shown that the same value of U can represent different levels of evenness in response distributions over categories, depending on the number of categories/options actually used. These constraints and limitations are discussed in relation to how researchers might report on behavioral variability.