Postdecision consolidation in a trend prediction task

To be able to learn from experience it is necessary to correctly apprehend experienced feedback and the situation in which it is provided. The results indicate how post-decision consolidation in complex domains may affect learning. The problem may be particularly pertinent in recurrent decision making where considerable risk is involved. The study explores the changes in aspect (signal) importance from pre- to postdiction as a function of outcome information. By postdiction we mean the remembering of an earlier prediction (cf. Hawkins & Hastie, 1990). Subjects were asked to decide on which of four alternative future price developments would follow a historical price trajectory for different commodities, and to rate the importance of each of the chosen alternative's corresponding aspects. The subjects revealed a bias in their support ratings of aspects--seeing support in aspects that traditionally (by themselves and in many contexts) would be seen as neutral or even counter-indicative of the alternative chosen. After an intermission, the subjects were also given information about what was indicated to be the actual development of the market. One group was told that their decisions were correct (irrespective of what the decisions were), another group that they were incorrect but close, a third group that they were incorrect by far, while a fourth group served as a control. Following this information the subjects were again asked to judge the importance of the aspects for their own prior decision on the most likely future development. The results indicated that outcome feed-back had an effect on post decision restructuring of facts. Subjects in the correct condition showed an average consolidation that increased the support, while the wrong conditions lead to negative consolidation (in retrospect indicating that they never found as much support for their decision in the past as they actually did). Thus, in a choice between consolidating their own initial prediction and the price trajectory they would have to live with, the decision makers consolidated the outcome. Therefore, the results of the study were related to the hindsight bias phenomenon (Fischhoff, 1975) and to Kahneman and Miller's (1986) mutability concept.     

Effects of reinforcement history on response rate and response pattern in periodic reinforcement

Several researchers have suggested that conditioning history may have long-term effects on fixed-interval performances of rats. To test this idea and to identify possible factors involved in temporal control development, groups of rats initially were exposed to different reinforcement schedules: continuous, fixed-time, and random-interval. Afterwards, half of the rats in each group were studied on a fixed-interval 30-s schedule of reinforcement and the other half on a fixed-interval 90-s schedule of reinforcement. No evidence of long-term effects attributable to conditioning history on either response output or response patterning was found; history effects were transitory. Different tendencies in trajectory across sessions were observed for measures of early and late responding within the interreinforcer interval, suggesting that temporal control is the result of two separate processes: one involved in response output and the other in time allocation of responding and not responding.     

Prefrontal brain activity predicts temporally extended decision-making behavior

Although functional neuroimaging studies of human decision-making processes are increasingly common, most of the research in this area has relied on passive tasks that generate little individual variability. Relatively little attention has been paid to the ability of brain activity to predict overt behavior. Using functional magnetic resonance imaging (fMRI), we investigated the neural mechanisms underlying behavior during a dynamic decision task that required subjects to select smaller, short-term monetary payoffs in order to receive larger, long-term gains. The number of trials over which the longterm gains accrued was manipulated experimentally (2 versus 12). Event-related neural activity in right lateral prefrontal cortex, a region associated with high-level cognitive processing, selectively predicted choice behavior in both conditions, whereas insular cortex responded to fluctuations in amount of reward but did not predict choice behavior. These results demonstrate the utility of a functional neuroimaging approach in behavioral psychology, showing that (a) highly circumscribed brain regions are capable of predicting complex choice behavior, and (b) fMRI has the ability to dissociate the contributions of different neural mechanisms to particular behavioral tasks.     

Fitting Psychometric Models with Methods Based on Automatic Differentiation

Quantitative psychology is concerned with the development and application of mathematical models in the behavioral sciences. Over time, models have become more complex, a consequence of the increasing complexity of research designs and experimental data, which is also a consequence of the utility of mathematical models in the science. As models have become more elaborate, the problems of estimating them have become increasingly challenging. This paper gives an introduction to a computing tool called automatic differentiation that is useful in calculating derivatives needed to estimate a model. As its name implies, automatic differentiation works in a routine way to produce derivatives accurately and quickly. Because so many features of model development require derivatives, the method has considerable potential in psychometric work. This paper reviews several examples to demonstrate how the methodology can be applied. From the Presidential Address delivered at the 70th Annual Meeting of the Psychometric Society, Tilburg University, The Netherlands, July 5–8, 2005.     

时间记忆层次网络模型的实验检验

实验采用分类性实验材料,使被试在对词单项目加工时形成较清晰的群集,进而形成时间组块,目的在于对时间记忆层次网络进行直接检验。其中时序判断用反时间时和正确率两种反应指标,时距估计用再现和口头估计两种方法。结果发现,时间信息记忆既存在层次网络的特征,又存在线性结构的特征。     

Support for a theory of memory for event duration must distinguish between test-trial ambiguity and actual memory loss

Staddon and Higa's (1999) trace-strength theory of timing and memory for event duration can account for pigeons' bias to "choose short" when retention intervals are introduced and to "choose long" when, following training with a fixed retention interval, retention intervals are shortened. However, it does not account for the failure of pigeons to choose short when the intertrial interval is distinct from the retention interval. That finding suggests that stimulus generalization (or ambiguity) between the intertrial interval and the retention interval may result in an effect that has been attributed to memory loss. Such artifacts must be eliminated before a theory of memory for event duration can be adequately tested.     

Individual differences in vulnerability to subjective time distortion1

Time duration is perceived to be longer when accompanied by dynamic sensory stimulation than when accompanied by static stimulation. This distortion of time perception is thought to be due to the acceleration of an internal pacemaker that has been assumed to be the main component of temporal judgments. In order to investigate whether the function of the internal pacemaker is modality dependent or independent, we examined the correlation of visual flicker and auditory flutter effects on a temporal production task. While seeing a 10‐Hz visual flicker or hearing a 10‐Hz auditory flutter, participants estimated a duration of 2500 ms as accurately as possible by pressing a button. The results showed a significant within‐individual correlation between the time distortion due to visual flicker and that due to auditory flutter. Additionally, we found that time distortion due to auditory flutter tended to be larger in female participants than in male participants. These results suggest that the mechanisms underlying subjective time dilation are similar between vision and audition within individuals, but that they vary across individuals.     

The interaction of temporal generalization gradients predicts the context effect

In a temporal double bisection task, animals learn two discriminations. In the presence of Red and Green keys, responses to Red are reinforced after 1-s samples and responses to Green are reinforced after 4-s samples; in the presence of Blue and Yellow keys, responses to Blue are reinforced after 4-s samples and responses to Yellow are reinforced after 16-s samples. Subsequently, given a choice between Green and Blue, the probability of choosing Green increases with the sample duration-the context effect. In the present study we asked whether this effect could be predicted from the stimulus generalization gradients induced by the two basic discriminations. Six pigeons learned to peck Green following 4-s samples (S(+)) but not following 1-s samples (S(-)) and to peck Red following 4-s samples (S(+)) but not following 16-s samples (S(-)). Temporal generalization gradients for Green and Red were then obtained. Finally, the pigeons were given a choice between Green and Red following sample durations ranging from 1 to 16 s. Results showed that a) the two generalization gradients had the minimum at the S(-) duration, an intermediate value between the S(-) and the S(+) durations, and the maximum at the S(+) as well as more extreme durations; b) on choice trials, preference for Green over Red increased with sample duration, the context effect; and c) the two generalization gradients predicted the average context effect well. The Learning-to-Time model accounts for the major trends in the data.