Abstract
Current behavior is largely determined by behavioral history and it’s consequences. How environmental rewards can shape the frequency and quality of behavior, has, however, only recently come under the scrutiny of attention researchers. In this domain, it is not obvious which components of attention are affected by reward, and whether the effect involves general enhancement or is specific to discrete components of attention. Observers viewed brief displays of differentially colored letters and reported their identity. Each color signified a consistent monetary value and we measured the accuracy of identification under different color-pairing conditions. At the end of the session, observers were paid the balance earned during the experiment. By fitting a model based on the Theory of Visual Attention (Bundesen, 1990) to the data, we estimated processing speed, selectivity, visual short-term memory and the threshold for perception. Our primary hypothesis was that observers could, under data-limited conditions via brief exposure, distribute their attentional resources according to the value of the stimuli, i.e. that selectivity would be higher for high-value over lower-value targets. Importantly, our design was balanced so that the expected utility of uninformed guessing was zero, yielding no incentive to employ value-dependent response criteria. We also tested value-dependent effects on the capacity of visual-short term memory. Finally, we tested for motivational salience effects, by including conditions with color-contingent negative values. This gave an opportunity to compare high-gain with high-loss conditions. We found clear effects of value on selectivity when comparing high- and low-value conditions. When comparing equally valuable high-loss and high-gain conditions there were indications of risk-aversion, consistent with results from behavioral economics. We show that the expected value of target selection shapes the deployment of resources at very low exposure durations and can increase the capacity of VSTM in a paradigm untainted by post-perceptual effects.
Meeting abstract presented at VSS 2014