Image credit: Public domain (CC0)
How do you decide whether to buy a new car? One factor to consider is how well the economy is doing. During an economic boom, you might happily commit to buying a new vehicle that goes on sale, but prefer to sit on your savings during a financial crisis, despite how good the offer may be. Adjusting how you make decisions in situations like this can help you optimize choices in an ever-changing world.
It’s currently thought that when deciding, we accumulate evidence for each of the available options. When evidence for one of the options passes a threshold, we choose that option. External factors – such as a booming economy when considering buying a car – could bias this process in two different ways. The standard view is that they move the starting point of evidence accumulation towards one of the two choices, so that the threshold for choosing that option is more easily reached. Alternatively, they could bias the accumulation process itself, so that evidence builds up more quickly towards one of the choices.
To distinguish between these possibilities, Kloosterman et al. asked volunteers to press a button whenever they detected a target hidden among a stream of visual patterns. To bias their decisions, volunteers were penalized differently in two experimental conditions: either when they failed to report a target (a ‘miss’), or when they ‘detected’ a target when in fact nothing was there (a ‘false alarm’). As expected, punishing participants for missing a target made them more liberal towards reporting targets, whereas penalizing false alarms made them more conservative.
Computational modeling of behavior revealed that when participants used a liberal strategy, they did not move closer to the threshold for deciding target presence. Instead, they accumulated evidence for target presence at a faster rate, even when in fact no target was shown. Brain activity recorded during this task reveals how this bias in evidence accumulation might come about. When a volunteer adopted a liberal response strategy, visual brain areas showed a reduction in low-frequency ‘alpha’ waves, suggesting increased attention. This in turn triggered an increase in high-frequency ‘gamma’ waves, reflecting biased evidence accumulation for target presence (irrespective of whether a target actually appeared or not).
Overall, the findings reported by Kloosterman et al. suggest that we can strategically bias perceptual decision-making by varying how quickly we accumulate evidence in favor of different response options. This might explain how we are able to adapt our decisions to environments that differ in payoffs and punishments. The next challenge is to understand whether such biases also affect high-level decisions, for example, when purchasing a new car.
