Thursday, February 12, 2009

Pavlovian conditioning can transfer from the virtual world to the real world

McCabe et al. offer an intriguing experiment showing that conditioning-dependent motivational properties can transfer from a computer game to the real world and also be expressed in terms of brain responses measured using functional magnetic resonance imaging (fMRI). They studied healthy participants conditioned with aversive and appetitive drinks in the context of a virtual cycling race. Three days after conditioning, participants returned for a fMRI session. They used this opportunity to observe the impact of incidental presentation of conditioned stimuli on a real-world decision (seat choice, see the figures below). They found a significant influence of conditioning on seat choice and, moreover, noted that individual susceptibility to this influence was reflected in differential insula cortex responses during subsequent scanning. Thus a stimulus in a virtual environment can acquire motivational properties that persist and modify behavior in the real world.

Figure - Day 1: Pavlovian conditioning in virtual environment. Participants in the virtual cycle race were overtaken by competitors. The stimuli on the competitors' jerseys acted as CSs predicting the delivery of either pleasant or unpleasant juice. The stimulus-to-juice assignment was counterbalanced across participants.

Figure - Day 4: Real-world decision when asked to take a seat in an unoccupied waiting room before scanning. Sixteen participants chose the seat bearing a towel with the CS+app.


  1. Let's be fair here: both of these happened in the real world. One also happened in the fake world. But the fake world is a part of the real world.

    Let's say you display a logo on a screen, and deliver the conditioning based on which logo is shown. How is this any different from displaying the logo on the back of an animated bicyclist, shown on the screen? It's not.

    Colour me unimpressed.

  2. Anonymous2:41 AM

    It's significant because the context is different.

    The subject existed in the real world, yes. But the context was the virtual world. The fact the subjects displayed preference outside the virtual context is the element that was important.

    Having said this, you're point is still valid - but it presents opportunities for more interesting research.