Abstract
Every day we face the world with beliefs about the rules that govern our environment and what will happen if we take particular actions. When incoming information does not match our predictions, we either need to abandon or update these beliefs. There is accumulating evidence that the right hemisphere is responsible for processing the statistical properties of an uncertain environment, which is important for building accurate representations of our environment, and adapting those representations when necessary. However, mental models may be more than 'look-up tables’ of conditional probabilities. Hence, updating failures should also be found when probability learning is less critical. To evaluate this hypothesis, we used a series of pictures that were based on well-known ambiguous figures (e.g., rabbit/duck). Participants saw pictures of unambiguous objects (e.g., rabbit) that incrementally changed over successive presentations to eventually show different unambiguous figures (e.g., duck). The point of transition from reporting the first object to reporting the second, provided an index of updating. RBD patients (n = 16) took significantly longer to switch their reports from the first unambiguous picture (e.g., "it is a rabbit") to the second (e.g., "now it is a duck") than did healthy controls (n = 18) [F(1,31) = 22.55, p <.001, h[sup]2[/sup] = .42]. This failure of updating occurred over a short time scale (15 pictures) and was not dependent on statistical learning. Other tests confirmed that results were neither due to a higher tendency to perseverate nor due to general cognitive impairment of the RBD patients. These findings are in accordance with a more generic role for the right hemisphere in model building and updating beyond what comes from the simple amalgamation of probabilities.
Meeting abstract presented at VSS 2013