Mechanisms of irrational preference construction
Konstantinos Tsetsos
Marie Curie Fellow, UKE Hamburg
There is nothing wrong in believing that A (e.g. salad) is better than B (e.g. fish); but there is something awry in believing that A is both better and worse than B. Years of research within the cognitive and decision sciences has demonstrated time and again that human choices are at odds with rational choice theory. For instance, people reverse their initial preference for A (e.g. fruit salad) over B (e.g. today’s cheesecake) when a third irrelevant alternative C (e.g. yesterday’s cheesecake) becomes available. Similarly, people show inconsistent (or intransitive) preference patterns, preferring A over B, B over Cbut C over A. What do these paradoxical behavioural patterns tell us about the neurocognitive processes that underlie preference formation? And, importantly, why have preference formation processes that give rise to irrational decision-making survived natural evolution? These questions have not been sufficiently addressed, partly because preferential decisions (such as choosing to buy or to rent a house) involve sampling and integration over rich value information that is rarely under experimental control. To circumvent this problem, and inspired by research in perception and visual psychophysics, I will present a novel paradigm (termed “value psychophysics”) that abstracts complex decision problems into simple, well-controlled, information-integration experiments. I will show that classical paradoxes such as violations of regularity and transitivity as well as framing and risk biases can be obtained in this simple psychophysical task. Surprisingly, a single mechanism, based on selective attention towards relatively more valuable samples of incoming information, underlies all these phenomena. I will demonstrate that this selective integration mechanism outperforms the statistically optimal choice algorithm (in terms of choice accuracy) under the assumption that moderate levels of noise corrupt the information integration process downstream of sensory representation. Thus, violations of rational choice theory do not disclose fundamental limitations in human processing capacity but they rather reflect adaptive neural computations. I will close by alluding to the neurocomputational and neurochemical basis of selective integration, showing that the pharmacological enhancement of cortical inhibition results in stronger decision irrationality patterns in healthy individuals.