Experimental Investigations on risk taking in sensorimotor decision making

It has long been known that humans display a number of cognitive biases when they are asked to select from a series of possible choice alternatives in economic decision-making tasks. In stark contrast however, the last two decades of research on sensorimotor control have shown that humans are able to exhibit statistically optimal behaviours (i.e. select the most appropriate action from a repertoire of options) when they act on the environment. Given that many critical decisions in the real world require not only selection but also action, it is surprising that there has been little crosstalk between these areas of research. This thesis bridges this gap in the literature by introducing a new experimental framework that allows the manipulation of features related to action selection and execution to understand how these processes interact to manifest in decision-making. Specifically, a novel multi-trial decision-making task is developed where participants are asked to select from a series of options with varying levels of risk and reward (equivalent to choice selection in economic decision-making) and implement their choices with actions that place demands on the sensorimotor system. One particularly prevalent bias in the decision-making literature is risk aversion under uncertainty, but recent studies have shown that this phenomenon can be reversed when the task is reframed as a sensorimotor reaching task. This thesis examines this bias in detail by experimentally manipulating key component parts in choice selection and action execution. There is a particular focus on the role of agency, feedback (either veridical or predetermined), motor competence and learning. Visual execution error has consistenly resulted in riskier behaviour as well as better error correction. Together, this work demonstrates the interplay between cognitive and sensorimotor systems in choice selection by illustrating the bilateral relationship between parameters driving action selection and execution interact to produce decision-making.