A hierarchical system for learning and control

点击翻译按钮获取中文摘要

Even the simplest of movements are monitored and evaluated by multiple systems within the brain. For instance, if we try to pick up a cup of coffee and fail for some reason, we need a neural system that can identify this and generate a learning signal that would be used to modify subsequent behaviour. However, this is not the only type of error we might encounter while reaching for a cup of coffee. Unbeknownst to us, the motor plan - the set of instructions that the brain issues to the motor system to carry out the movement - may contain errors. Additionally, errors in the motor plan may be brought about by unexpected changes in the environment - imagine the cup of coffee is poorly balanced and that it begins to slip as we reach for it. These types of errors - errors in planning and/or errors brought about by unexpected changes in the environment need to be rapidly mediated within the movement itself in order for us to successfully achieve our movement goal. Recently, I have proposed the hierarchical error processing hypothesis which states that there are two distinct error processing systems in the brain: a system that is responsible for evaluating the success or failure of our actions, and a system that is tasked with resolving errors in the motor plan within the movement. A series of recent studies by my research group supports the hierarchical hypothesis. In sum, the results demonstrate that a learning system within frontal-central parts of the brain plays a key role in the evaluation of motor output. Existing experimental evidence also supports the hierarchical hypothesis as it demonstrates the role of error systems in posterior regions of the brain in the resolution of errors in the initial motor plan within an ongoing movement. The research that forms the basis of this proposal seeks to further develop the hierarchical hypothesis by identifying further neural markers of error evaluation and examining their impact on behaviour. Further, I hope to improve the theoretical framework which underlies the hierarchical hypothesis. Importantly, the proposed research will further our understanding of how learning occurs at the neural level, thus providing results which will be of interest to the scientific community and the general population.