The neural basis of expert anticipation in a real-world skill

In many sports such as tennis, football, and basketball, the ability to anticipate the actions of an opponent is a key component of expert performance. Behavioural research has shown that expert athletes are able to identify and comprehend anticipatory cues present in the kinematics of an opposing player and use them to anticipate the end goal of an action. The aim of the work reported here was to investigate the neural basis of this anticipation skill in athletes. Specifically, the thesis focuses on the role of the sensorimotor system in facilitating this ability, based on evidence showing that the sensorimotor system is involved in the understanding of other people’s actions.
First, two behavioural experiments are presented showing the development of an anticipation test that is able to distinguish expert and novice participants based upon their response accuracy. Then the role of sensorimotor activity in facilitating anticipation skill is investigated using EEG. Event-related power changes in cortical sensorimotor oscillations in the mu (8-13Hz) and beta (15-25Hz) frequency bands are used as indices of sensorimotor activity, based on the findings of previous work. It was found that earlier and greater event-related desynchronisation (ERD) occurred in the expert group, compared to the novices, in both the mu and beta frequency bands. This suggests greater use of the sensorimotor system during action anticipation in athletes, whilst viewing domain specific actions. However, traditional channel-based analyses of this measure are flawed in that volume conduction effects mean mu and non-mu alpha activity can become mixed. This means it is unclear the extent to which mu activity specifically indexes the sensorimotor system, as opposed to other processes such as attentional demand.
As a potential solution to this issue, the data was re-analysed using independent component analysis (ICA) to separate out the underlying brain processes ongoing during the anticipation task. Expertise-related differences in mu and beta ERD were then analysed on independent component (IC) activity. The ICA analysis largely replicated the channel analysis, with earlier and greater ERD in expert athletes in ICs relating the sensorimotor activity. No group differences were found in ICs relating to non-mu, alpha activity. This suggests group differences were specific to sensorimotor activity, providing evidence that sensorimotor activity is key in distinguishing expert from novice athletes on an action anticipation task.
In Chapter 5, a novel analysis method was used. Mu and beta ERD in the two groups were contrasted between trials that were subsequently anticipated correctly versus trials that were subsequently anticipated incorrectly. In the experienced group only, it was found that there was greater beta ERD for correctly anticipated trials compared to incorrectly anticipated trials. There were no differences in mu ERD between correct and incorrect trials. Furthermore, overall mu and beta ERD significantly negatively correlated with anticipation accuracy in the experienced group, suggesting a higher accuracy is associated with greater sensorimotor ERD during prior observation. Finally, in Chapter 6 the results are summarised and their implications within the wider literature considered, as well as a discussion of limitations and future directions.