The Relationship Between Reward-Related Motor Adaptation and Sensitivity to Punishment and Reward

The Relationship between Reward-Related Motor Adaptation and Sensitivity to Punishment and Reward.

Julien Khoury

Unit: PSY2061 (Biological Psychology)

Due Date: 18/05/17

Tutor: Oscar Murphy

Lab Class: Tuesday 10am-12pm

Word Count: 1998

Abstract

Contrary to past research, motor adaptation (MA) is not insensitive to motivational feedback. This study explored the effects of both reward and punishment on motor adaptive learning as well as investigating how individual differences in sensitivity to reward (SR) and punishment (SP) can predict how someone will adapt with reinforcement learning. A sample of 151 PSY2061 students, male (n=28) and female (n=123) was used, undergoing testing using the synchronous bimanual force task (MAT). Participants also completed the Sensitivity to Reward and Sensitivity to Punishment Questionnaire to determine whether they were approach or avoidance learners. It was found that approach learners adapted better (M=154.59, SD=43.48) than avoidance learners (M=139.58, SD=39.07). It was also established that SR scores significantly predicted scores on the MAT while there was no correlation between SP scores and corresponding scores on the MAT. The results suggest that pathways associated with reward-based learning have more influence over MA.

Motor adaptation (MA) has been studied and analysed through many different aspects of learning. It is defined as the modification of a movement through trials based on error feedback, whereby the change occurs with repetition and is gradual, and the action remains consistent throughout the trials (Bastian, 2008). Contrary to previous opinions, it does not just involve a simple error cancellation process, rather, it includes calibrations of the brain’s prediction of how the body will move and considers costs associated with the new task demand (Bastian, 2008). Although applications of MA have been heavily researched, it has seldom been considered whether reinforcement learning has any influence over MA, which was previously believed to be insensitive to rewards and punishments.

Reinforcement learning is a branch of learning that focuses on an individual’s use of information from the environment, and how they modify their behaviour to either attain a reward or avoid punishment (Therrien, Wolpert, & Bastian, 2015). People who are more sensitive to reward often learn better through positive consequences (also called approach learners), whereas those said to be more sensitive to punishment learn better through negative consequences (avoidance learners) (Aberg, Doell, & Schwartz, 2016).

Some major structures involved in MA and its associations with reinforcement learning are the basal ganglia and the cerebellum. The ventral striatum, a principal component of the basal ganglia (a group of structures integrated with the thalamus), is involved with the dopaminergic pathway and the planning of motor movement. Furthermore, the cerebellum, an important structure in the back of the brain, plays a role in the mediation of conditioned/reinforced motor responses (Doya, 2000).

Recent studies have explored the idea of independence between different feedback. Doya (2000) and Tseng, Diedrichsen, Krakauer, Shadmehr, and Bastian (2007) established that there were significant differences in areas of brain activity utilised when learning through either reinforcement or error-based adaptation. It was discovered that the basal ganglia were strongly associated with reinforcement learning, with higher brain activity identified around the thalamus. Contrastingly, the cerebellum was coupled with error-based learning. Furthermore, in a study focusing on the associations of the dopamine-reward system and motor adaptation, it was discovered that the cerebellum was involved in externally driven (e.g. visually guided) movement, whereas the basal ganglia were involved in internally generated (memory guided) movement (Doya, 2000). These structures have been heavily researched with results indicating they play a major role in motor adaptation, however investigations into how an individual’s sensitivity to motivational feedback affects motor adaptation, have been a focus in recent studies.

A study conducted by Awenowicz and Porter (2002) investigated which structures in the brain received the densest dopamine innervation. It was found that primarily the motor cortex and smaller motor regions were heavily integrated with dopaminergic pathways, indicating that dopamine has a significant influence over motor activity. It was also shown that dopamine-depleted subjects presented impairments in skilled motor tasks as well as diminished accuracy of skilled movements. These results suggest that, as dopamine is released after receiving a reward, or during reinforcement learning, that motor adaptation would be more heavily influenced and enhanced by reward (Wickens, Reynolds, & Hyland, 2003).

Despite these controversial results, multiple studies including those by Wachter, Lungu, Liu, Willingham, and Ashe (2009) and Aberg, Doell, and Schwartz (2016) have explored the idea that punishment and reward, as well as occupying two different pathways in the brain, influence two different aspects of learning. It has been discovered that due to the ventral striatum and the substantia nigra in the basal ganglia, which is associated with the dopamine-reward system, positive reinforcement in response to a correct motor adaptation improves motor ‘retention’, on average; whereas punishment in response to an incorrect motor adaptation proved to increase the rate or speed of learning (Wachter et al. 2009).

This study aimed to investigate the effects of reinforcement feedback on motor adaptation learning, and further consider to what extent approach or avoidance traits influence sensitivity to motivational feedback.

It is hypothesised that those who are more sensitive to reward, or approach learners will display better motor adaptation, as measured on a motor adaptation task than participants displaying increased sensitivity to punishment (avoidance learners). Furthermore, it is hypothesised that scores on the sensitivity to reward questionnaire will significantly predict corresponding scores on the motor adaptation task. Conversely, it is theorised that scores on the sensitivity to punishment questionnaire will not accurately predict comparable scores on the motor adaptation task.

Method

Participants

This study consisted of 151 participants. As shown in Table 1.1, participants formed two groups, with most consisting of AUS/NZ/European descent. Both consisted of conveniently sampled Monash University students enrolled In the PSY2061 unit. One group contained 56 participants who were deemed more sensitive to reward (SR) and another group of 95 participants who were deemed to have greater sensitivity to punishment (SP). Students did not have to adhere to any inclusionary criteria. The study was approved by the Monash University Human Research Ethics Committee and no informed written consent was needed as all participants were above 18.