Alterations in the amplitude and burst rate of beta oscillations impair reward-dependent motor learning in anxiety

Elife. 2020 May 19:9:e50654. doi: 10.7554/eLife.50654.

Abstract

Anxiety results in sub-optimal motor learning, but the precise mechanisms through which this effect occurs remain unknown. Using a motor sequence learning paradigm with separate phases for initial exploration and reward-based learning, we show that anxiety states in humans impair learning by attenuating the update of reward estimates. Further, when such estimates are perceived as unstable over time (volatility), anxiety constrains adaptive behavioral changes. Neurally, anxiety during initial exploration increased the amplitude and the rate of long bursts of sensorimotor and prefrontal beta oscillations (13-30 Hz). These changes extended to the subsequent learning phase, where phasic increases in beta power and burst rate following reward feedback were linked to smaller updates in reward estimates, with a higher anxiety-related increase explaining the attenuated belief updating. These data suggest that state anxiety alters the dynamics of beta oscillations during reward processing, thereby impairing proper updating of motor predictions when learning in unstable environments.

Keywords: anxiety; beta oscillations; human; motor learning; neuroscience; reward.

Plain language summary

Feeling anxious can hinder how well someone performs a task, a phenomenon that is sometimes called “choking under pressure”. Anxiety may also impair a person’s ability to learn a new manual task, like juggling or playing the piano; however, it remains unclear exactly how this happens. People learn manual tasks more quickly if they can practice first, and the more someone varies their movements during these trial runs, the faster they learn afterwards. Yet, anxiety can affect movement; for example, anxious people often make repetitive motions like hand-wringing or fidgeting. There is also evidence that very anxious people may learn less from the outcomes of their actions. To understand how anxiety may affect the learning of manual tasks, Sporn et al designed experiments where people learned to play a short sequence of notes on a piano. The main experiment involved 60 participants and was split over two phases. In the first ‘exploration’ phase, participants had to play the piano sequence using any timing they liked and were encouraged to explore different rhythms. In the second ‘learning’ phase, participants were rewarded with a higher score the closer they got to playing the notes with a certain rhythm, without being told that this was their specific goal. To see how anxiety affected performance, the participants were split into three groups. One group were told in the initial exploration phase that they would give a public talk after they completed the piano task, which reliably made them more anxious. A second group were told about the anxiety-inducing public speaking only during the learning phase; while a third group – the controls – were not aware of any public speaking task. People in the second group could learn the rhythm as well as the controls. Participants who were made anxious during the exploration phase, however, scored fewer points and were less likely to learn the piano sequence in the second phase. They also varied their movements less in the first phase. As a follow-up, Sporn et al. repeated the experiment with 26 people but without the initial exploration phase. This time the anxious participants were less able to learn the piano sequence and scored fewer points. This suggests that the initial exploration in the previous experiment had enabled later anxious participants to succeed in the learning phase despite being anxious. Finally, Sporn et al. also used a technique called electroencephalography (or EEG for short) to record brain activity and observed differences in participants with and without anxiety, particularly when they received their scores. The EEG signals showed that anxiety altered rhythmic patterns of brain activity called “sensorimotor beta oscillations”, which are known to be involved in both movement and learning.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adolescent
  • Adult
  • Anxiety / diagnosis
  • Anxiety / physiopathology
  • Anxiety / psychology*
  • Beta Rhythm*
  • Case-Control Studies
  • Cerebral Cortex / physiopathology*
  • Female
  • Humans
  • Learning*
  • Male
  • Models, Psychological
  • Motor Skills*
  • Reward*
  • Time Factors
  • Young Adult