Distinct Neural Signatures of Outcome Monitoring following Selection and Execution Errors

Abstract

1Losing a point playing tennis may result from poor shot selection or poor stroke execution. To 2 explore how the brain responds to these different types of errors, we examined EEG 3 signatures of feedback-related processing while participants performed a simple decision-4 making task. In Experiment 1, we used a task in which unrewarded outcomes were framed as 5 selection errors, similar to how feedback information is treated in most studies. Consistent 6 with previous work, EEG differences between rewarded and unrewarded trials in the medial 7 frontal negativity (MFN) correlated with behavioral adjustment. In Experiment 2, the task 8 was modified such that unrewarded outcomes could arise from either poor execution or 9 selection. For selection errors, the results replicated that observed in Experiment 1. However, 10 unrewarded outcomes attributed to poor execution produced larger amplitude MFN, 11 alongside an attenuation in activity preceding this component and a subsequent enhanced 12 error positivity (Pe) response in posterior sites. In terms of behavioral correlates, only the 13 degree of the early attenuation and amplitude of the Pe correlated with behavioral adjustment 14 following execution errors relative to reward; the amplitude of the MFN did not correlate with 15 behavioral changes related to execution errors. These results indicate the existence of distinct 16 neural correlates of selection and execution error processing and are consistent with the 17 hypothesis that execution errors can modulate action selection evaluation. More generally, 18 they provide insight into how the brain responds to different classes of error that determine 19 future action. 20 Significance Statement 21To learn from mistakes, we must resolve whether decisions that fail to produce rewards are 22 due to poorly selected action plans or badly executed movements. EEG data were obtained to 23 identify and compare the physiological correlates of selection and execution errors, and how 24 these are related to behavioral changes. A neural signature associated with reinforcement 25 learning, a medial frontal negative (MFN) ERP deflection, correlated with behavioral 26 adjustment after selection errors relative to reward outcomes, but not motor execution 27 errors. In contrast, activity preceding and following the MFN response correlated with 28 behavioral adjustment after execution errors relative to reward. These results provide novel 29 insight into how the brain responds to different classes of error that determine future action. 30

Feedback processing in cognitive and motor tasks: A meta‐analysis on the feedback‐related negativity

“…For example, when throwing a ball to a specific target, the wrong target could have been chosen, but missing the target could also be induced by a wrong release point, a wrong estimation of the ball's flight curve, the distance to the target area, the control of the muscular system, the wrong standing position, a slippery surface, etc. It has been shown that feedback focusing on execution errors induces higher FRN amplitudes compared to feedback that is related to selection errors (Mushtaq et al, 2022). This supports the present finding that the FRN amplitude was higher in motor tasks, where feedback is related to movement execution (and action selection) than in cognitive tasks, where feedback is related exclusively to action selection.…”

“…Errors in cognitive tasks are commonly assigned to action selection for highly automatized (and simple) button presses. In contrast, errors in motor tasks can arise for different task elements during action selection, as well as during movement execution (De Bruijn et al, 2003;Maurer et al, 2021;Mushtaq et al, 2022). For example, when throwing a ball to a specific target, the wrong target could have been chosen, but missing the target could also be induced by a wrong release point, a wrong estimation of the ball's flight curve, the distance to the target area, the control of the muscular system, the wrong standing position, a slippery surface, etc.…”

“…This supports the present finding that the FRN amplitude was higher in motor tasks, where feedback is related to movement execution (and action selection) than in cognitive tasks, where feedback is related exclusively to action selection. Mushtaq et al (2022) argued that the higher FRN amplitude following execution errors could also be explained by a higher amount of agency following these errors, such that participants assume a higher level of control during action execution errors. In addition, errors are divided into low-level errors and high-level errors.…”

“…Lastly, it is possible that domain-specific differences could have been induced by P300 effects, a component which overlaps with the FRN. It has been previously shown that execution and selection errors are driven by reward-related differences in both domains (De Bruijn et al, 2003;Mushtaq et al, 2022). Additional research with motor tasks and cognitive tasks or a meta-analysis on P300 effects is needed to shed more light on this issue.…”

“…Neural correlates of error processing are reflected in event-related potentials of the electroencephalogram (EEG) signal and help to understand the neuro-cognitive underpinnings of feedback-dependent learning. The feedback-related negativity (FRN), which is also called reward positivity (RewP; Krigolson, 2018), and less often called medialfrontal negativity (MFN; e.g., Mushtaq et al, 2022), is the most prominent component associated to reinforcement learning Miltner et al, 1997;Schultz, 2002). The FRN can be elicited at frontal sites 200-300 ms after feedback onset, being more negative after negative feedback than after positive feedback.…”