Performance monitoring beyond choice tasks: The time course of force execution monitoring investigated by event-related potentials and multivariate pattern analysis

Accurate force production is an essential motor function which, in most cases, requires continuous performance monitoring. Unlike choice-response tasks with two response alternatives, the accuracy in a force production paradigm is defined as an area between an upper and lower limit on the force continuum. In the present study, we investigated the neural mechanisms underlying force production. We used a force production task in which the participants (n = 48) were asked to exert a brief force pulse within a specific force range. This allowed: (1) investigation of action monitoring activity during force execution using response-locked and feedback-locked event-related potential (ERP) components known to be involved in error monitoring; (2) multivariate pattern analysis (MVPA) for ERPs. We found that the different force production ranges (characterised as too low, correct, and too high with respect to the target force range) showed no clear error-specific variations in the ERP components of interest. MVPA, on the other hand, allowed for successful classification, not only between the correct and the incorrect outcome conditions, but also between the two incorrect outcome conditions. This suggests that the classifier identified neural patterns reflecting the force magnitude rather than the correctness of a response. Moreover, additional support-vector regression (SVR) analyses showed that single-trial response parameters (i.e. peak force and time-to-peak) could be decoded from the brain activity pattern starting from 140 ms (for peak force) and 270 ms (for time-to-peak) before the response onset. These results indicate that the motor program defined the magnitude and timing of the force pulse before response execution, while the correctness of that response (in relation to the “default force” required) was not yet foreshadowed in neural signals. Finally, this study presents the first evidence of a post-error force adjustment mechanism, for which participants produced a higher force in trials after under-producing the required force, and a lower force in trials after over-producing the required force. •Neural activity predicts performance parameters in force execution.•Distributed patterns of ERPs predictive for specific force rather than correctness.•Peak force (PF) and time-to-peak (TTP) were decodable already before response onset.•Systematic force adjustments were observed after committing errors.