Errors in Imagined and Executed Typing

Abstract

In motor imagery (MI), internal models may predict the action effects. A mismatch between predicted and intended action effects may result in error detection. To compare error detection in MI and motor execution (ME), ten-finger typists and hunt-and-peck typists performed a copy-typing task. Visibility of the screen and visibility of the keyboard were manipulated. Participants reported what type of error occurred and by which sources they detected the error. With covered screen, fewer errors were reported, showing the importance of distal action effects for error detection. With covered screen, the number of reported higher-order planning errors did not significantly differ between MI and ME. However, the number of reported motor command errors was lower in MI than in ME. Hence, only errors that occur in advance to internal modeling are equally observed in MI and ME. MI may require more attention than ME, leaving fewer resources to monitor motor command errors in MI. In comparison to hunt-and-peck typists, ten-finger typists detected more higher-order planning errors by kinesthesis/touch and fewer motor command errors by vision of the keyboard. The use of sources for error detection did not significantly differ between MI and ME, indicating similar mechanisms.

Are observed effects of movement simulated during motor imagery performance?

“…Results on a dart-throwing task whereby participants were to imagine themselves “hitting the bulls-eye” revealed that self-report of the position of their dart on the board after motor imagery performance was inaccurate (Dahm & Rieger, 2019b). This finding demonstrated that motor imagery performance can result in a subjectively experienced error, in line with past work, including that demonstrating the commission of errors in imagined typing (Dahm & Rieger, 2019a). Additionally, the degree of reported error on the task when performed via motor imagery was correlated with error produced in overt execution and was additionally modulated by experience with the task (Dahm & Rieger, 2019b).…”

“…Therefore, learning can only occur through comparison of predicted against intended effect (Circle 2). C) Theorized changes to the Dahm and Rieger (2019a) model of motor learning in motor imagery to account for the errors in the predicted effects of movement (gray dashed lines). In this model observed effects are simulated via motor imagery performance alongside the predicted effects arising from the forward model enabling comparisons represented by Circles 1 (predicted vs. observed effects) and 3 (observed vs. intended effects) to contribute to learning occurring via motor imagery.…”

Expectations affect the contribution of tonic global inhibition, but not of phasic global inhibition to motor imagery.

“…However, inhibitory aftereffects seem to decay quickly (Bart et al, 2021a). Further, despite the requirement to inhibit actual actions, forward modeling still occurs during motor imagery (Dahm & Rieger, 2019a, 2019b; Kilteni et al, 2018). Thus, motor and sensory consequences of the action are emulated (Kilteni et al, 2018), which is part of the reason why motor imagery practice is effective.…”

“…This prediction can be compared with intended action consequences or actual action consequences to monitor one's performance and detect or correct movement errors (Frith et al, 2000). Similarly, it seems that in motor imagery, even though no actual actions occur, an efference copy is generated and the end state of the body after the imagined action as well as the sensory consequences of the imagined action are predicted (Dahm & Rieger, 2019a, 2019bDavidson & Wolpert, 2005;Grush, 2004;Kilteni et al, 2018;Miall & Wolpert, 1996). This raises the question at which time point inhibition during motor imagery sets in, that is, prior or after the generation of an efference copy (cf.…”

“…Similarities between motor imagery and motor execution on neurophysiological (Beyer et al, 1990; Decety et al, 1993; Lotze & Halsband, 2006; Munzert et al, 2009) and behavioral (Dahm & Rieger, 2016b; Guillot & Collet, 2005; Guillot, Hoyek, et al, 2012) levels have been reported. However, differences between motor imagery and motor execution also exist (Dahm & Rieger, 2019a, 2019b; Guillot & Collet, 2005; Guillot, Di Rienzo, et al, 2012). One apparent difference is the necessity to prevent actual actions in motor imagery (Bart et al, 2021b; Guillot, Di Rienzo, et al, 2012; Jeannerod, 2001; O’Shea & Moran, 2017, 2018; Rieger et al, 2017; Scheil & Liefooghe, 2018; Scheil et al, 2020).…”