The role of conflict, feedback, and action comprehension in monitoring of action errors: Evidence for internal and external routes

Paradoxical decrease of imitation performance with age in children

Imitation development was studied in a cross-sectional design involving 174 primary-school children (aged 6-10), focusing on the effect of actions' complexity and error analysis to infer the underlying cognitive processes. Participants had to imitate the model's actions as if they were in front of a mirror ('specularly'). Complexity varied across three levels: movements of a single limb; arm and leg of the same body side; or arm and leg of opposite body sides. While the overall error rate decreased with age, this was not true of all error categories. The rate of 'side' errors (using a limb of the wrong body side) paradoxically increased with age (from 9 years). However, with increasing age, the error rate also became less sensitive to the complexity of the action. This pattern is consistent with the hypothesis that older children have the working memory (WM) resources and the body knowledge necessary to imitate 'anatomically', which leads to additional side errors. Younger children might be paradoxically free from such interference because their WM and/or body knowledge are insufficient for anatomical imitation. Yet, their limited WM resources would prevent them from successfully managing the conflict between spatial codes involved in complex actions (e.g. moving the left arm and the right leg). We also found evidence that action side and content might be stored in separate short-term memory (STM) systems: increasing the number of sides to be encoded only affected side retrieval, but not content retrieval; symmetrically, increasing the content (number of movements) of the action only affected content retrieval, but not side retrieval. In conclusion, results suggest that anatomical imitation might interfere with specular imitation at age 9 and that STM storages for side and content of actions are separate.

Action imitation via trajectory-based or posture-based planning

Imitation is a significant daily activity involved in social interaction and motor learning. Imitation has been theorized to be performed in at least two ways. In posture-based imitation, individuals reproduce how the body should look and feel, and are sensitive to the relative positioning of body parts. In trajectory imitation, individuals mimic the spatiotemporal motion path of the end effector. There are clear anecdotal situations in which one might benefit from imitating postures (when learning ballet) or trajectories (when learning to reach around objects). However, whether these are in fact distinct methods of imitation, and if so, whether they may be applied interchangeably to perform the same task, remain unknown. If these are indeed separate mechanisms that rely on different computational and neural resources, a cost should be incurred when switching from using one mechanism to the other within the context of a single task. Therefore, observing a processing cost would both provide evidence that these are indeed two distinct mechanisms, and that they may be used interchangeably when trying to imitate the same stimulus. To test this, twenty-five healthy young adults performed a sequential multitasking imitation task. Participants were first instructed to pay attention to the limb postures or the hand path of a video-recorded model, then performed a neutral, congruent, or incongruent intervening motor task. Finally, participants imitated the modeled movement. We examined both spatial and temporal imitation accuracy as well as individual spatial consistency. When the primary task involved imitating trajectories, analysis of individual consistency suggested a processing cost: movements following the posture-matching intervening task were less consistent with baseline (neutral) performance, suggesting performance may be disrupted by the incongruence. This effect was not observed when imitating limb postures. In summary, we present initial evidence for a difference between posture matching and trajectory imitation as a result of instructions and intervening tasks that is consistent with the existence of two computationally distinct imitation mechanisms.

Proprioception-based movement goals support imitation and are disrupted in apraxia

The ability to imitate observed actions serves as an efficient method for learning novel movements and is specifically impaired (without concomitant gross motor impairments) in the neurological disorder of limb apraxia, a disorder common after left hemisphere stroke. Research with apraxic patients has advanced our understanding of how people imitate. However, the role of proprioception in imitation has been rarely assessed directly. Prior work has proposed that proprioceptively sensed body position is transformed into a visual format, supporting the attainment of a desired imitation goal represented visually (i.e., how the movement should look when performed). In contrast, we hypothesized a more direct role for proprioception: we suggest that movement goals are also represented proprioceptively (i.e., how a desired movement should feel when performed), and the ability to represent or access such proprioceptive goals is deficient in apraxia. Using a novel imitation task in which a robot cued meaningless trajectories proprioceptively or visually, we probed the role of each sensory modality. We found that patients with left hemisphere stroke were disproportionately worse than controls at imitating when cued proprioceptively versus visually. This proprioceptive versus visual disparity was associated with apraxia severity as assessed by a traditional imitation task, but could not be explained by general proprioceptive impairment or speed-accuracy trade-offs. These data suggest that successful imitation depends in part on the ability to represent movement goals in terms of how those movements should feel, and that deficits in this ability contribute to imitation impairments in patients with apraxia.

Preserved performance monitoring and error detection in left hemisphere stroke

Depending on the lesion site, a stroke typically affects various aspects of cognitive control. While executing a task, the performance monitoring system constantly compares an intended action plan with the executed action and thereby registers inaccurate actions in case of any mismatch. When errors occur, the performance monitoring system signals the need for more cognitive control, which is most efficient when the subject notices errors rather than processing them subconsciously. The current study aimed to investigate performance monitoring and error detection in a large sample of patients with left hemisphere (LH) stroke. In addition to clinical and neuropsychological tests, 24 LH stroke patients and 32 healthy age-matched controls performed a Go/Nogo task with simultaneous electroencephalography (EEG) measurements. This set-up enabled us to compare performance monitoring at the behavioral and the neural level. EEG data were analyzed using event-related potentials [ERPs; e.g., the error-related negativity (Ne/ERN) and error positivity (Pe)] and additionally more sensitive whole-brain multivariate pattern classification analyses (MVPA). We hypothesized that LH stroke patients would show behavioural deficits in error detection when compared to healthy controls, mirrored by differences in neural signals, in particular reflected in the Pe component. Interestingly, despite clinically relevant cognitive deficits (e.g., aphasia and apraxia) including executive dysfunction (trail making test), we did not observe any behavioral impairments related to performance monitoring and error processing in the current LH stroke patients. Patients also showed similar results for Ne/ERN and Pe components, compared to the control group, and a highly similar prediction of errors from multivariate signals. ERP abnormalities during stimulus processing (i.e., N2 and P3) demonstrated the specificity of these findings in the current LH stroke patients. In contrast to previous studies, by employing a relatively large patient sample, a well-controlled experimental paradigm with a standardized error signaling procedure, and advanced data analysis, we were able to show that performance monitoring (of simple actions) is a preserved cognitive control function in LH stroke patients that might constitute a useful resource in rehabilitative therapies for re-learning impeded functions.