RESUMEN

Introduction: Robotic assembly tasks require precise manipulation and coordination, often necessitating advanced learning techniques to achieve efficient and effective performance. While residual reinforcement learning with a base policy has shown promise in this domain, existing base policy approaches often rely on hand-designed full-state features and policies or extensive demonstrations, limiting their applicability in semi-structured environments. Methods: In this study, we propose an innovative Object-Embodiment-Centric Imitation and Residual Reinforcement Learning (OEC-IRRL) approach that leverages an object-embodiment-centric (OEC) task representation to integrate vision models with imitation and residual learning. By utilizing a single demonstration and minimizing interactions with the environment, our method aims to enhance learning efficiency and effectiveness. The proposed method involves three key steps: creating an object-embodiment-centric task representation, employing imitation learning for a base policy using via-point movement primitives for generalization to different settings, and utilizing residual RL for uncertainty-aware policy refinement during the assembly phase. Results: Through a series of comprehensive experiments, we investigate the impact of the OEC task representation on base and residual policy learning and demonstrate the effectiveness of the method in semi-structured environments. Our results indicate that the approach, requiring only a single demonstration and less than 1.2 h of interaction, improves success rates by 46% and reduces assembly time by 25%. Discussion: This research presents a promising avenue for robotic assembly tasks, providing a viable solution without the need for specialized expertise or custom fixtures.

RESUMEN

Practice not only improves task performance, but also changes task execution from rule- to memory-based processing by incorporating experiences from practice. However, how and when this change occurs is unclear. We tested the hypothesis that strategy transition in task learning results from cost-benefit analysis. Participants learned two task sequences and were then queried about the task type at a cued sequence and position. Behavioral improvement with practice can be accounted for by a computational model implementing cost-benefit analysis. Model-predicted strategy transition points are related to behavioral slowing and changes in fMRI activation patterns in the dorsolateral prefrontal cortex. Strategy transition is also related to increased pattern separation in the ventromedial prefrontal cortex. The cost-benefit analysis model outperforms alternative models (e.g., both strategies racing for being expressed in behavior) in accounting for empirical data. These findings support cost-benefit analysis as a mechanism of practice-induced strategy shift.

RESUMEN

Thalamocortical interaction is a ubiquitous functional motif in the mammalian brain. Previously (Hwang et al., 2021), we reported that lesions to network hubs in the human thalamus are associated with multi-domain behavioral impairments in language, memory, and executive functions. Here, we show how task-evoked thalamic activity is organized to support these broad cognitive abilities. We analyzed functional magnetic resonance imaging (MRI) data from human subjects that performed 127 tasks encompassing a broad range of cognitive representations. We first investigated the spatial organization of task-evoked activity and found a basis set of activity patterns evoked to support processing needs of each task. Specifically, the anterior, medial, and posterior-medial thalamus exhibit hub-like activity profiles that are suggestive of broad functional participation. These thalamic task hubs overlapped with network hubs interlinking cortical systems. To further determine the cognitive relevance of thalamic activity and thalamocortical functional connectivity, we built a data-driven thalamocortical model to test whether thalamic activity can be used to predict cortical task activity. The thalamocortical model predicted task-specific cortical activity patterns, and outperformed comparison models built on cortical, hippocampal, and striatal regions. Simulated lesions to low-dimensional, multi-task thalamic hub regions impaired task activity prediction. This simulation result was further supported by profiles of neuropsychological impairments in human patients with focal thalamic lesions. In summary, our results suggest a general organizational principle of how the human thalamocortical system supports cognitive task activity.

Asunto(s)

RESUMEN

Routine action sequences can share a great deal of similarity in terms of their stimulus response mappings. As a consequence, their correct execution relies crucially on the ability to preserve contextual and temporal information. However, there are few empirical studies on the neural mechanism and the brain areas maintaining such information. To address this gap in the literature, we recently recorded the blood-oxygen level dependent (BOLD) response in a newly developed coffee-tea making task. The task involves the execution of four action sequences that each comprise six consecutive decision states, which allows for examining the maintenance of contextual and temporal information. Here, we report a reanalysis of this dataset using a data-driven approach, namely multivariate pattern analysis, that examines context-dependent neural activity across several predefined regions of interest. Results highlight involvement of the inferior-temporal gyrus and lateral prefrontal cortex in maintaining temporal and contextual information for the execution of hierarchically organized action sequences. Furthermore, temporal information seems to be more strongly encoded in areas over the left hemisphere.

Asunto(s)

RESUMEN

Selection and integration of information based on current goals is fundamental for goal-directed behavior. Reward motivation has been shown to improve behavioral performance, yet the neural mechanisms that link motivation and control processes, and in particular its effect on context-dependent information processing, remain unclear. We used functional magnetic resonance imaging (fMRI) in 24 human volunteers (13 females) to test whether reward motivation enhances the coding of task-relevant information across the frontoparietal cortex, as would be predicted based on previous experimental evidence and theoretical accounts. In a cued target detection task, participants detected whether an object from a cued visual category was present in a subsequent display. The combination of the cue and the object visual category determined the behavioral status of the objects. To manipulate reward motivation, half of all trials offered the possibility of a monetary reward. We observed an increase with reward in overall univariate activity across the frontoparietal control network when the cue and subsequent object were presented. Multivariate pattern analysis (MVPA) showed that behavioral status information for the objects was conveyed across the network. However, in contrast to our prediction, reward did not increase the discrimination between behavioral status conditions in the stimulus epoch of a trial when object information was processed depending on a current context. In the high-level general-object visual region, the lateral occipital complex, the representation of behavioral status was driven by visual differences and was not modulated by reward. Our study provides useful evidence for the limited effects of reward motivation on task-related neural representations and highlights the necessity to unravel the diverse forms and extent of these effects.

Asunto(s)

RESUMEN

Functional connectivity (FC) studies have identified at least two large-scale neural systems that constitute cognitive control networks, the frontoparietal network (FPN) and cingulo-opercular network (CON). Control networks are thought to support goal-directed cognition and behavior. It was previously shown that the FPN flexibly shifts its global connectivity pattern according to task goal, consistent with a "flexible hub" mechanism for cognitive control. Our aim was to build on this finding to develop a functional cartography (a multimetric profile) of control networks in terms of dynamic network properties. We quantified network properties in (male and female) humans using a high-control-demand cognitive paradigm involving switching among 64 task sets. We hypothesized that cognitive control is enacted by the FPN and CON via distinct but complementary roles reflected in network dynamics. Consistent with a flexible "coordinator" mechanism, FPN connections were varied across tasks, while maintaining within-network connectivity to aid cross-region coordination. Consistent with a flexible "switcher" mechanism, CON regions switched to other networks in a task-dependent manner, driven primarily by reduced within-network connections to other CON regions. This pattern of results suggests FPN acts as a dynamic, global coordinator of goal-relevant information, while CON transiently disbands to lend processing resources to other goal-relevant networks. This cartography of network dynamics reveals a dissociation between two prominent cognitive control networks, suggesting complementary mechanisms underlying goal-directed cognition.SIGNIFICANCE STATEMENT Cognitive control supports a variety of behaviors requiring flexible cognition, such as rapidly switching between tasks. Furthermore, cognitive control is negatively impacted in a variety of mental illnesses. We used tools from network science to characterize the implementation of cognitive control by large-scale brain systems. This revealed that two systems, the frontoparietal (FPN) and cingulo-opercular (CON) networks, have distinct but complementary roles in controlling global network reconfigurations. The FPN exhibited properties of a flexible coordinator (orchestrating task changes), while CON acted as a flexible switcher (switching specific regions to other systems to lend processing resources). These findings reveal an underlying distinction in cognitive processes that may be applicable to clinical, educational, and machine learning work targeting cognitive flexibility.

Asunto(s)

RESUMEN

Researchers have recognized the role that task representation plays in our behavior for many years. However, the specific influence that the structure of one's task representation has on executive functioning has only recently been investigated. Prior research suggests that adjustments of cognitive control are affected by subtle manipulations of aspects of the stimulus-response pairs within and across task sets. This work has focused on examples of cognitive control such as response preparation, dual-task performance, and the congruency sequence effect. The current study investigates the effect of task representation on another example of control, post-error slowing. To determine if factors that influence how people represent a task affect how behavior is adjusted after an error, an adaptive attention-shifting task was developed with multiple task delimiting features. Participants were randomly assigned to a separate task set (two task sets) or an integrated task set (one task set) group. For the separate set group, the task sets switched after each trial. Results showed that only the integrated set group exhibited post-error slowing. This suggests that task representation influences the boundaries of cognitive control adjustments and has implications for our understanding of how control is organized when adjusting to errors in performance.

Asunto(s)

RESUMEN

By following explicit instructions, humans instantaneously get the hang of tasks they have never performed before. We used a specially calibrated multivariate analysis technique to uncover the elusive representational states during the first few implementations of arbitrary rules such as 'for coffee, press red button' following their first-time instruction. Distributed activity patterns within the ventrolateral prefrontal cortex (VLPFC) indicated the presence of neural representations specific of individual stimulus-response (S-R) rule identities, preferentially for conditions requiring the memorization of instructed S-R rules for correct performance. Identity-specific representations were detectable starting from the first implementation trial and continued to be present across early implementation trials. The increasingly fluent application of novel rule representations was channelled through increasing cooperation between VLPFC and anterior striatum. These findings inform representational theories on how the prefrontal cortex supports behavioral flexibility specifically by enabling the ad-hoc coding of newly instructed individual rule identities during their first-time implementation.

Asunto(s)

RESUMEN

The valence of stimuli can influence performance in the spatial stimulus-response compatibility task, but this observation could arise from the process of selecting responses or selecting stimulus-response mappings. The response-selection account proposes that spatial compatible and incompatible keypress responses serve as approaching and avoiding actions to a valenced target. The mapping-selection account suggests that there is congruence between stimulus valence and stimulus-response mappings; positive-compatible/negative-incompatible is more congruent than negative-compatible/positive-incompatible. Whereas affective valence was part of the target stimuli to which participants responded in previous studies, the present study isolated affective valence from the target by presenting an additional mapping cue separately from the target, so that spatially compatible and incompatible keypress responses could no longer serve as approaching and avoiding actions to valenced target stimuli. The present results revealed that responses were still faster when positive and negative mapping cues were assigned to the spatially compatible and incompatible mappings than when the assignment was reversed. The finding supports the mapping-selection account, indicating that positive and negative cues influence performance without approach-avoidance actions to valenced stimuli. The experiment provides important implications as to how tasks are represented and are dependent on affective processing.

Asunto(s)

RESUMEN

Sequential modulation between two task congruencies has been examined to investigate the nature of the cognitive control mechanism underlying the congruency sequence effect (CSE). Previous results regarding what consecutive tasks must have in common to engender the cross-task CSE are inconsistent. The present study examined the roles of stimulus-response (S-R) mappings and response mode as critical factors in determining the scope of control. Two flanker-compatibility tasks having different stimulus and response sets alternated in turn, and the arbitrariness of S-R mappings alone (Experiment 1) or the arbitrariness of stimulus set and the distinctiveness of response modes (Experiment 2) were manipulated. Experiment 1 showed that non-arbitrary S-R mappings engendered a cross-task CSE even when the response modes were different. However, when S-R mappings were arbitrary in Experiment 2, sequential modulation was evident across two tasks only when their response modes were same, irrespective of the arbitrariness of the stimulus set. These results suggest that the arbitrariness of S-R mappings and response mode are salient task features that reconfigure task representation and consequently determine the scope of the control underlying the CSE.

Asunto(s)

RESUMEN

This paper explores issues of task representation in choice reaction time tasks. How is it possible, and what does it take, to represent such a task in a way that enables a performer to do the task in line with the prescriptions entailed in the instructions? First, a framework for task representation is outlined which combines the implementation of task sets and their use for performance with different kinds of representational operations (pertaining to feature compounds for event codes and code assemblies for task sets, respectively). Then, in a second step, the framework is itself embedded in the bigger picture of the classical debate on the roles of contingency and similarity for the formation of associations. The final conclusion is that both principles are needed and that the operation of similarity at the level of task sets requires and presupposes the operation of contingency at the level of event codes.

Asunto(s)

RESUMEN

We present a framework for representing scenarios with complex object interactions, where a robot cannot directly interact with the object it wishes to control and must instead influence it via intermediate objects. For instance, a robot learning to drive a car can only change the car's pose indirectly via the steering wheel, and must represent and reason about the relationship between its own grippers and the steering wheel, and the relationship between the steering wheel and the car. We formalize these interactions as chains and graphs of Markov decision processes (MDPs) and show how such models can be learned from data. We also consider how they can be controlled given known or learned dynamics. We show that our complex model can be collapsed into a single MDP and solved to find an optimal policy for the combined system. Since the resulting MDP may be very large, we also introduce a planning algorithm that efficiently produces a potentially suboptimal policy. We apply these models to two systems in which a robot uses learning from demonstration to achieve indirect control: playing a computer game using a joystick, and using a hot water dispenser to heat a cup of water.

RESUMEN

Co-actors can facilitate the achievement of a shared goal by accurately anticipating each other's needs and subsequently planning actions to accommodate those needs. The purpose of the present study was to determine if co-actors plan and execute their movements to accommodate the difficulty of their partners' action. We hypothesized that information derived from shared task representations could influence the simulation of other's actions and that motor experience would enhance the ability of co-actor's to anticipate their co-actor's needs. Partners performed a sequential aiming task. The initiator of the sequential action placed a dowel on a line between two potential targets that varied in size across trials. The initiator did not know the actual target location prior to placing the dowel. The finisher then grasped the dowel and moved it to whichever target was signaled, from wherever their partner had placed the dowel. Participants completed the partner task before and after completing an individual task in which they completed both the initiating and the finishing movements. Consistent with the prediction that co-actors represent the difficulty of their partners' actions, the dowel was placed closer to the smaller target of a pair. Further, it was found that motor experience influenced dowel placement - there was a shift in dowel placement following the completion of the individual task. These results indicate that co-actors plan their movements based on features of their co-actor's movements and that motor experience provides information that allows people to better plan movements for their partners.

Asunto(s)

RESUMEN

The ability to coordinate the simultaneous movements of our arms is limited by a coalition of constraints. Some of these constraints can be overcome when the task conceptualisation is improved. The present study investigated how the movement preparation of bimanual reaching movements was affected by integrated visual feedback of the responses. Previous research has shown that the preparation of bimanual asymmetric movements takes longer than bimanual symmetric movements. The goal of the present study was to determine whether integrated, Lissajous feedback could eliminate this bimanual asymmetric cost. Fifteen participants made unimanual and bimanual symmetric and asymmetric reaches with separate feedback, where there was a cursor and a target for each hand. Participants also made bimanual symmetric and asymmetric movements with integrated feedback; a single cursor and a single target represented the locations and goals of both arms in this condition. The results showed a bimanual asymmetric cost with separate feedback, and that this cost persisted with integrated feedback. We suggest that integrated feedback improved continuous and discrete bimanual movements in other experiments by facilitating error detection and correction processes. We hypothesise that the bimanual asymmetric cost persisted in the present experiment because the uncertainty associated with choice reaction time prevented the facilitated error processing from improving the preparation of the next trial.

Asunto(s)

RESUMEN

The present study attempts to give an account of how students represent writing task in an EAP course. Further, the study is intended to discover if learners' mental representation of writing would contribute to their written performance. During a 16-week term, students were instructed to practice writing as a problem solving activity. At almost the end of the term, they were prompted to write on what they thought writing task was like and also an essay on an argumentative topic. The results revealed that students could conceptualize the instructed recursive model of writing as a process-based, multi-dimensional and integrated activity inducing self-direction and organization while holding in low regard the product view of writing. The findings also demonstrated that task representation was related to the students' writing performance, with process oriented students significantly outperforming the product-oriented ones. Also, it was found that task representation components (ideational, linguistic, textual, interpersonal) had a significant relationship with the written performance ([Formula: see text]; Sig.: 0.006). The study can have both theoretical and practical implications with regard to the factors involving the students' writing internal processes and their effects on written performance.

Asunto(s)

RESUMEN

When engaging in joint activities, humans tend to sacrifice some of their own sensorimotor comfort and efficiency to facilitate a partner's performance. In the two experiments reported here, we investigated whether ownership-a socioculturally based nonphysical feature ascribed to objects-influenced facilitatory motor behavior in joint action. Participants passed mugs that differed in ownership status across a table to a partner. We found that participants oriented handles less toward their partners when passing their own mugs than when passing mugs owned by their partners (Experiment 1) and mugs owned by the experimenter (Experiment 2). These findings indicate that individuals plan and execute actions that assist their partners but do so to a smaller degree if it is the individuals' own property that the partners intend to manipulate. We discuss these findings in terms of underlying variables associated with ownership and conclude that a self-other distinction can be found in the human sensorimotor system.

Asunto(s)

RESUMEN

The congruency effect observed in distracter interference tasks is usually smaller after incongruent relative to congruent trials. However, the nature of control processes underlying this congruency sequence effect (CSE) remains a topic of active debate. For example, while some researchers have suggested that these processes are recruited only when participants utilize the same response mode (e.g., the same hand) to respond in consecutive trials, others have argued that these processes can operate independently of response mode. To distinguish between these views, we investigated whether changes of response mode across consecutive trials influence the CSE in a prime-probe task (Experiment 1) or a flanker task (Experiment 2). Such changes did not influence the CSE in either task. Further, the CSE was significant even when participants utilized different response modes (i.e., different hands) to respond in consecutive trials. These findings indicate that control processes underlying the CSE can operate independently of response mode and thereby clarify the nature of control processes that minimize distraction from irrelevant stimuli.

Asunto(s)

RESUMEN

The execution of a task necessitates the use of a specific response modality. We examined the role of different response modalities by using a task-switching paradigm. In Experiment 1, subjects switched between two numerical judgments, whereas response modality (vocal vs. manual vs. foot responses) was manipulated between groups. We found judgment-shift costs in each group, that is irrespective of the response modality. In Experiment 2, subjects switched between response modalities (vocal vs. manual, vocal vs. foot, or manual vs. foot). We observed response-modality shift costs that were comparable in all groups. In sum, the experiments suggest that the response modality (combination) does not affect switching per se. Yet, modality-shift costs occur when subjects switch between response modalities. Thus, we suppose that modality-shift costs are not due to a purely motor-related mechanisms but rather emerge from a general switching process. Consequently, the response modality has to be considered as a cognitive component in models of task switching.

RESUMEN

The social Simon effect (SSE) occurs if two participants share a Simon task by making a Go/No-Go response to one of two stimulus features. If the two participants perform this version of the Simon task together, a Simon effect occurs (i.e., performance is better with spatial stimulus-response correspondence), but no effect is observed if participants perform the task separately. The SSE has been attributed to the automatic co-representation of the co-actor's actions, which suggests that it relies on online information about the other's actions. To test this implication, we investigated whether the SSE varies with the presence and amount of online action-related feedback from the other person. Experiment 1 replicated the SSE with auditory stimuli. Experiment 2, in which participants were blindfolded, demonstrated that visual feedback from the other's actions is not necessary for the SSE to occur. Experiment 3 replicated Experiment 2 with a regular and a soundless keyboard. A comparable SSE was obtained in both conditions, suggesting that even auditory online input from the other's actions is not necessary. Taken together, our data suggest that the SSE does not rely on online information about the co-actor's actions but that a priori offline information about another actor's presence is sufficient to generate the effect.