Search

When grasping and manipulating a hand-held object, the grip force (GF) a person applies has been typically reported as finely tuned to the object’s load force (LF) as LF varies due to object accelerations, with GF adjustments exhibiting continuously synchronous and proportionate coordination with LF variations. Considering the existence of inherent sensorimotor feedback delays, synchronous GF-LF coordination thus indicates that GF anticipates future changes in LF. Prevailing theoretical accounts of GF control account for this anticipation via internal models that generate explicit predictions of future sensorimotor states on the basis of efference copies of outbound motor commands to enable feedforward (i.e., preprogrammed) anticipatory GF control.There are several theoretical issues with the internal model approach, however. In particular, it fails to properly acknowledge various computational limitations faced by the CNS. With regard to GF control, the internal model approach has not addressed the act of manipulating mechanically complex objects, which are objects that exhibit complex internal dynamics due to nonlinear, underactuated degrees of freedom. The pervasive ability for humans to skillfully manipulate complex objects raises a critical issue for the internal model approach as the underactuated, nonlinear dynamics of a complex object cannot be encoded by an efference copy, which makes internally modeling those dynamics unfeasible. An internal model account of GF control would therefore expect that GF would be unable to anticipate the changing LF exerted by a complex object as it is actively manipulated. However, other lines of research, such as the anticipatory synchronization framework, offer evidence that anticipatory control of complex dynamics can arise without needing to invoke internal model constructs.The current study aimed to critically test the internal model approach’s pivotal claim—that efference copies of motor commands are essential to anticipation—by evaluating the degree to which GF could anticipate the nonlinear, underactuated LFs that resulted from manipulations of a complex object. These LFs, in principle, cannot be estimated on the basis of an efference copy. GF control with the complex object was tested in comparison to the degree to which GF anticipated the (traditionally studied) linear, fully actuated LFs that resulted from equivalent manipulations of a simple object.Participants grasped a simple or complex object between the thumb and forefinger and moved the object to manually track a moving target presented in virtual reality. The simple object was a disc-shaped, weighted, rigid object while the complex object employed a weighted pendulum mechanism to elicit underactuated, nonlinear LFs resulting from object manipulations.Comparisons of observed (i.e., real) complex object LF profiles to LF profiles that were simulated to assume the mechanics of the simple object confirmed that complex object LF deviated substantially from what would have been experienced with equivalent manipulations of the simple object (and from what could have been predicted about LF on the basis of efference copies of the underlying motor commands). Despite this, GF was observed to coordinate very strongly with and robustly anticipate the nonlinear, underactuated variations in complex object LF alongside anticipating the linear, fully actuated variations in simple object LF, which strongly suggest that anticipatory GF control can be achieved without reliance on efference copy-based predictions of upcoming LF changes. These findings are substantially incompatible with the internal model approach to GF control and suggest that a dynamical systems-based approach, such as the anticipatory synchronization framework, may be more suitable to account for anticipatory phenomena during the ongoing control of action.

Tightly coordinated grip force adaptations in response to changing load forces have been reported as continuous, stable, and proportional to the load force changes. Considering the existence of inherent sensorimotor feedback delays, current accounts of grip force-load force coupling invoke explicit predictive mechanisms in the form of internal models for feedforward control to account for anticipatory grip force modulations. However, recent findings show that grip force-load force coupling is less stable and regular than previously thought. When load force variations are more diminished, grip force is adjusted only intermittently and with less responsiveness, rather than adjusted continuously and proportionally. This has several implications for current accounts of grip force-load force coupling, and more broadly for accounts of anticipation in motor control. The objective of the current study was to quantify the time-varying characteristics of grip force in response to differentially changing load force, and to investigate how the intermittent responsiveness of grip force is tuned to changes in load force magnitude and predictability. A second aim of the study was to investigate grip force-load force coupling in a task where the participant’s focus was not directed toward explicitly controlling grip force; given that the ability to maintain grasp of hand-held objects persists in daily life without constant explicit control, it is noteworthy that no existing studies have examined GF-LF coupling in an implicit context.Twenty-three undergraduate students from the University of Cincinnati grasped a cylindrical object with their right thumb and forefinger and moved the object in order to track a moving target presented in virtual reality. In moving the object to track the target, participants produced changes in the load force exerted by the object on their grasp, for which their grip force had to be adjusted to prevent the object from slipping. The target moved in either a highly regular (sinusoidal), stochastic (fractional Brownian motion (fBm), i.e., random walk), or deterministic but chaotic fashion (i.e., the chaotic Lorenz system), which resulted in load force changes with varying degrees of predictability. Different amounts of mass (60 g and 120 g) were added to the object as a second manipulation to influence the magnitude of load force variation.The results confirmed the finding of Grover, Lamb, Bonnette, Silva, Lorenz, & Riley (submitted) that grip force-load force coupling was not continuous and stable over the course of a trial, and that the strength and stability of the coupling increased as load force changes became less predictable and their magnitudes became more enhanced. In addition, the intermittency observed in grip force-load force coupling was found to extend to the temporal relation between grip force to load force, which varied from anticipatory, to lagged, to complete coordination with load force, over the course of individual trials. These insights raise several important issues for the internal model account of grip force-load force coupling and suggest instead that a complex dynamical system approach, such as an anticipatory synchronization or delay differential equation model, may be necessary to further develop a comprehensive account of the anticipatory nature of grip force-load force coupling and other anticipatory phenomena in sensorimotor control.