Forward modeling allows feedback control for fast reaching movements

“But yet, despite these delays, we are able to respond to properties for which signals have not yet been received (see e.g. Desmurget and Grafton 2000).”

“Additionally, recent studies have shown that “efference copy” and a forward model of the limb and its environment are significant actors in determining the sense of body position [14,15].”

“During the bimanual-load lifting task, the prediction of the consequences of the motor command might rely on a forward predictive model [17,21,22], that could then produce APAs.”

“Thereafter, an estimate of the new ‘state’ and associated reafferent signals is, respectively, generated (Wolpert and Ghahramani 2000; Wolpert et al. 1998, see also Desmurget and Grafton 2000; Elliott et al. 2010).”

“The minimum delay for sensory signals to influence an ongoing movement is 80-100 ms [1], and even reflex control involves at least 40-70 ms of loop delay [2], [3].”

“Primate species evolved distinct neural architecture for planning abstract action goals and converting them into actual motor actions (Hoshi & Tanji, 2007; Tanji & Hoshi, 2008) and for online motor control of the actual motor execution (Desmurget & Grafton, 2000).”

“La notion de modèle interne renvoie à la capacité de l’enfant à se représenter un ensemble de transformations d’états du corps propre qui pourront être utilisées de façon anticipée et rétroactive pour contrôler le mouvement (Desmurget & Grafton, 2000).”

“…the learning that we observed; however, in light of evidence that online error corrections can be driven by eVerence copy-based error signals (Bard et al. 1999; Desmurget and Grafton 2000), it is intriguing to consider that these same error signals might lead to modiWcations in movement planning.”

“However, many organisms, ranging from insect to human, also have the ability to predict the outcome of their own movements (Ebner & Pasalar, 2008; Webb, 2004), and this ability is fundamental for skilled motor behavior (Desmurget & Grafton, 2000).”

“…ballistic, manual-aiming movement relies on visual feedback to achieve accuracy, and that this feedback can be processed in as little as 100 ms (Desmurget & Grafton, 2000; Elliott & Allard, 1985; Goodale, Pelisson, & Prablanc, 1986; Saunders & Knill, 2003; Yoshida, Cauraugh, & Chow, 2004;…”

“…velocity of object and effector can be estimated with 378 negligible delays or even predicted in advance, thus allowing feedback for fast movements (e.g., Desmurget and Grafton 2000; Miall and Wolpert 1996). general), the main focus has been on the optical information used for timing the movement.”

“Whilst resolving redundancy in the CNS remains a key research focus [34], it is widely accepted that the CNS employs both feedforward and feedback internal models for motor control [35, 36] with signals descending from motor cortical areas to the spinal neural structures [21].”

“This is a well-known problem in sensorimotor control, and it has been proposed that the prediction of future states may solve it [32-37].”

“The two controls require different neurophysiological processes: the proactive being more driven by a feedforward control while the reactive more by a feedback control [20].”

“to progressively nullify the distance between the ongoing state of the motor plant, indirectly estimated by integrating sensory inflows and motor outflows (Desmurget and Grafton, 2000; Wolpert et al., 1995) and the goal of the movement (Diedrichsen et al.”

“, a deviation from a designated movement trajectory) are mostly connected to activity in the posterior parietal cortex (Desmurget et al., 1999; Desmurget and Grafton, 2000; Krigolson and Holroyd, 2006, 2007b; Krigolson et al., 2008).”

“…either from a delayed comparison between the predicted sensory feedback (i.e., the output of the forward model) and the actual feedback (Diedrichsen et al., 2005), or from an instantaneous comparison of the goal representation, and the expected sensory feedback (Desmurget and Grafton, 2000).”

“Consequently, the differences typically observed between reaches performed without, as compared to with visual feedback, are taken to reflect the activity of feedforward control (Desmurget & Grafton, 2000).”

“Considerable evidence points to internal models in the PPC (and cerebellum) that integrate multisensory inflow and efference copy (replicas of motor commands) to enable fast online corrections (Desmurget & Grafton, 2000; Buneo & Andersen, 2006; Mulliken et al., 2008).”

“MIP neurons are able to detect mismatches between reaching movements and target location (Desmurget & Grafton, 2000; Kalaska, Cisek, & Gosselin-Kessiby, 2003).”

“To explain the stable movements of a human, recent studies [3]–[5] have instead emphasized the importance of feedforward control and”

“The forward model is updated/finetuned during movement execution by incoming proprioceptive and visual inputs (for a review, see Desmurget & Grafton, 2000; Shadmehr, Smith, & Krakauer, 2010).”

“Finally, their role could be related with the optimization of the internal feedback-loops to continuously adapt motion at different frequencies [Desmurget and Grafton, 2000; Miall et al., 2001].”

“The brain uses this sensory prediction error in real-time to adjust motor commands, in order to reach the movement goal (Desmurget and Grafton, 2000).”

“However, there is evidence that humans are capable of making a prediction by integrating motor efference copies using an internal model in order to compensate this delay (Wolpert et al. 1995; Kawato 1999; Desmurget and Grafton 2000; Mehta and Schaal 2002; Milton et al. 2016; Insperger and Milton 2021).”

“635) and Condition × Session × ID (F(12,276) = 2.”

“…et al., 1995; Todorov & Jordan, 2002), Saunders and Knill’s findings were consistent with the recently developed integrative hybrid movement model (Desmurget & Grafton, 2000; Schmidt & Wrisberg, 2008), maintaining that sensory feedback and motor program preplanning are both used in motor control.”

“This implies that the CNS ‘‘knows” in advance the mechanical effects of a motor command (Desmurget and Grafton, 2000).”

“…have also analyzed how movements are controlled and how precision is achieved (feed-forward models suggested by Meyer et al. (1988) and Milner, 1992; feedback models proposed by Flanagan, Feldman, and Ostry (1993) and Todorov and Jordan (2002); hybrid models by Desmurget and Grafton (2000)).”

“A second comparison is posited between the prediction outputted by the forward model and the goal state (Blakemore et al. 2002; Desmurget and Grafton 2000).”

“Current theories in motor control attribute an important role to predictive (feedforward) and reactive (feedback) processes [2-3].”

“This choice is consistent with the usage of goal states considered in motor planning in [5] and in motor control in [13], who define the goal as a particular desired state of a given external stimulus.”

“Strikingly, the kinematic modifications applied to the ongoing response are not consciously detected by the subjects, who remain completely unaware of the profound changes that affect path curvature and individual joint trajectories ( Desmurget and Grafton 2000; Desmurget et al. 2001, 2004).”

“…response to probe tones during speech movement planning (i.e., PSAM; [11–13, The notion that the central nervous system generates sensory predictions by means of an internal forward model is a key concept in contemporary models of human sensorimotor control across effector systems (e.g., [34–40]).”

“…successful execution of voluntary movements is supported by the interactions between the somatosensory and motor systems (Prochazka et al., 1997; Desmurget and Grafton, 2000; Schieber, 2001; Davis et al., 2022), the neural mechanisms by which tactile inputs in fl uence motor circuits in humans…”

“…initial ballistic phase and a subsequent controlling phase in which fine adjustments are made to successfully achieve the task goal (Desmurget and Grafton, 2000), it is reasonable to think that tasks with higher precision requirements demand more sensory feedback control during the movement…”

“…MT constraint, participants would have had sufficient time to process and use visual feedback online to reduce their reaching errors when reaching with both aligned and rotated cursor feedback (Desmurget & Grafton, 2000; Shadmehr, Smith, & Krakauer, 2010; Zelaznik, Hawkins, & Kisselburgh, 1983).”

“…1745 capacity to read the feedforward commands (i.e., based on the open loop circuitery) and consequently over-rely on auditory feedback (for a review on sensory feedback and forward modeling, see Desmurget and Grafton, 2000; Perkell, 2012; Guenther and Hickok, 2016; Parrell and Houde, 2019).”

“Evidence of sensorimotor loops within the posterior parietal cortex (PPC), at the area of the intraparietal sulcus (IPS), and the cerebellum suggest the use of both feedforward and feedback control mechanisms associated within these structures [41].”

“Also, hand perception in the early phase of action impacts on a more accurate motor program [6].”

“However, successful motor control relies on communication between motor and nonmotor areas, such as integrating sensory information to update the movement trajectory or to attenuate movement or to respond to unexpected perturbations [3-6].”

“The loop of two sensorimotor transformations allows the internal feedback control to achieve a reaching movement (Desmurget & Grafton, 2000).”

“, conduction delays from axonal connection lengths) [32, 33].”

“The learning and adaptation capabilities of the human brain have fascinated control engineers over the past several decades [4], [5].”

“…and Pasalar, 2008; Miall and King, 2008; Imamizu and Kawato, 2012), as well as in the parietal lobe (Wolpert et al., 1998a; Sirigu et al., 1999; Desmurget and Grafton, 2000; Andersen and Buneo, 2002; Desmurget and Sirigu, 2009) (for a review tackling these two views, see Blakemore and Sirigu,…”

“To take into account accuracy constraints, we propose to consider goal-directed movements such as arm pointing as the superposition of an open-loop motion (the planned trajectory) and of a feedback process whose role is to provide on-line corrections and in particular to stabilize the hand around the target [12, 51].”

“Thus it is argued that there must exist strategies that anticipate, or compensate, for the effects of the time delay [5], such as feedforward controllers [6], ”inverse dynamics” [7], and so on.”

“…ballistic, manual-aiming movement relies on visual feedback to achieve accuracy, and that this feedback can be processed in as little as 100 ms (Desmurget & Grafton, 2000; Elliott & Allard, 1985; Goodale, Pelisson, & Prablanc, 1986; Saunders & Knill, 2003; Yoshida, Cauraugh, & Chow, 2004;…”

“…be adjusted, it was chosen based on typical neural processing times, such as reaction times to sensory stimuli (150–250 ms) [33] cortical processing in sensory pathways (100–200 ms) [34], and feedback loops in motor control (around 200 ms) [35], which are common in neural integration and response.”

“However, these interactions are so closely linked in time that are often taken as evidence for predictive models (Desmurget and Grafton, 2000; Wolpert and Flanagan, 2001; Flanagan et al., 2003; Franklin and Wolpert, 2011).”

“However, studies link forward models to the cerebellum, parietal cortex, and primary motor cortex, obscuring where PD pathology may exist relative to forward models [5], [14], [28].”

“These models are then used to update the initial motor plan, defined before the onset of the movement [10]; in turn, motor plans of point-to-point reaching tasks in the three-dimensional space could be approximated as straight movements with sigmoidal trajectories (i.e., with bell-shaped velocity…”