{"@context":{"@vocab":"https://cir.nii.ac.jp/schema/1.0/","rdfs":"http://www.w3.org/2000/01/rdf-schema#","dc":"http://purl.org/dc/elements/1.1/","dcterms":"http://purl.org/dc/terms/","foaf":"http://xmlns.com/foaf/0.1/","prism":"http://prismstandard.org/namespaces/basic/2.0/","cinii":"http://ci.nii.ac.jp/ns/1.0/","datacite":"https://schema.datacite.org/meta/kernel-4/","ndl":"http://ndl.go.jp/dcndl/terms/","jpcoar":"https://github.com/JPCOAR/schema/blob/master/2.0/"},"@id":"https://cir.nii.ac.jp/crid/1360302866833609088.json","@type":"Article","productIdentifier":[{"identifier":{"@type":"DOI","@value":"10.1016/j.neunet.2023.10.049"}},{"identifier":{"@type":"URI","@value":"https://api.elsevier.com/content/article/PII:S0893608023006147?httpAccept=text/xml"}},{"identifier":{"@type":"URI","@value":"https://api.elsevier.com/content/article/PII:S0893608023006147?httpAccept=text/plain"}},{"identifier":{"@type":"DOI","@value":"10.1101/2022.09.07.506913"}},{"identifier":{"@type":"PMID","@value":"38029719"}}],"resourceType":"学術雑誌論文(journal article)","dc:title":[{"@value":"Theoretical limits on the speed of learning inverse models explain the rate of adaptation in arm reaching tasks"}],"description":[{"notation":[{"@value":"<jats:title>Abstract</jats:title><jats:p>An essential aspect of human motor learning is the formation of inverse models, which map desired actions to motor commands. Inverse models can be learned by adjusting parameters in neural circuits to minimize errors in the performance of motor tasks through gradient descent. However, the theory of gradient descent establishes an upper limit on the learning speed, above which learning becomes unstable. Specifically, the eigenvalues of the Hessian of the error surface around a minimum determine the maximum speed of learning in a given task. Here, we use this theoretical framework to analyze the speed of learning in different inverse model learning architectures in a set of isometric arm-reaching tasks. We show theoretically that, for these isometric tasks, the error surface and, thus the speed of learning, are determined by the shapes of 1) the force manipulability ellipsoid of the arm and 2) the distribution of targets in the task. In particular, rounder manipulability ellipsoids generate a rounder error surface, allowing for faster learning of the inverse model. Rounder target distributions have a similar effect. We tested these predictions experimentally in a virtual quasi-isometric reaching task with a visuomotor transformation. The experimental results were consistent with our theoretical predictions. Furthermore, our analysis accounts for the speed of learning in previous experiments with incompatible and compatible virtual surgery tasks, and in visuomotor rotation tasks with different numbers of targets. By identifying aspects of a task that influence the speed of learning, our results provide theoretical principles for the design of motor tasks that allow for faster learning.</jats:p><jats:sec><jats:title>Author Summary</jats:title><jats:p>When facing a new or changing environment, humans need to learn a new internal inverse model to generate fast and accurate movements. Different computational architectures based on supervised learning via gradient descent have been proposed to explain the acquisition of these inverse models. Although the theory of gradient descent and its results regarding the speed of learning are well developed in Machine Learning, this framework has seldom been applied to computational models of human motor learning. In this study, we found that applying this theoretical framework to a set of isometric reaching tasks clearly reveals aspects of the motor task that can speed up or slow down learning. In particular, we found that, in isometric tasks, the force manipulability ellipsoid of the arm and the distribution of force targets determine the speed of learning. These theoretical results allowed us to generate testable hypotheses about the speed of learning in different motor task conditions, which we successfully confirmed experimentally. We believe that our methods and results could open new lines of research to systematically identify aspects of motor learning tasks that can be exploited to enhance the speed of learning and to design new tasks that are easy to learn.</jats:p></jats:sec>"}]}],"creator":[{"@id":"https://cir.nii.ac.jp/crid/1380302866833609097","@type":"Researcher","foaf:name":[{"@value":"Victor R. Barradas"}]},{"@id":"https://cir.nii.ac.jp/crid/1420282801182646272","@type":"Researcher","personIdentifier":[{"@type":"KAKEN_RESEARCHERS","@value":"10302978"},{"@type":"NRID","@value":"1000010302978"},{"@type":"NRID","@value":"9000312259308"},{"@type":"NRID","@value":"9000415225927"},{"@type":"NRID","@value":"9000318613181"},{"@type":"NRID","@value":"9000017682168"},{"@type":"NRID","@value":"9000347040258"},{"@type":"NRID","@value":"9000045444010"},{"@type":"NRID","@value":"9000241156066"},{"@type":"NRID","@value":"9000415199977"},{"@type":"NRID","@value":"9000004808440"},{"@type":"NRID","@value":"9000307254508"},{"@type":"NRID","@value":"9000014659646"},{"@type":"NRID","@value":"9000004833822"},{"@type":"NRID","@value":"9000255862248"},{"@type":"NRID","@value":"9000415412185"},{"@type":"NRID","@value":"9000318613791"},{"@type":"NRID","@value":"9000415178686"},{"@type":"NRID","@value":"9000241155920"},{"@type":"NRID","@value":"9000002100524"},{"@type":"NRID","@value":"9000341541760"},{"@type":"RESEARCHMAP","@value":"https://researchmap.jp/yaskoike"}],"foaf:name":[{"@value":"Yasuharu Koike"}]},{"@id":"https://cir.nii.ac.jp/crid/1380302866833609111","@type":"Researcher","foaf:name":[{"@value":"Nicolas Schweighofer"}]}],"publication":{"publicationIdentifier":[{"@type":"PISSN","@value":"08936080"}],"prism:publicationName":[{"@value":"Neural Networks"}],"dc:publisher":[{"@value":"Elsevier BV"}],"prism:publicationDate":"2024-02","prism:volume":"170","prism:startingPage":"376","prism:endingPage":"389"},"reviewed":"false","dcterms:accessRights":"http://purl.org/coar/access_right/c_abf2","dc:rights":["https://www.elsevier.com/tdm/userlicense/1.0/","https://www.elsevier.com/legal/tdmrep-license","http://creativecommons.org/licenses/by-nc/4.0/"],"url":[{"@id":"https://api.elsevier.com/content/article/PII:S0893608023006147?httpAccept=text/xml"},{"@id":"https://api.elsevier.com/content/article/PII:S0893608023006147?httpAccept=text/plain"}],"createdAt":"2023-10-31","modifiedAt":"2025-10-12","project":[{"@id":"https://cir.nii.ac.jp/crid/1040003825673333120","@type":"Project","projectIdentifier":[{"@type":"KAKEN","@value":"19H05728"},{"@type":"JGN","@value":"JP19H05728"},{"@type":"URI","@value":"https://kaken.nii.ac.jp/grant/KAKENHI-PLANNED-19H05728/"}],"notation":[{"@language":"ja","@value":"身体変容への超適応のモデル化"},{"@language":"en","@value":"Modeling of hyper-adaptability to altered musculoskeletal system"}]},{"@id":"https://cir.nii.ac.jp/crid/1040569382210216960","@type":"Project","projectIdentifier":[{"@type":"KAKEN","@value":"21K17789"},{"@type":"JGN","@value":"JP21K17789"},{"@type":"URI","@value":"https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-21K17789/"}],"notation":[{"@language":"en","@value":"Enhancing motor skill learning by manipulating extrinsic and intrinsic components of the motor task"}]},{"@id":"https://cir.nii.ac.jp/crid/1040581301858240384","@type":"Project","projectIdentifier":[{"@type":"KAKEN","@value":"23K28123"},{"@type":"JGN","@value":"JP23K28123"},{"@type":"URI","@value":"https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-23K28123/"}],"notation":[{"@language":"ja","@value":"筋シナジーを基にした力制御技能の定量化"}]}],"relatedProduct":[{"@id":"https://cir.nii.ac.jp/crid/1050575373777491072","@type":"Article","resourceType":"学術雑誌論文(journal article)","relationType":["references"],"jpcoar:relatedTitle":[{"@language":"en","@value":"Computational role of exploration noise in error-based de novo motor learning"}]},{"@id":"https://cir.nii.ac.jp/crid/1050579089766762112","@type":"Article","resourceType":"学術雑誌論文(journal article)","relationType":["references"],"jpcoar:relatedTitle":[{"@language":"en","@value":"Reinforcement learning establishes a minimal metacognitive process to monitor and control motor learning performance"}]},{"@id":"https://cir.nii.ac.jp/crid/1360011143567987584","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Motor Adaptation as a Greedy Optimization of Error and Effort"}]},{"@id":"https://cir.nii.ac.jp/crid/1360011144432686976","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Motor-output variability: A theory for the accuracy of rapid motor acts."}]},{"@id":"https://cir.nii.ac.jp/crid/1360011144476096768","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"<b>nparLD</b>: An<i>R</i>Software Package for the Nonparametric Analysis of Longitudinal Data in Factorial Experiments"}]},{"@id":"https://cir.nii.ac.jp/crid/1360011144571406336","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Unsupervised learning of granule cell sparse codes enhances cerebellar adaptive control"}]},{"@id":"https://cir.nii.ac.jp/crid/1360011145829574272","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Optimal feedback control as a theory of motor coordination"}]},{"@id":"https://cir.nii.ac.jp/crid/1360013172441345280","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Factors Governing the Form of the Relation Between Muscle Force and the EMG: A Simulation Study"}]},{"@id":"https://cir.nii.ac.jp/crid/1360021396345197824","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Internal models for interpreting neural population activity during sensorimotor control"}]},{"@id":"https://cir.nii.ac.jp/crid/1360021396524298368","@type":"Article","resourceType":"preprint","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Moving outside the lab: The viability of conducting sensorimotor learning studies online"}]},{"@id":"https://cir.nii.ac.jp/crid/1360283694350806528","@type":"Article","resourceType":"学術雑誌論文(journal article)","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Neural Substrates Related to Motor Memory with Multiple Timescales in Sensorimotor Adaptation"}]},{"@id":"https://cir.nii.ac.jp/crid/1360285709894510720","@type":"Article","resourceType":"学術雑誌論文(journal article)","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Modularity speeds up motor learning by overcoming mechanical bias in musculoskeletal geometry"}]},{"@id":"https://cir.nii.ac.jp/crid/1360292618806499456","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Role of the cerebellum in reaching movements in humans. I. Distributed inverse dynamics control"}]},{"@id":"https://cir.nii.ac.jp/crid/1360294645320598144","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"A deep learning framework for neuroscience"}]},{"@id":"https://cir.nii.ac.jp/crid/1360294648218873984","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Backpropagation and the brain"}]},{"@id":"https://cir.nii.ac.jp/crid/1360298762022701952","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Interactions between sensory prediction error and task error during implicit motor learning"}]},{"@id":"https://cir.nii.ac.jp/crid/1360302867629308672","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Both fast and slow learning processes contribute to savings following sensorimotor adaptation"}]},{"@id":"https://cir.nii.ac.jp/crid/1360574095700530432","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Cerebellar complex spikes encode both destinations and errors in arm movements"}]},{"@id":"https://cir.nii.ac.jp/crid/1360574095879875712","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"What are the computations of the cerebellum, the basal ganglia and the cerebral cortex?"}]},{"@id":"https://cir.nii.ac.jp/crid/1360574095944007808","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"A hierarchical neural-network model for control and learning of voluntary movement"}]},{"@id":"https://cir.nii.ac.jp/crid/1360580235887946240","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Corticospinal correlates of fast and slow adaptive processes in motor learning"}]},{"@id":"https://cir.nii.ac.jp/crid/1360580238039800704","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Explorative learning of inverse models: A theoretical perspective"}]},{"@id":"https://cir.nii.ac.jp/crid/1360585256671131008","@type":"Article","resourceType":"学術雑誌論文(journal article)","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"When 90% of the variance is not enough: residual EMG from muscle synergy extraction influences task performance"}]},{"@id":"https://cir.nii.ac.jp/crid/1360855570969424384","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Motor Memory and Local Minimization of Error and Effort, Not Global Optimization, Determine Motor Behavior"}]},{"@id":"https://cir.nii.ac.jp/crid/1360855571350827648","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"A more biologically plausible learning rule for neural networks."}]},{"@id":"https://cir.nii.ac.jp/crid/1360861709568509824","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"The dynamics of motor learning through the formation of internal models"}]},{"@id":"https://cir.nii.ac.jp/crid/1360861711803574528","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Reexposure to a sensorimotor perturbation produces opposite effects on explicit and implicit learning processes"}]},{"@id":"https://cir.nii.ac.jp/crid/1360869855128486272","@type":"Article","resourceType":"学術雑誌論文(journal article)","relationType":["isReferencedBy"],"jpcoar:relatedTitle":[{"@value":"Synergy quality assessment of muscle modules for determining learning performance using a realistic musculoskeletal model"}]},{"@id":"https://cir.nii.ac.jp/crid/1361137043969341952","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Cerebellar supervised learning revisited: biophysical modeling and degrees-of-freedom control"}]},{"@id":"https://cir.nii.ac.jp/crid/1361137044378755072","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Differences in Adaptation Rates after Virtual Surgeries Provide Direct Evidence for Modularity"}]},{"@id":"https://cir.nii.ac.jp/crid/1361137044971507968","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Learning Curves for Stochastic Gradient Descent in Linear Feedforward Networks"}]},{"@id":"https://cir.nii.ac.jp/crid/1361137045603418368","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Internal models for motor control and trajectory planning"}]},{"@id":"https://cir.nii.ac.jp/crid/1361418520453617536","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Inverse-dynamics model eye movement control by Purkinje cells in the cerebellum"}]},{"@id":"https://cir.nii.ac.jp/crid/1361418520593716992","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Questioning the role of sparse coding in the brain"}]},{"@id":"https://cir.nii.ac.jp/crid/1361699994104342528","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"A memory of errors in sensorimotor learning"}]},{"@id":"https://cir.nii.ac.jp/crid/1361699994332092416","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Learning of action through adaptive combination of motor primitives"}]},{"@id":"https://cir.nii.ac.jp/crid/1361981468701601664","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Dual Adaptation Supports a Parallel Architecture of Motor Memory"}]},{"@id":"https://cir.nii.ac.jp/crid/1361981470815715712","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Simple Statistical Gradient-Following Algorithms for Connectionist Reinforcement Learning"}]},{"@id":"https://cir.nii.ac.jp/crid/1361981471190853120","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Forward Models: Supervised Learning with a Distal Teacher"}]},{"@id":"https://cir.nii.ac.jp/crid/1362262943546171008","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Failure of Motor Learning for Large Initial Errors"}]},{"@id":"https://cir.nii.ac.jp/crid/1362544419284584448","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Toward an Integration of Deep Learning and Neuroscience"}]},{"@id":"https://cir.nii.ac.jp/crid/1362544419466257536","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Manipulability of Robotic Mechanisms"}]},{"@id":"https://cir.nii.ac.jp/crid/1362544419751921664","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Learning the parts of objects by non-negative matrix factorization"}]},{"@id":"https://cir.nii.ac.jp/crid/1362825896147436672","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Encoding of Sensory Prediction Errors in the Human Cerebellum"}]},{"@id":"https://cir.nii.ac.jp/crid/1363107368351710720","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Learning of Visuomotor Transformations for Vectorial Planning of Reaching Trajectories"}]},{"@id":"https://cir.nii.ac.jp/crid/1363107368645187712","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Interacting Adaptive Processes with Different Timescales Underlie Short-Term Motor Learning"}]},{"@id":"https://cir.nii.ac.jp/crid/1363107371266788608","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Purkinje cell activity during motor learning"}]},{"@id":"https://cir.nii.ac.jp/crid/1363388845018800128","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Neural Model of Adaptive Hand-Eye Coordination for Single Postures"}]},{"@id":"https://cir.nii.ac.jp/crid/1363388845228484352","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Adaptation to Visuomotor Transformations: Consolidation, Interference, and Forgetting"}]},{"@id":"https://cir.nii.ac.jp/crid/1363670319399171968","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Chaos may enhance information transmission in the inferior olive"}]},{"@id":"https://cir.nii.ac.jp/crid/1364233271195002240","@type":"Article","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Internal models in the cerebellum"}]},{"@id":"https://cir.nii.ac.jp/crid/1370302866833608962","@type":"Product","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Characteristics of implicit sensorimotor adaptation revealed by task-irrelevant clamped feedback"}]},{"@id":"https://cir.nii.ac.jp/crid/1370302866833609096","@type":"Product","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"A mathematical approach to the mechanical capabilities of limbs and fingers"}]},{"@id":"https://cir.nii.ac.jp/crid/1370302866833609102","@type":"Product","relationType":["references"],"jpcoar:relatedTitle":[{"@value":"Feedback-error-learning neural network for supervised motor learning"}]}],"dataSourceIdentifier":[{"@type":"CROSSREF","@value":"10.1016/j.neunet.2023.10.049"},{"@type":"KAKEN","@value":"PRODUCT-25157063"},{"@type":"KAKEN","@value":"PRODUCT-25306840"},{"@type":"KAKEN","@value":"PRODUCT-24795655"},{"@type":"OPENAIRE","@value":"doi_dedup___::409498fa45747280798e067d0ac01aa7"},{"@type":"CROSSREF","@value":"10.3389/fncom.2024.1355855_references_DOI_JOZx1yHojPY0pGPaeda8LWGcR7s"}]}