Tokyo, Japan ー Sony Computer Science Laboratories (hereinafter “Sony CSL,” President: Hiroaki Kitano) has discovered a training method to improve the complex, high-speed hand and finger movements of pianists, through its work under the Moonshot Research and Development Program (Moonshot Goal 1) and the Core Research for Evolutionary Science and Technology (CREST) Strategic Creative Research Promotion Program of the Japan Science and Technology Agency (JST). This result was achieved by a team led by Shinichi Furuya, Research Director of Tokyo Research at Sony CSL.

Musicians, athletes, surgeons, and other skilled people continually work to refine their already exceptional skills. However, after tremendous amounts of practice, skills sometimes level off, and the methods for breaking through such limitations have not been completely solved.
Sony CSL’s research team developed a training method using an exoskeleton robot that moves the five fingers independently and at high speed to experience complex, high-speed finger movements that humans cannot execute themselves. Training conducted after one hand was fitted with the robot revealed improvement in skills that had plateaued. Skill improvement also occurred in the other hand, which had not been trained.
This discovery suggests that, depending on practice quality, there is leeway to extend a skilled person’s skill limitations. These research findings are thus expected to be useful for creating new training programs and clarifying mechanisms of learning that are unique to skilled players and the function of the central nervous system behind such learning, as well as for developing practice methods that avoid excessive training and prevent disorders and injuries.

These results were published in the international scientific journal Science Robotics on January 15, 2025.

＜Background behind the research＞

Performing artists, athletes, surgeons, traditional craftspeople, and other experts in various fields acquire their skills through extensive training over many years. However, the "ceiling effect," in which trained skills eventually plateau, has long posed a major challenge to further refining these skills.
Until now, many skilled individuals have increased their amount of training in order to break through the ceiling effect. However, doing this increases the risk of a disorder or injury. In recent years, there has also been a series of reports stating that even boosting the amount of training is not effective enough to break through the ceiling effect (Figure 1).
One obstacle to improving training quality is the fact that the further advanced skills that skilled persons want to acquire cannot be experienced in advance because they are complex and high-speed, and it is therefore difficult to understand the movements through language. Because of this, it had been considered impossible to learn such exceptional skills with the body.

＜Content of the research＞

Sony CSL has discovered that a new training method using an exoskeleton robot that moves individual fingers independently and at high speeds stably for long periods of time can break through the ceiling effect on the motor skills of a pianist's fingers.
The skilled subjects that the research team trained were highly skilled pianists wearing an exoskeleton robot to experience high-speed and complex finger movements that would be impossible to perform on their own. As a result, the trained pianists broke through the limits of performance skills that had plateaued under previous practicing and were able execute complex skills even more quickly (Figure 2). Moreover, skill improvement was also observed in the opposite hand, which had not received any exoskeleton robot training. These results appeared when the pianists played after the exoskeleton robot was removed, suggesting that the pianists' own skills had improved.

The researchers conducted several experiments to clarify the mechanisms at work in breaking through the skill ceiling effect. First, they manipulated the complexity and speed of the exoskeleton robot's movements to determine which of the movements that it generated resulted in improving post-training performance skills. The results showed no improvement in skill after experiencing simple, quick hand movements or slow, complex movements. Next, tests of muscle strength and dexterity of each of the five fingers revealed no change in motor function of these fingers after training with the exoskeleton robot. Similarly, the sensory functions of the finger movements showed no differences before and after training. These results suggest that experience through robotic technology of complex, high-speed skills that are impossible for one to perform oneself will improve one's subsequent performance skill, and that accompanying this improvement are changes in the central nervous system, which controls complex movements (Figure 3).

To clarify the changes in brain function that are behind the benefits of training, a non-invasive transcranial magnetic stimulation test was implemented before and after training to evaluate the function of the cerebral cortex by magnetic stimulation. Finger movements generated by magnetically stimulating the motor area of the cerebral cortex, which is in charge of forming hand and finger movements, were measured with a sensor called a dataglove, and the movement patterns were analyzed using a data science technique known as tensor decomposition. The results showed that after the experience of complex finger movements with the exoskeleton robot, the magnetic stimulation test produced finger movements that more closely resembled the experienced movement patterns. This suggests that the function of the motor cortical area changed plastically to facilitate the forming of the complex movements that had been experienced.

The findings of these experiments on 90 pianists confirmed that training using an exoskeleton robot attached to the hand and fingers to experience complex, high-speed hand movements resulted in improving skill—even after the robot was removed—in both the trained hand and the opposite hand, and clarified that plastic changes in the motor cortical area are involved in this process.

＜Future Developments＞

This research suggests that skill plateauing, i.e., lack of improvement even after increasing the amount of training, is not equivalent to a ceiling effect on skill, but rather includes the possibility that there is room for further improvement through new experiences. Although robotics and virtual reality technology to date have mainly been used to augment a person’s abilities while fitted with a device (e.g., human augmentation), a new direction for future study has been indicated—the use of training with these advanced technologies to break through the conventional limits of our abilities. This is expected to lead to the development of unprecedented human abilities and the further deepening of our understanding of the mechanism of brain plasticity behind such abilities. The excitement of being able to improve and do new things with one's body is common not only to piano performance but also sports, surgery, and many other fields. Therefore, the results of this research can be expected to contribute to the realization of a society where people are freed from constraints on their bodies and brains, and to create the butterfly effects in many fields.

Compared to sports, learning to play music lags far behind in qualitative contributions that involve science and technology. As a result, many of the challenges that physical and mental limitations pose to embodying creativity remain unsolved. Providing training that uses the results of this research will contribute to the creation of a future society where artists embody their creativity free from physical and mental constraints by establishing new forms of music education based on evidence from the science of music performance (Musical Dynaformics). Moreover, since the skills for quickly achieving complex movements make up only a small portion of the skills for realizing diverse expressions in performance, there are expectations for the future creation of training that enables the embodiment and selection of a wide variety of expressions, as well as work on verifying the effectiveness of training that is tailored to artists' needs.

Title : Surmounting the ceiling effect of motor expertise by novel sensory experience with a hand exoskeleton
Journal : Science Robotics Vol 10, Issue 98
Authors : Furuya Shinichi, Oku Takanori, Nishioka Hayato, Hirano Masato
https://doi.org/10.1126/scirobotics.adn3802

This achievement was obtained through the following projects, research areas, and research topics.

Moonshot R&D Program
Moonshot Goal 1
Realization of a society in which human beings can be free from limitations of body, brain, space, and time by 2050.
Liberation from Biological Limitations via Physical, Cognitive and Perceptual Augmentation
Project manager:KANAI Ryota (Advanced Telecommunications Research Institute International)

CREST (Strategic Basic Research Programs)
[Symbiotic Interaction] Creation and Development of Core Technologies Interfacing Human and Information Environments
A study on skill acquisition mechanism and development of skill transfer systems
Research Director:Hideki Koike (Professor, School of Computing Institute of Sience Tokyo)