Investigating sensory-motor interactions to shape rehabilitation
Author(s)
Melendez-Calderon, Alejandro
Type
Thesis
Abstract
Over the last decades, robotic devices for neurorehabilitation have been developed with the aim of
providing better and faster improvement of motor performance. These devices are being used to help
patients repeat movements and (re)learn different dynamic tasks. Over the years, these devices have
become bigger and more complex, so as to provide the end user with a more realistic and
sophisticated stimuli while still allowing the experimenter to have control over the interaction forces
that can potentially shape the motor behaviour. However, experimental results have shown no clear
advantage of these complex devices over simpler versions. In this context, this thesis investigates
sensory-motor processes of human interaction, which can help us understand the main issues for
rehabilitation devices and how to overcome the limitations of simple devices to train particular motor
behaviours.
Conventional neurorehabilitation of motor function relies on haptic interaction between the
patient and physiotherapist. However, how humans deal with human-human interactions is largely
unknown, and has been little studied. In this regard, experiments of the first section of the thesis
investigate the mechanisms of interaction during human-human collaborative tasks. It goes from
identifying the different strategies that dyads can take to proposing methods to measure and
understand redundancy and synchrony in haptic interactions. It also shows that one can shape the
interaction between partners by modifying only the visual information provided to each agent.
Learning a novel skill requires integration of different sensory modalities, in particular vision and
proprioception. Hence, one can expect that learning will depend on the mechanical characteristics of
the device. For instance, a device with limited degrees of freedom will reduce the amount of
information about the environment, modify the dynamics of the task and prevent certain error-based
corrections. To investigate this, the second section of the thesis examines whether the lack of
proprioceptive feedback that is created due to mechanical constraints or haptic guidance can be
substituted with visual information.
Psychophysical experiments with healthy subjects and some preliminary experiments with stroke
patients presented in this thesis support the idea that by incorporating task-relevant visual feedback
into simple devices, one could deliver effective neurorehabilitation protocols. The contributions of the
thesis are not limited to the role of visual feedback to shape motor behaviour, but also advance our
understanding on the mechanisms of learning and human-human interaction.
providing better and faster improvement of motor performance. These devices are being used to help
patients repeat movements and (re)learn different dynamic tasks. Over the years, these devices have
become bigger and more complex, so as to provide the end user with a more realistic and
sophisticated stimuli while still allowing the experimenter to have control over the interaction forces
that can potentially shape the motor behaviour. However, experimental results have shown no clear
advantage of these complex devices over simpler versions. In this context, this thesis investigates
sensory-motor processes of human interaction, which can help us understand the main issues for
rehabilitation devices and how to overcome the limitations of simple devices to train particular motor
behaviours.
Conventional neurorehabilitation of motor function relies on haptic interaction between the
patient and physiotherapist. However, how humans deal with human-human interactions is largely
unknown, and has been little studied. In this regard, experiments of the first section of the thesis
investigate the mechanisms of interaction during human-human collaborative tasks. It goes from
identifying the different strategies that dyads can take to proposing methods to measure and
understand redundancy and synchrony in haptic interactions. It also shows that one can shape the
interaction between partners by modifying only the visual information provided to each agent.
Learning a novel skill requires integration of different sensory modalities, in particular vision and
proprioception. Hence, one can expect that learning will depend on the mechanical characteristics of
the device. For instance, a device with limited degrees of freedom will reduce the amount of
information about the environment, modify the dynamics of the task and prevent certain error-based
corrections. To investigate this, the second section of the thesis examines whether the lack of
proprioceptive feedback that is created due to mechanical constraints or haptic guidance can be
substituted with visual information.
Psychophysical experiments with healthy subjects and some preliminary experiments with stroke
patients presented in this thesis support the idea that by incorporating task-relevant visual feedback
into simple devices, one could deliver effective neurorehabilitation protocols. The contributions of the
thesis are not limited to the role of visual feedback to shape motor behaviour, but also advance our
understanding on the mechanisms of learning and human-human interaction.
Date Issued
2011
Date Awarded
2012-02
Copyright Statement
Attribution NoDerivatives 4.0 International Licence (CC BY-ND)
Advisor
Burdet, Etienne
Sponsor
CONACYT (Mexico) and EU grant HUMOUR
Grant Number
FP7-ICT- 231724
Publisher Department
Bioengineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)