Quantifying and Predicting Upper Limb Stroke Recovery in a Computer Game ■
Subject ■
We have developed in previous research a game based rehabilitation
framework for upper limb rehabilitation in stroke patients: patients are tracked with a
Kinect camera and are trying to virtually move objects on the screen in a reaching task. A
cohort of stroke patients has been evaluated over many weeks with this task, as well as
with a battery of validated clinical tests.
While performing the task, the subject was tracked with the Kinect camera, the hand
movement is tracked with an IMU sensor and trunk, shoulder and arm movements are
captured by means of EMG sensors.
This constitutes a unique dataset to investigate two open questions:
- Can the Kinect, the IMU or the EMG capture the improved quality of the hand
reaching task over the rehabilitation trajectory and in how far do metrics from
these sensors correlate with the evolution of clinical scores?
- Can the game based performance (task performance in the game, or any sensory
modality) shortly after stroke be used to predict stroke recovery?
framework for upper limb rehabilitation in stroke patients: patients are tracked with a
Kinect camera and are trying to virtually move objects on the screen in a reaching task. A
cohort of stroke patients has been evaluated over many weeks with this task, as well as
with a battery of validated clinical tests.
While performing the task, the subject was tracked with the Kinect camera, the hand
movement is tracked with an IMU sensor and trunk, shoulder and arm movements are
captured by means of EMG sensors.
This constitutes a unique dataset to investigate two open questions:
- Can the Kinect, the IMU or the EMG capture the improved quality of the hand
reaching task over the rehabilitation trajectory and in how far do metrics from
these sensors correlate with the evolution of clinical scores?
- Can the game based performance (task performance in the game, or any sensory
modality) shortly after stroke be used to predict stroke recovery?
Kind of work ■
To answer the two questions above, by implementing a state of the art signal
processing pipeline on Kinect, IMU and/or EMG, exploiting the task information to
compute reliable metrics of the quality of movement and task completion. The main
challenge is robustness of a variety of unpredicted events/movements during the task.
processing pipeline on Kinect, IMU and/or EMG, exploiting the task information to
compute reliable metrics of the quality of movement and task completion. The main
challenge is robustness of a variety of unpredicted events/movements during the task.
Framework of the Thesis ■
Literature Review (ETOC: 2 months): Familiarize with existing literature on
generation of upper limb stroke rehabilitation, game approaches for the tasks
and sensory processing for such task.
- Define focus on one or more of the sensor modalities, the development of a
reliable algorithm for measuring a relevant metric of motion quality.
- Analysis of temporal evolution and comparison with clinical scores.
generation of upper limb stroke rehabilitation, game approaches for the tasks
and sensory processing for such task.
- Define focus on one or more of the sensor modalities, the development of a
reliable algorithm for measuring a relevant metric of motion quality.
- Analysis of temporal evolution and comparison with clinical scores.
Expected Student Profile ■
(Mandatory) qualifications:
• Following an MSc in a field related to one or more of the following: Computer
Science, Biomedical Engineering, Applied Computer Science - Digital Health.
• Strong programming skills (Python).
• Ability to write scientific reports and communicate research results at
conferences in English.
• Following an MSc in a field related to one or more of the following: Computer
Science, Biomedical Engineering, Applied Computer Science - Digital Health.
• Strong programming skills (Python).
• Ability to write scientific reports and communicate research results at
conferences in English.