Internship / thesis project: Philips Research Capacitive ECG and sleep staging in preterm infants

Introduction:
Infants born prematurely sleep about 70% of 24 hours. They do this to create a safeguarded setting which is needed for proper brain development. To trace their brain development sleep staging performed manually mainly by neonatologists or neurologists. We are working on an algorithm which can automatically determine the sleep states in preterm infants to provide continuous development monitoring. Until now, we use classic adhesive electrodes to obtain the vital signs. This is very obtrusive for the preterm infants. The goal is to obtain the vital signs unobtrusively and use them for automated sleep staging with our algorithm. To achieve this goal we need you.

Task I (3-4 month) Journal/conference publishable results are expected:
• Can motion of preterm infant data be categorized into movement, caretaking, technical noise, etc.?
• Use motion to stabilize ECG for analysis algorithm (Noise filter).
• Use these information on a capacitive ECG dataset to identify noise generated by capacitive electrodes.
• Use motion to stabilize cECG for analysis algorithm (Noise filter).
• Repair clipped/noisy signals.
• Use statistical analysis to show separability of preterm infant sleep states based on movement.

Task I + Task II in the following (6 month):
• Based on Task I, the student should create new features for sleep stage classification.
• The features should be implemented in an existing classifier to improve performance.

Provided: We will provide a dataset with several vital sign measurements. Next to others, the vital signs contain capacitive ECG waveform data and the corresponding gold standard ECG waveform data.

Requirements: Signal processing skills. Knowledge of coding with Matlab and Python. Reading existing Matlab (and Pyhton) code. Basic statistic knowledge.

Skill level: The student should be from electrical engineering or computer science. Master student are preferred. The tasks can be adapted to the length of stay and the level of task (internship / Thesis). The student should feel comfortable in Matlab. Basic knowledge in Python would be optimal.

Location: The Student will be based at Philips Research, High Tec Campus 34

Supervision: The Student will be supervised by a Ph.D. student and a research scientist of Philips.

Contact: werth.jan@philips.com