Estimating quality of multimodal biosignals in real-time

Company

Open Mind Neurotechnologies (Open Mind) is a 3-year old French neurotechnology startup dedicated to wellness and healthcare applications of virtual reality (VR) and quantitative neurophysiology for the general public. Its first B2C service offers a 360-degree assessment of individual mental wellbeing. It uses a combination of psychometric questionnaires with physiological and behavioral measurements taken during an immersive experience in VR to quantify several dimensions of mental health in a dynamical approach. Open Mind employs a multidisciplinary team of 20 people including 5 PhDs. The scientific team includes data scientists, experimental researchers and psychologists, as well as software engineers. The creative studio team includes game developers and programmers specialized in VR as well as graphic and sound designers. Through partnerships with clinical and academic institutions in applied research and product development, Open Mind aims at transforming the mental health framework by creating new avenues, within and beyond healthcare, to maximize the duration of a healthy, functional and meaningful living.

Work objectives

Characterize real-time signal quality indices for biosignals obtained during specific Virtual Reality (VR) protocols.

Description

Open Mind consumer services are based on the collection and analysis of multimodal biometric data using research level measurement tools. In this context, we want to ensure that data acquisition is reliable (robustness and replicability in different recording conditions) and qualitative (highest possible SNR, minimal undesirable activities measured). Therefore at every moment it is essential to have a clear idea of the validity and noise level in the collected biosignals.

This internship aims to develop quality indices of physiological signals, especially respiratory movement (resp), heart pulse (PPG) and electrodermal activity (GSR). These indices must enable tracing information related to, for example, improper installation of the sensors or to movement artefacts.

Inspired by the literature and adapted to our specific software-hardware tools, biosignal processing pipelines for quality index estimation will first be tested on existing databases by approximating the constraints of real-time processing. After adjustments and optimization, the obtained indices will be implemented as modules within TimeFlux, an open-source framework for the real-time processing of biosignals (https://timeflux.io/).

If the results are convincing, the indexes implemented during this internship will be directly applied and used within our services, and might also be shared to benefit the community at large.

* Note that the subject may be extended to other measurement methods (EEG, eyetracking, etc.) and be published in peer-reviewed journals according to the interest, skills and
motivation of the candidate.

Materials and skills

• The student will work with Python to develop pipelines of data analyses.
• Spoken language is French and/or English.

Masters specialization

• Machine learning, signal processing, computer science, neurosciences

Your profile

• Strong skills in scientific computing (Python) and previous projects in data mining with time series. Interest on signal processing and machine learning applied to biological signals.
• Curiosity for new technologies, IoT, self quantified, cognitive sciences, and their applications for well-being.
• Openness and ability to collaborate with various collaborators in a multidisciplinary environment.
• Versatility, hacker spirit and DiY: you know how to inform yourself, interact and think out of the box to work out new problems.

Tutors

• Raphaëlle Bertrand-Lalo, Ms, data scientist / raphaelle@omind.me
• David Ojeda, PhD, data scientist / david@omind.me

Practical info

• Duration: 3 to 6 months (flexible)
• Location: Caen or Paris (France)

How to apply

Send us your resume and cover letter to internship@omind.me with email subject “application internship biosignal quality”