The power of digital technology at the service of physiological data processing
Physip, founded in 2002, is a spin off of Telecom ParisTech: the company project derives from a PhD work on the digital processing of sleep EEG signals. Physip is pioneer in the use of digital technologies for the analysis of EEG signals.
Physip has a complete complementary know-how: signal processing applied to EEG, learning and classification methods, processing of massive data, identification of markers.
We have developed an exceptional portfolio of technologies
- Automated staging of sleep ASEEGA
- Automated evaluation of drowsiness MEEGAWAKE
- Evaluation of sleep pressure: MEEGAFIT
- A platform for the expertise of physiological signal
2 sensor EEG:
an exclusive approach, a decisive technological advance
Since the beginning, Physip developed all its innovative module with the strong constraint of reducing the number of sensors. All our analyses need only 2 EEG electrodes. Why this constraint? To open the possibilities of EEG, to take EEG out of the lab. Reducing the number of sensors makes it faster for clinicians and technicians to prepare the patient or subject, makes ambulatory applications easier – even professional applications in a near future. Reducing the number of sensors reduces the volume of data for distant applications, for telemedicine.
Physip conducts an demanding R&D. It allows us to keep developing new products and improving existing ones to address new scientific and industrial needs.
Efficiency and rigor
R&D at Physip is mostly collaborative. We work with academic and industrial partners. We have a network of expert researchers supporting our R&D in the field of sleep and vigilance, cognitive sciences, performance, signal processing. We work with industrial partners in the field of transportation, aeronautics, automotive, the military…
With a dozen of successful R&D projects, we have a unique experience in developing new algorithms for the identification of significant EEG markers, from the exploratory phase to the clinical validation or validation in simulated conditions.
Ongoing or completed research projects
Scoal : Sleep, Cognition and COgnitive decline in ALzheimer’s Disease, 2011-2015
Objectifve – The SCOAL project was pursued to better understand sleep disorders (changes in the macro and micro architecture of sleep) that are a precursor of Mild Cognitive Impairments and their potential contribution to a progressive deterioration of cognitive function. For Physip, the project involved looking for early diagnostic and prognostic markers for cognitive decline and Alzheimer’s Disease.
Support – SCOAL was supported by the MALZ ANR program.
Results – Results were presented in the ESRS congress in 2014 in Tallinn.
Sagaspe P, Taillard J, Chaufton C, Berthomier C, Brandewinder M, Amiéva H, Dartigues J-F, Philip P. Polysomnographic data in patients with isolated memory complaints or mild cognitive impairment. J. Sleep Res., 23 Suppl. 1:240, 2014
Objective – Meegaperf aims to identify automatically and in real time, EEG markers indicating a risk of impaired cognitive and physical performance. The project consists of recording the EEG of subjects involved in physical and cognitive tasks conducive to a reduction or breakdown in performance, in order to identify the cerebral signs making it possible to predict theses performance breakdowns.
Support – Meegaperf is a Rapid Project, supported by the DGA and DGE. Didier Bazalgette, in charge of Human Factor Research at the DGA, is the technical pilot for this project.
Partners – Working on physical performance imposed to imagine a complex system, allowing the synchronized recording of multiple physiological signals, in relation to physical performance evaluated through the pedaling frequency on a cycle ergometer. The design and realization of the recording system brought together all the partners and various skills to produce exceptional data.
Meegaperf, led by Physip, brought together a team involving the following contributors:
- Inserm Unit U1028, CRNL (Lyon Neuroscience Research Center): Jérémie Mattout, Aline Bompas, Emmanuel Maby, Patrick Bouchet, worked on the project, on the cognitive performance aspects and on the recording of the EEG signal during physical performance
- The LIBM (ex-CRIS, Inter-University Laboratory of Human Movement Biology): Christian Collet, Vincent Pialoux, Michel Clémençon, Thomas Creveaux on physical performance.
- The Signal and Image Processing Department of Telecom ParisTech: Slim Essid, Cecilia Damon, François Rigaud, Alexis Benichoux worked on original artefact processing methods.
Objective – To develop an automatic method for analyzing sleep pressure, measured using the Karolinska Drowsiness Test (KDT), based on an expert artifact rejection solution.
Partners – The project was conducted with the CNRS/ Bordeaux University Unit Sanpsy, with Jacques Taillard.
Support – The project was supported by BPI and Paris Region Entreprises (former “Centre Francilien de l’Innovation”).
Results – An innovative automated method for artifact rejection was developed and validated. It allows fast, objective and reproducible preprocessing of data before the power spectrum is calculated on the 6‑9 Hz frequency band.
Results were presented in oral communication in Le Congrès du Sommeil (the Annual Meeting of the French Sleep Society) in Marseille in 2013.
DAM : Drowsiness Automatic Monitoring, 2006-2008
Objectives – DAM originally aimed to evaluate the feasibility of an automated and real-time measurement of drowsiness based on a single EEG channel. It became a full development study part of the way through, given the excellent preliminary results.
Partners – Dr. Alain Muzet and Forenap participated in the DAM project, for the recording of data in a driving simulator, and visual expertise of EEG recordings.
Support – DAM was supported by the French Ministry of Defense (DGA).
Results – The DAM project resulted in the DAM algorithm, capable of describing automatically and in real-time the level of drowsiness of an active subject, on a scale ranging from 0 – full alertness, to 4 heavy drowsiness.
Results were presented in ESRS congress and in Le Congrès du Sommeil (the Annual Meeting of the French Sleep Society) in 2008.
Berthomier C, Muzet A, Berthomier P, Prado J, Mattout J, Real-Time Automatic Measurement of Drowsiness based on a Single EEG Channel
Objectives – The study aimed to validate the performance of the ASEEGA® technology, by comparing the visual analysis of 2 expert clinicians, and automated analysis, on 15 PSG recording on healthy subjects.
Partners – Sleep recordings and visual analysis of PSGs were conducted by the team of the Functional Exploration Department of the Henri Mondor University Hospital: Prof. Marie-Pia d’Ortho, Dr. Xavier Drouot, Dr. Maria Stoïca, RPSGT Djibril Bokar-Thire.
Support – The study was supported by Paris Region Entreprises (formerly Bio-CRITT) funding.
A detailed evaluation of the performance of the ASEEGA® technology was published in the journal SLEEP, in 2007.
Automatic analysis of single-channel sleep EEG: validation in healthy individuals. Berthomier C, Drouot X, Herman-Stoïca M, Berthomier P, Prado J, Bokar-Thire D, Benoit O, Mattout J, d’Ortho MP. Sleep. 2007 Nov;30(11):1587-95.
Co-founder of Physip, CEO
co-founder of Physip, Director of IT
co-founder, General Manager
Frequently Asked Questions
What is Physip?
Physip is a French company located in Paris, in the 11th arrondissement, that has developed innovative methods for analysing sleeping and waking brain activity using electroencephalography. We offer these methods to doctors, researchers, pharmaceutical and medical device companies to optimise their sleeping and waking analysis data.
Is the ASEEGA® analysis reproducible?
Yes, ASEEGA® is 100% reproducible: A same single trace analysed twice by ASEEGA® will produce exactly the same result.
What format are the results provided in?
Pdf reports are available. Results are provided as Excel reports, Matlab structure or XML files.
Does ASEEGA® analyse EOG and EMG signals?
No, ASEEGA® analyses EEG signals exclusively. However, for practical purposes, if you have recorded PSG traces, you can send the entire trace and we will perform for you the extraction of the EEG channel, which we will then analyse.
Are there any exclusion criteria for the traces?
The traces must comply with the ASEEGA®, specifications, in terms of duration, sampling frequency, amplitude resolution and bandwidth. For more details regarding the specifications for the recordings and the continuation of the polysomnographs and polygraphs, see here.
Berthomier C. & Brandewinder M., EOG-based auto-staging: less is more, Sleep Breath, 2015
This issue of Sleep and Breathing presents the validation results of a new automated wake/sleep staging method based on EOG activity, developed by Jussi Virkkala from the Finnish Institute of Occupational Health. Classically, the automated method is compared to visual analysis, on an epoch by epoch basis. It reaches a level of global concordance of 88 % with a Kappa of 0.57. In other words, on the 248,696 epochs of the validation dataset, 212,138 were scored correctly in wake/sleep, that is as the human expert did it, and on 36,558 epochs, the two scorings differ.
Berthomier C. & Brandewinder M., Sleep scoring: man vs. machine ? Sleep Breath., 2013
Berthomier C, Brandewinder M, Sleep scoring: man vs. machine ?, Sleep Breath. 17 (2):461-462, 2013 The automated analysis of sleep has grown in interest in the past decade. Advances in computing have brought the needed intensive calculations within reach ; while simultaneously, there is an increasing demand for sleep diagnosis and analysis. The prevalence of sleep troubles is high, and the awareness of their consequences is spreading among patients, health authorities, and clinicians. This awareness is directing more and more patients to sleep centers. The upward trend in demand for sleep evaluations concerns not only sleep specialists. Sleep appears to be an extremely promising territory for other fields, such as cardiology and nutrition for example . Needs exceed capacities by this far. Data analysis has been identified as one of the bottlenecks in the sleep evaluation process, making clear the importance of developing tools to facilitate analysis. These developments have an impact that is medical, as well as economical and social. > Accéder à l’article complet
Schmidt C, et al., Circadian preference modulates the neural substrate of conflict processing across the day, PLoS One 2012
Schmidt C, Peigneux P, Leclercq Y, Sterpenich V, Vandewalle G, Philips C, Berthomier P, Berthomier C, Tinguely G, Gais S, Schabus M, Desseilles M, Dang-Vu T, Salmon E, Degueldre C, Balteau E, Luxen A, Cajochen C, Maquet P, Collette F. Circadian preference modulates the neural substrate of conflict processing across the day. PLoS One. 2012;7(1):e29658. 2012 Jan 4. PLoS One. 2012;7(1):e29658. 2012 Jan 4.
Human morning and evening chronotypes differ in their preferred timing for sleep and wakefulness, as well as in optimal daytime periods to cope with cognitive challenges. Recent evidence suggests that these preferences are not a simple by-product of socio-professional timing constraints, but can be driven by inter-individual differences in the expression of circadian and homeostatic sleep-wake promoting signals. Chronotypes thus constitute a unique tool to access the interplay between those processes under normally entrained day-night conditions, and to investigate how they impinge onto higher cognitive control processes. Using functional magnetic resonance imaging (fMRI), we assessed the influence of chronotype and time-of-day on conflict processing-related cerebral activity throughout a normal waking day. Sixteen morning and 15 evening types were recorded at two individually adapted time points (1.5 versus 10.5 hours spent awake) while performing the Stroop paradigm. Results show that interference-related hemodynamic responses are maintained or even increased in evening types from the subjective morning to the subjective evening in a set of brain areas playing a pivotal role in successful inhibitory functioning, whereas they decreased in morning types under the same conditions. Furthermore, during the evening hours, activity in a posterior hypothalamic region putatively involved in sleep-wake regulation correlated in a chronotype-specific manner with slow wave activity at the beginning of the night, an index of accumulated homeostatic sleep pressure. These results shed light into the cerebral mechanisms underlying inter-individual differences of higher-order cognitive state maintenance under normally entrained day-night conditions.