EEG PROCESSING, CLINICAL & MOBILE APPLICATIONS, CUSTOM SOLUTIONS, AND MORE
Learn about the Rekonas
Application Platform
Primary Services
Our primary services address general problems that are prevalent in daily clinical EEG workflows. They aim to be the most efficient and precise tools for providing solutions to clinically relevant issues within time-constrained day-to-day operations.
Our primary offering includes the following services:
Our primary offering includes the following services:
Automated EEG Pre-processing
Raw EEG recordings are contaminated by signals of non-neural origin such as heart beat, eye blinks, muscle activity or the occasional cell phone ringing. EEG analysis therefore begins with automatically identifying those artifacts.
Smart EEG Viewer & Sleep Analysis
Accelerating time-consuming visual inspection of EEG by automatically detecting segments of diagnostic value and automatically quantifying the standard clinical parameters of EEG.
Quantified Cognition
The Signal power and signal complexity of EEG contain information about future cognition in Parkinson's disease. The worsening of cognition, i. e. the risk of dementia within the next three years, can be estimated.
Mobile EEG Devices
On our platform, combining periodic neurological in-patient assessments with the possibility of near-continuous mobile data collection provides superior treatment-monitoring for patients and pharmacological studies.
Custom Application Development
Research projects and pharmacological studies often have very specific needs. We develop custom applications for our platform, integrate mobile and clinical applications and prepare the necessary steps for certification.
Automated EEG Preprocessing
The essence of EEG is the amplification of tiny currents of neural origin into a graphic representation that can be interpreted. Unfortunately, while intending to record only brain oscillations, other non-neural potentials are likewise amplified. These unwanted signals may be due to body movement, eye blinking or heart beat. The neural signal, which is often many times smaller in amplitude compared to likely sources of interference, gets obscured and, as a consequence, any downstream analysis - whether computerized or through visual inspection - may suffer both in accuracy as well as the time required to complete the analysis.
Heart beat artifact. Diffuse sharp potentials (arrows) coincident with the electro-cardiogram (ECG). ECG artifact is particularly prominent in the obese and those with hypertension [2].
Muscle artifact. The muscle artifact shown here (arrows) is most prominent in the frontal head region and obsures the underlying neural information which is still present in the contaminated signal [2].
Eye blink artifact. The figure shows eye opening (thin arrow) and eye closure (thick arrow) [2].
Processing and interpreting EEG is challenging as it requires taking into account the fact that all EEG signals are mixtures of both electro-cortical and extracerebral signals. Our early preprocessing pipeline can significantly increase the signal-to-noise ratio and reduce unwanted artifacts in a fully automated fashion. Three different possibilities on how to handle artifacts are available to any user:
HIGHLIGHT
Artifacts, including their likely origin, may simply be highlighted in a given EEG recording. This makes the downstream task of visual inspection by a trained EEG reader much less time consuming.
REJECT
After an artifact has been identified, the contaminated sequence can be removed completely from the EEG recording. After rejecting all contaminated parts of a recording, the remaining sequences can be stitched together. The resulting EEG will be artifact free but shorter in overall duration.
RECOVER
Advanced machine learning methods allow to recover the neural signals hidden underneath the artifacs. Based on this method, the recording duration will remain unchanged, which is especially of advantage in developmental populations where a high degree of artifact contamination is present and the loss of data using a rejection approach would be unsustainable.
Smart EEG Viewer & Sleep Analysis
Inspecting EEG can be time consuming. The perceived level of abnormality of a recording is highly correlated with inspection time: The more unusual an EEG is, the longer it will take to review. But reviewer experience also plays a significant role: an inexperienced reviewer will naturally be more cautious and take more time to finalize a review compared to a more seasoned neurologist. Brogger et al. have shown that for standard EEG recordings as well as recordings taken after sleep deprivation, reading times can increase up to threefold due to inexperience [1]. Our smart EEG viewer aims to make inspection times less dependent on reviewer profile by providing a computerized, preliminary EEG analysis highlighting the sections of the EEG recording that have the potentially highest diagnostic value.
The more abnormal an EEG looks, the longer it will take to review. In any case, time to complete a review significantly depends on reviewer experience and can be up to 3 times higher for less experienced reviewers.
Medically certified smart EEG viewer.
Our smart EEG viewer - in combination with automated preprocessing - quantifies posterior dominant rhythm, reactivity, anterior-posterior gradients, potential diffuse slow-wave activity, and hemispheric imbalance. This provides greater objectivity compared to purely visual inspection and descriptive reporting of findings. An automated numerical evaluation of EEG is central to quantitative EEG (qEEG) and provides a more reliable basis for comparisons. Furthermore, our smart EEG viewer offers spike detection and classification, source localization, automated sleep staging and the quantification of cognition. Importantly, our smart EEG viewer is accompanied by a normative database of healthy individuals (N=200, young & old) for which the features quantitified by our viewer are available for reference (Note: This database is not yet medically certified).
Neurological disorders may be accompanied by elevated levels of daytime sleepiness and patients undergoing resting state EEG might be more prone to micro-sleep during the recording session. Our EEG viewer can automatically highlight these portions of micro sleep. It is then up to the reviewer to either automatically remove these sections from the EEG or to treat their frequency of occurence as an additional source of diagnostic information.
Quantified Cognition
Depending on both the disease and the time scale under consideration, neurodegeneration does not automatically imply cognitive worsening: In Parkinson's disease (PD) the most debilitating manifestations of disease progression is cognitive decline and is a key determinant of a patient's quality of life and independence. Based on the CamPaIGN study [3], the dementia prevalence rate 5 years after diagnosis is 15 to 20% and increases to 46% at 10 years.
For recruitment into and stratification of clinical trials, particularly those designed to slow disease progression and to prevent dementia, the ability to predict this complication is highly relevant.
Resting state EEG is a predictive biomarker with high accuracy of defining patient cohorts at risk of cognitive decline, leading to dementia, within 3 years after diagnosis.
For recruitment into and stratification of clinical trials, particularly those designed to slow disease progression and to prevent dementia, the ability to predict this complication is highly relevant.
Resting state EEG is a predictive biomarker with high accuracy of defining patient cohorts at risk of cognitive decline, leading to dementia, within 3 years after diagnosis.
EEG recordings taken in a resting state denote spontaneous neural activity. This spontaneous activity contains the information necessary to predict cognitive decline. The most significant difference between cognitively stable patients and patients at risk of cognitive decline manifest in the 4-8Hz range, also known as the theta band. If the spectral power in this band is elevated and the signal complexity is reduced compared to age matched healthy controls (HC), then the risk of cognitive decline within the next 3 years is significant.
Mobile EEG Devices
Intermittent vs Continuous Monitoring is the main difference between reactive and proactive healthcare: We will provide an integrated and certified device for mobile multimodal neurological assessments aimed at continuous or near continuous collection of brain physiological data. This will allow continuous risk assessment, as well as critical event detection, and generate more data points, in turn permitting more accurate trending of patient status.
Table 1: Comparison of drug development within and outside the central nervous system [4].
Especially of interest - both for research and in pharmacological studies - are monitoring and progression biomarkers in neurodegeneration based on mobile EEG devices. Here, Multiple Sclerosis and Alpha-Synucleinopathies such as Parkinson's disease and dementia with Lewy bodies, are situations where mobile monitoring can provide added value regarding risk assessment, stratification and outcome prediction.
Currently, our mobile EEG device is being tested at the University Hospital Basel. We welcome interested collaboration partners.
Please contact us for more information.
Custom application development
Our cloud-based EEG analysis platform can host your custom developed application, collect, host and analyze your study or research data, and integrate mobile and clinical EEG devices into your workflow.
Our application development process While our primary sevices address a wide range of problems, specific projects often come with very specific requirements. From implementing and automating custom pattern recognition algorithms or neural networks, to specific artifact detection techniques, and the design of evoked potential workflows. Such needs may arise as part of large multi-center research projects in order to ensure that each centre has access to identical data processing pipelines, or as part of pharmacological studies, where EEG might be used as an additional biomarker for monitoring progress in a mobile fashion from within the patient's home environment.
Application development is a multi-step process. Each step defines an important milestone. In this sequential model, development will continue only once the previous milestone has been fully achieved.
1
Product Vision
3
User Journey
5 Deployment
Ideation & Research
2
Development 4
Maintenance 6
1 Product Vision
Our custom development process begins with understanding the intention with which your apllication is being created, and the goal it aims to achieve for your customers or users. Once we fully understand your vision we can enter the Research and Ideation phase of your project.
2 Ideation & Research
EEG, like any other technology, is not a one size fits all solution. We asses your product vision and provide an opinion whether EEG is the technology best suited to your needs. Based on our experience, we compile current research relevant to your goals, compare to other available technology and provide you with a well founded opinion about how to proceed best.
3 User Journey
A user journey describes how a user interacts with a product and allows designers to see a product from a user's point of view. Who is your user - a patient, a medical professional or a member of your team? What are the expectations of a user when interacting with your application? What are tyical scenarios and user expectations? Together, we will answer these questions and come up with specific designs - in software and in hardware, if necessary.
4 Development
Your application will be deployed on the Rekonas platform for digital health applications. After the initial release, we take care of the productivity and performance for the lifetime of your application. We establish continuous server monitoring, provide general support services and allocate experts to fix bugs.
5 Deployment
Your application will be deployed on the Rekonas platform for digital health applications. After the initial release, we take care of the productivity and performance for the lifetime of your application. We establish continuous server monitoring, provide general support services and allocate experts to fix bugs.
6 Maintenance
Maintenance of your application for the full duration of your research project or study is ensured through experts allocated to your project. We focus on delivering updates in form of rolling releases, ensuring optimal data quality through randomized plausibility checks and responding to user feedback.
Our application development process
While our primary sevices address a wide range of problems, specific projects often come with very specific requirements. From implementing and automating custom pattern recognition algorithms or neural networks, to specific artifact detection techniques, and the generation of non-standard evoked potential workflows. Such needs may arise as part of large multi-center research projects in order to ensure that each centre has access to identical data processing pipelines, or as part of national and international pharmacological studies, where EEG might be used as a mobile progression biomarker in the patient home environment.
1 Product Vision
Our solution discovery experts analyze your requirements and estimate the scope of work and timelines. We set up a supervised and cross-functional team involving designers to conduct profound market and user research to align your solution with customers’ real needs.
2 Ideation & Research
Our solution discovery experts analyze your requirements and estimate the scope of work and timelines. We set up a supervised and cross-functional team involving designers to conduct profound market and user research to align your solution with customers’ real needs.
3 User Journey
Our solution discovery experts analyze your requirements and estimate the scope of work and timelines. We set up a supervised and cross-functional team involving designers to conduct profound market and user research to align your solution with customers’ real needs.
4 Development
Our solution discovery experts analyze your requirements and estimate the scope of work and timelines. We set up a supervised and cross-functional team involving designers to conduct profound market and user research to align your solution with customers’ real needs.
5 Deployment
Our solution discovery experts analyze your requirements and estimate the scope of work and timelines. We set up a supervised and cross-functional team involving designers to conduct profound market and user research to align your solution with customers’ real needs.
6 Maintenance
Our solution discovery experts analyze your requirements and estimate the scope of work and timelines. We set up a supervised and cross-functional team involving designers to conduct profound market and user research to align your solution with customers’ real needs.
References
- Visual EEG reviewing times with SCORE EEG [link]
Brogger J, Eichele T, Aanestad E, Olberg H, Hjelland I, Aurlien H. (2018). Clin Neurophysiol Pract. , 3:59-64 - Rowan's Primer of EEG E-Book [link]
Marcuse, L. V., Fields, M. C., & Yoo, J. Y. J. (2015). Elsevier Health Sciences. - The CamPaIGN study of Parkinson's disease: 10-year outlook in an incident population-based cohort. [link]
Williams-Gray, C. H., Mason, S. L., Evans, J. R., Foltynie, T., Brayne, C., Robbins, T. W., & Barker, R. A. (2013). Journal of Neurology, Neurosurgery & Psychiatry, 84(11), 1258-1264. - ... [link]
Marcuse, L. V., Fields, M. C., & Yoo, J. Y. J. (2015). Elsevier Health Sciences.