Automated, Intelligent Seizure Detection
In 2017 Neuro Event Labs created Nelli
Nelli is an intelligent algorithm that analyses video and audio to deliver quantitative and objective diagnostic insights for movement disorders such as epilepsy.
No wearables, no capital investments, just simple intuitive diagnostics.
Why is this important & what is the challenge
Patients with Epilepsy are guided through an established pathway for treatment. During that journey, there are several instances of diagnostic video electro encephalography (VEEG). Because VEEG is such an integral part of the pathway and instrumental in the decision making process, anything we can do to improve on this existing solitary method will help to improve patient outcomes for patients whose lives are touched by Epilepsy.
Video EEG & EMU is Costly
The average cost for VEEG (Video Electro Encephalography) is €5000 per patient including technical equipment, staff and analysis.
Seizure diaries are less than 30% accurate
The gold standard for seizure history is the patient reported seizure diary. Most treatment decisions and drug changes are based on the patient reported seizure diary. Unfortunately, several studies over the last 3 decades have proven that this data is highly unreliable.
EMU Queues Risk Patient Outcomes
On average, a patient will wait for 6 - 12 months for an EMU (epilepsy monitoring unit) admission. The EMU is crucial for capturing seizures and often patient treatment is put on hold waiting for results from the EMU. Time to treatment is a key factor in successful epilepsy treatment.
Clinicians often debate seizure classification
The nature of seizures is complex, seizure classification is the holy grail for diagnosis leading to appropriate treatment. Experience is a dominating factor in the classification of seizures and as such, debates are common during patient review.
50% of EMU stays fail
The average EMU stay is 1-2 days. During this limited time, half of all patients do not experience seizures. The reasons for this are mostly attributed to the length of the EMU stay, but environmental factors are also contributing to this phenomenon.
PNES - Psychogenic Non-Epileptic Seizures
A challenging differentiator in the treatment of Epilepsy is understanding if a patient has PNES or Epilepsy. 40% of adults and 30% of children admitted to the EMU are later diagnosed with PNES. Only after a true diagnosis is administrated are patients adequately treated for their symptoms.
Hello Nelli
What is it?
Nelli is a set of algorithms stored in the cloud which analyse video and audio in order to detect and classify events as seizures. There are no wearables, only a camera which observes the patient in hospital or at home for an unlimited amount of time and continuously reports seizures (events) to the doctor.
How do you collect video and audio?
Using a small camera, mounted on a micro computer. This device is easily mounted anywhere, in patient homes and hospitals. The only thing it requires is a power source.
When the camera is installed, the patient will have a remote control in order to easily turn the camera on and off.
How does the cloud get the data?
Data is transfered securely using a data connection built in to the micro computer. Neuro Event Labs follows all regulations stipulated by the HIPPA and GDPR bodies in both Europe and North America.
What happens in the cloud?
Magic. When the cloud receives the data, it is then analysed by more than 50 algorithms which look at every single pixel and listen to each and every sound wave. This is a very heavy process which is why we have super computers to do the heavy lifting. Each event is individually analysed by the algorithms and decisions are made as to the events classification based on what the algorithms have been taught, what they know and how sure they are that this event correlates with what they have seen before. As each algorithm operates independently, they all deliver individual scoring to a master classifier which then takes the final decision on that events individual classification.
How does the doctor get the results
When the results are ready, and interactive report is sent to the doctor where they are able to see a complete overview of all events for the recording period and dive deep into each individual event. Probably one of the more useful tools is the fact that doctors are also able to play back each event - they can watch the seizures.
A look through Nelli’s eyes
One of Nelli’s many algorithms analyses the movement of pixels on the Z axis - also known as depth. This allows Nelli to decipher, amongst many other things, breathing oscillation.
During a test in the sleep lab, this algorithm was put to the test agains a RIP belt (respiratory inductance plethysmography) used is a method of evaluating pulmonary ventilation by measuring the movement of the chest and abdominal wall.
The test concluded that Nelli’s oscillation algorithm was 99% specific when compared to the RIP belt.
The video on the right shows one of our engineers sleeping whilst the algorithmic calculations are analysing the breathing oscillation.
Nelli is effective
Capture Rate
Nelli is able to capture more seizures than both VEEG and Seizure Diaries combined. This is achieved due to the fact that the camera is continuously collecting data on the patient supported by the fact that the algorithms are highly tuned not to miss anything.
Cost Per Day
At a fraction of the cost compared to VEEG and hospital stays, Nelli is able to operate cost effectively. This supported with the fact that Nelli applies the value based health care (VBHC) principle, means that insurance companies, governments and hospital systems have a guaranteed value proposition.
99%
Sensitivity
The camera used in Nelli has a lot of important features including:
120m pixels
dual lenses
depth sensors
4 microphones
This means that micro movements, not even visible to the human eye are caught by the camera and reported to the algorithm.
>90%
Specificity
Nelli has been developed using prospective data from over 300 patients and 10 000 seizures.
Validation studies have been completed on more than 50 patients with 1200 seizures.
More validation studies are currently ongoing and results are expected soon.
Implementing Nelli
It is very easy to implement Nelli in your current practice. Contact Go North Medical by email or give us a call at +46733319271 and we will be happy give you more information about Nelli and how we can help.
More than 30 hospitals worldwide use Nelli today in their current practice, with outstanding results, saving money and improving patient outcomes.