Home / Update: 23rd April 2018

Are we there yet?

Dear Backers,      

In this update we will be covering our activities from December ‘17 through to April ‘18. Here’s an outline of what we will be discussing:

  • Overview
  • Team
  • Digital Therapies
  • Industrial design
  • Mechanical design
  • Algorithm development
  • Clinical testing
  • Mobile app development

There’s a lot to fill you in on, so as always, we hope you enjoy the read!

 

OVERVIEW

 

As I am sure a lot of you are probably wondering, where is my order? Over the last 4 months, we’ve been working extremely hard on all fronts and unfortunately, we’ve faced a few setbacks during our final engineering stages. Namely, getting our light panel to diffuse light evenly. The good news is the solution is fixable and we’re currently working and testing a solution to ensure you get a high-quality product. The reason why we’ve delayed updating everyone on numerous occasions is that we wanted to be transparent and have full knowledge before committing to our timelines. The setback means that we’re currently 6-8 weeks behind on our hardware manufacturing schedule, which pushes back our delivery window to summer this year.

Frankly speaking, we apologize and we promise to you that we are 100% dedicated to delivering a quality product. At Circadia, we’ve always pushed for the extra mile and while we can ship our solution in its current state very easily, we have decided to hold back to deliver a much more cohesive and beautiful product. Form follows function they say and while we’ve nailed the functionality, we are now obsessive over the form. We want to make sure that when we ship something, it’s something everyone on the team can be proud of putting our name on. We’ll expand on this in more depth during the update. 

The good news is that while the hardware team have been working day and night to achieve a beautiful, homogenous light diffusion solution, we’ve been expanding our software features. We will now be offering a range of digital therapies within our mobile app to compliment the light therapy feature. These therapies are evidence-based, clinically tested and developed by leading researchers, physicians, and psychologists in the field of sleep medicine from around the world.

In addition to expanding our mobile app feature set, we’ve also been doing serious comparative clinical testing. We’ve put our tracker to the test against the current gold standard, polysomnography and frankly speaking, have achieved remarkable results that have by far, exceeded our own expectations. Our latest accuracy figures put us a mile ahead of all consumer wearable sleep trackers (including the latest Fitbit Alta HR) and even medical wearables like the Philips Actiwatch. We’ll be releasing another whitepaper to share the results soon but here’s a summary of our results:

  • Our respiration tracking accuracy is 95%
  • Our Deep and REM sleep accuracy is 75% from a maximum of 80% (due to the inconsistency if medical equipment)

What does this mean for me?

Clinical grade sleep monitoring data is critical to evaluating what type of sleep improvement tool or exercise you are better suited to. It is key to our AI-powered algorithms which personalize the therapy for you based on analyzing your sleep quality and body clock. So we’ve spent months running these tests and have partnered with a leading university in Europe to continue our testing.

Not happy about the delay and I want a refund?

We understand that the delay will frustrate a lot of our backers, hence, we’ve decided that it is only fair that we reward each backer with our premium app (usually around $100 per year) along with a free 30-minute workshop delivered by our sleep team (typically $25 per person), which includes access to all of the below:

  • All mindfulness & meditation exercises
    • Deep breathing
    • Focused breathing
    • Progressive muscle relaxation
  • All cognitive behavioural therapy exercises
    • Sleep hygiene
    • Stimulus control
  • All circadian exercises
    • Circadian entrainment
  • Personalised sleep insights

Unfortunately, given the stage of the project, we are at, we cannot offer backers refunds. Kickstarter’s policy is to allow backers to file for a refund 14-days after the end of the project. We’ve invested a lot of capital in our development and production (to purchase supplies and materials). Offering refunds at this point will mean that 95% of the other backers that understand the how challenging it is to develop a product with several moving parts, will face even more significant delays, something that we ourselves, will not accept.

So while we can’t offer you a refund, if you are still not satisfied by the end of reading this update, message us directly and we will talk to you personally to explain what is happening behind the scenes.

  

TIMELINE

  

As you can see, the good news is that we're almost 70% of our way there! We need your support for the final 30% push. Here are our updated timelines: 

 

 

TEAM

 

To kick-off the good news for this update - we wanted to announce two new additions to the Circadia Sleep Team. We are pleased to welcome Dr Daniel Blum and Dr Sara Nowakowski to the Circadia family, both practising and Behavioral Sleep Medicine certified, clinical sleep psychologists and experts in developing programmes based on cognitive behavioural therapy for insomnia (CBTi).

 

Daniel and Sara, join our extremely experienced and world-renowned sleep team which consists of:

  

Below are links to the bios of each individual: 

 

DIGITAL THERAPIES

 

Why you may ask?

So here’s the first BIG news for all Kickstarter backers!

We expanded the therapies that we offer in our mobile app to include clinically tested evidence-based sleep strategies such as cognitive behavioural therapy (CBTi), mindfulness and meditation (MBTi), and circadian exercises, all personalised to you, based on your sleep data.

From February to March, we did extensive user research. We surveyed over 3,000 people from around the world, conducted over 75 user interviews and 25 expert interviews.

What we learned from the exercise was that people need help with:

  1. Falling asleep faster
  2. Get deeper sleep
  3. Stay asleep longer
  4. Wake up refreshed
  5. Boost daytime energy
  6. Improve concentration

Now, the good news is light therapy can certainly help with 4 out of the 6 above but to make the entire experience more holistic and closed-loop, something we have always prioritised with our offering, the efficacy of our solution will be significantly bolstered by coupling the solution with cognitive behavioral therapy for insomnia (CBTi), mindfulness, and meditation.

During interviews, people constantly reported problems that fall into one or more of the following categories:

  1. Circadian (irregular sleep-wake schedules & morning grogginess)
  2. Onset insomnia (difficulty falling asleep)
  3. Maintenance insomnia (difficulty staying asleep, waking up in the middle of the night)
  4. Environment problems (sound, temperature, and comfort)
  5. Stress (negative thoughts)
  6. Anxiety (work rumination and negative thoughts)


This will enable our solution to be more appealing to a much wider group of people and ensure that we can help you with specific target areas with your sleep that you are not happy about. 

So what are sleep improvement tools you are offering?

Stimulus Control

This is the "standard of care" for those users who wake up in the middle of the night and cannot easily fall back asleep.  The basic idea is that these users often learn to associate the bed and bedroom with alertness and being awake.  Stimulus control helps teach them to associate the bed and bedroom with easily and quickly falling asleep.

Sleep Hygiene 

This is a set of behavioral "do's and don'ts" that promote a healthy sleep/wake cycle.  It is an educational-based treatment approach that focuses on both individual and environmental factors.  Data shows that those with poor sleep habits also often have the best knowledge of sleep hygiene issues.  Thus, the goal is not just to increase knowledge of healthy sleep habits, but also to increase the application of that knowledge.  The data does not support the use of sleep hygiene instructions alone, although many clinicians use it as a routine adjunct to any other treatment they choose. 

Relaxation Techniques

These techniques are most useful with users who report high levels of arousal or anxiety during both the night and the day.  Helping the patient physically and mentally relax increases the likelihood of a swift sleep onset.  There are, of course, many types of relaxation techniques.  The choice of a specific technique depends on whether somatic or cognitive arousal is the target of intervention, clinician expertise, and to which technique a specific patient is likely to respond.  This treatment strategy requires sufficient practice on the part of the patient and, at times, extra sessions with the clinician, to ensure the technique is learned well.  Some relaxation techniques that have been tested for insomnia include:

  • Progressive Muscle Relaxation:  This technique appears to have the most evidence in favor of efficacy.  It focuses largely on somatic arousal.
  • Biofeedback: The evidence here isn't as strong as for PMR, but it has been shown as effective.  It focuses largely on somatic arousal.
  • Imagery Training:  Insufficient evidence exists to fully critique the efficacy of imagery training, though its use has been reported.  It focuses largely on cognitive arousal.
  • Others:  Other relaxation techniques have been advocated, but not well tested empirically (e.g., deep breathing alone, meditation, hypnosis).

Sleep Restriction

This is a strategy used as an adjunct to stimulus control with users who spend much more time in bed than time actually asleep.  Sleep quality (and often quantity) is best when sleep is obtained in one large chunk rather than several small fits and spurts.  Conversely, increased time in bed leads to fragmented and shallow sleep. Thus, we want users in bed only when they are asleep - and asleep only when they are in bed.

Cognitive Behavioral Therapy

CBT is very efficacious and effective. Controlled studies and meta-analyses have shown that CBT takes longer to show initial efficacy than pharmacotherapy, but that by the end of a treatment period, the effects are similar with the two approaches. Furthermore, CBT has longer lasting effects, since users have been taught a skill upon which they can rely in the future. The "B" part of CBT is all the thing mentioned above. At a minimum, it would include stimulus control. The "C" part of CBT for insomnia focuses on unrealistic expectations for sleep, the results of poor sleep, the causes of insomnia, and/or performance anxiety surrounding sleep.  Typical thought challenging techniques are used as well as education.

Circadian Rhythm related Insomnia

Some users presenting with complaints of insomnia will actually have a circadian rhythm disorder.  These disorders include delayed sleep phase syndrome, jet lag, and shift work type.  Since the fundamental problem here is not insomnia but rather a dysfunction of the circadian clock, different treatments are utilized.  The most common therapies include bright light therapy, chronotherapy, and more recently, melatonin.  The insomnia treatments detailed above are sometimes combined with these more specific therapies, depending on the individual aetiology for the problem. 

We believe in going the extra mile. There are many gimmicky sleep products out there. So our unparalleled Sleep Team with the addition of Sara and Daniel is stronger than ever before to ensure we remain focused and scientific in our approach.

 

Here are a few visuals from our mobile app user interface to explain how we’re organising the therapies:

 

 

We’re not going to dive into the specific for the time being but as a little teaser, we’ve recruited the help of these 6 characters to be a part of our sleep exercises. We’d welcome your feedback, so if you have any comments please feel free to tell us in the comments section of this update below.

 

  

INDUSTRIAL DESIGN

 

Now moving on to the design of our physical product we’ve got some great visuals to share with you! This will give you a better sense of the final look and feel of the product that we intend on delivering.


 

 

 

 

MECHANICAL DESIGN

 

Now to the meat of our development update. Below you can see some of the issues we are facing with regards to achieving even light diffusion. 

 

  After consulting with lighting specialists from Germany, we were recommended to have three injection points in our tool for the front diffuser. 

 

 

 Further research revealed the assembly stack required to achieve even light dissipation, also explained by the sketch of concept 1 on the board below. 

  

 

  

FIRMWARE & ALGORITHM DEVELOPMENT

 

Radar Streaming

One of the first steps to our software infrastructure that we have nailed is to ensure seamless transmission of our radar sensor data over WiFi. This specific task took us 4 weeks but is absolutely crucial as it streams a large amount of data (8 hours of sleep data) completely wirelessly and without the user having to go in to the app and physically click sync. 

 


 

Digital Signal Processing

The next step is to take the raw sensor data and perform digital signal processing to ensure we can extract features such as respiration rate, respiration rate variability, and body movement (amongst 15 other variables). 

 Figure: Shows how we analyze sensor data in the time domain: We have identified the user range before, so we can extract the respiration signal and its features from the sensor data. The respiratory rate, as well as the intervals between individual respirations and a number of other features, vary significantly between different sleep stages. 

 

  
 

Figure: Shows an intermediate step in the signal processing of raw sensor data: Frequency analysis for different depths. We can observe a clear peak at a range of 1.3m, with a frequency of 15.3 cycles per minute. Because we know that respiration at rest usually varies between 10 and 25 respirations per minute, we can say with confidence that this is the user, located at 1.3m from the Circadia tracker. 

 

Using processed data from our signal processing algorithms we then analyze your sleep stages and circadian rhythm using advanced machine learning algorithms. The image below shows a typical graph of their sleep stats. Note, we have renamed REM to Dream to make it easier for people to understand. 

 

   

Circadia Sleep Index

From the very first day, one of our goals has always been to ensure people can understand their sleep data. So we've taken a lot of care and employed clinical practices to devise a completely new metric defining your quality of sleep. The Circadia Sleep Index is the first metric which factors in not just your quantitative data from the tracker but also uses self-report diary data from the sleep diary built-in to reveal how well you slept and what you can do sleep better. 

  

  

Circadia Insight

The final and most critical algorithm which connects our tracking with therapy is our magical Circadia Insight algorithm. Taking data from your sleep statistics, the Circadia Insight algorithm personalizes and automatically recommends the therapies most suited to help you sleep better. More to follow on this in our next update!

 

CLINICAL TESTING

 

At Circadia we have an incredible team of 6 PhDs and 2 Professors. One of the things we do not compromise on is ensuring we benchmark our device with the golden sleep lab standard, polysomnography. Most of all, unlike other sleep trackers, we are not afraid of sharing our data. In December, we conducted an extensive 24-night sleep study, putting our sleep tracker to the test and the results have by far exceeded our own expectations. Circadia has blown existing consumer-grade sleep trackers out of the park! We will be publishing a paper on this soon but for now here are a few teaser graphs to show you the performance: 

Body Vitals Tracking Accuracy

 

Figure: Respiratory rate from PSG data (yellow) and as predicted by the Circadia tracker, for two representative datasets. Overall accuracy was 93.16% (STD ± 1.82pp).

 

Machine Learning Sleep Staging

Figure: Hypnograms from PSG (the current golden standard) data (yellow) and as predicted by the Circadia tracker, for a representative dataset. As can be seen, Circadia predicts almost all major sleep events throughout the night.

 

MOBILE APP DEVELOPMENT

 

In our last update, we shared our mobile app blueprint which consisted of 167 screens. We have now developed all the UI screens and are in the process of deploying our algorithms and connecting our tracker to the app via WiFi.  

 

 

Here’s a video show our integration work with Fitbit and Alexa:

 

To-do:

  • Digital Signal Processing Deployment on Mobile - estimated time: 2 weeks
  • Final bits & bobs of our UI - estimated time: 2 weeks
  • Device connectivity - estimate time: 1 week
  • Digital therapies - estimated time: 4 weeks

So in 4 weeks from now, we will start bench testing our mobile app.

    

THANK YOU FOR READING & FUTURE UPDATES

  

Just as we wrap up this update, we would like to say a few words:

"We express our deepest apologies for being unresponsive and not fulfilling promises we have made. We can assure you that this is not a choice but a consequence of our position. We are a very small team and at times are particularly busy. Circadia has grown from a project in our garage to a company with 5 full-time employees, attracting significant partnership interest from corporates."

If you’d like to speak to us, we’re starting a new initiative where you can select a time to arrange for a callback. 

 

 

If you’ve moved houses or even countries, please hang tight for the next update with the survey, within the next 2 weeks. 

Thank you very much, we hoped you enjoyed the read and we will be in touch shortly with another update. 

Sleep tight!

-- Farés Siddiqui & Michal Maslik (Founders at Circadia)