Kinisi: The Moonshot Company Restoring Mobility to the Paralyzed
Running, brushing your teeth, giving a high-five, hugging your friends.
We don’t think about these movements, movements that enrich the human experience and make life feel effortless. Yet, we take these graceful movements for granted.
Imagine if your body was paralyzed — your arms immobile blocks that hung stiffly at your sides, your legs rigid and permanently fixed in a wheelchair.
In the United States alone, this is the reality for nearly 1 in 50 people who are paralyzed. Often the victims of car accidents, the spinal cord injuries (SCIs) they sustain will result in either upper body, lower body, or full-body paralysis.
Afterwards, life is completely transformed as the victims must relearn how to perform the most basic, essential functions without the freedom of movement, which the SCI ripped away.
It’s not just autonomy but lifespan that is lost. For victims who suffer spinal cord injuries, life expectancy post-injury has not improved within the past thirty years. For ventilator-dependent patients aged 60 and above, the SCI will slash life expectancy to just 1.5 after the injury. Younger patients, such as those 20 years or old with preserved motor function, can expect to live a longer 52.6 years.
Yet, in addition to the quantity of life decreasing exponentially, quality does as well. 94% of paralyzed patients spend the remainder of their life in chronic pain. Add in an increased risk of cardiovascular disease, pneumonia, and life-threatening blood clots, and it becomes visibly apparent why paralysis is considered one of the worst possible outcomes from a car accident.
Status Quo
But surely there’s hope, right? We’d like to think so. However, despite dramatic advancements in medical technology within recent decades, the outlook remains bleak for paralyzed patients.
According to a report on the Systemic Complications of Spinal Cord Injury, it costs between $320,000 and $985,000 to treat a spinal cord injury patient the first year and as much as $5 million during the patient’s lifetime.
Even though recovery is optimistically pushed as being “within reach,” the devastating physical, financial and mental impact of a spinal cord injury leaves many victims with a host of complicated health issues and even PTSD. Furthermore, the most common mobility aid — a wheelchair — exacerbates the risk of cardiovascular diseases and blood clots due to its sedentary nature.
With the alarming increase in the prevalence of car accidents, spinal cord injuries and thereby paralysis, it’s imperative that we find innovative solutions to restore mobility and life for paralyzed patients.
Wheelchairs are currently the most prevalent mobility aid, but Kinisi aims to create a healthier alternative.
Our Solution
Kinisi is harnessing the power of machine learning and brain-computer interfaces to restore mobility in paralyzed patients.
Our moonshot company focuses on using a combination of a Brain-Computer Interface (BCI), a Machine Learning model using a Bayesian classifier, and electrode pairs to enable paralyzed patients to move again.
Our team will utilize a Strenthrode BCI implanted in the frontal-medial area of the brain, where it will pick up Electroencephalogram (EEG) waves. These inputs are then passed through a Machine Learning model to determine if the brain activity signified walking or idle movement. The Machine Learning model will then communicate this output to the electrode pairs implanted on the legs to the patient to determine if a step will be taken.
Going In-Depth
For our solution to work, the patient must undergo medical procedures to implant the Strenthrode BCI in their brain and the electrode pairs in their legs. The Strenthrode BCI will then pick up brain activity when the patient attempts to walk. This data is then passed onto the app via the Bluetooth communication protocol, where the Machine Learning will process the data.
The Machine Learning model is a Bayesian classifier that will compare the probabilities of the idling and walking classes to determine if the patient will stay still or move. The Bayesian classifier will take in brain activity data collected from the Strenthrode BCI through a Bluetooth connection. The probabilities will be found through the Bayes rule:
Bayes’ rule essentially asks, “How often will the user want to remain idle?”. Using past data on how many times the user’s movements remained stationary, probabilistic calculations will allow a class to be predicted and sent as an output. The Bayesian classifier uses Bayes’ rules to determine if the brain signals result in idle or walking movements. It then picks the class that happens most often.
After the proper class is chosen and communicated to the receiver, the corresponding electrode pairs will be activated. Specifically, electrode pairs will be placed bilaterally over the femoral (immediately proximal to the knee) and the deep peroneal (immediately distal to the knee) nerve. To allow the patient to stand, simultaneous bilateral activation of the quadriceps will be used to maintain the required knee extension.
The Bluetooth connection will wirelessly facilitate all communication between the Strenthrode BCI, the mobile app, and the implanted electrode pairs wirelessly through a Bluetooth connection. Starting from the Strenthrode BCI, the data will be wirelessly sent to the mobile app where the data is processed through the Machine Learning model, where the output is then communicated to the implanted electrode pairs.
Any issues concerning the facets of our solution will be detected and reported through the app. The app’s backend will be directly connected to Kinisi’s lab technicians, where the app will send recommendations for our solution to the user, such as any mandatory visits to Kinisi lab technicians.
Implementation
Kinisi will be launched within a three-year timeline starting Q1 2022, with each phase consisting of a year, separated by pre-launch, pilot testing, and post-launch of the product. Product research and development and advertising and recruitment will be completed by Q1 2023, ready for pilot testing. By Q1 2024, Kinisi will recruit over 500 people with paralysis in North America to gather sample data, refine the Machine Learning model, and modify the sensors on an individual basis to tailor the customers’ needs. By Q1 2025, the product will be ready to launch in over 50 countries, with a starting price of $5,000 US dollars.
With this price tag comes benefits — for example, customers will be able to file the purchase of product into their insurance policies; often, customers with insurances would only have to pay for less than half of the price. These flexible policies allow customers to buy the product with zero hassle, making the product as accessible as possible.
Vision
Looking into the future, there are numerous opportunities for experimentation. With our FES-BCI and Machine Learning model, we can recondition the patient’s muscles to enable ambulation abilities. The first experimentation revolves around experimenting with the Bayesian model instead of the Support Vector Machine (SVM). Essentially, the difference between the two is that the Bayesian method focuses on a statistical approach with an emphasis on probabilistic calculations. SVMs are particular linear classifiers, focused on surpassing a particular numerical threshold, based on the margin maximization principle, which separates two different classes with a line while leaving the maximum margin.
Our vision for future experimentation dives into the specific location of the BCI. Our team’s current plan is to utilize the Strenthrode BCI, implanted in the front medial area of the brain, to pick up Electroencephalogram (EEG) waves. Certain regions of the brain have specific impacts our experiment’s outcome, such as the medulla, located at the base of your brain, where the brainstem connects the brain to your spinal cord. This area is the key centre for deploying messages from your brain to the spinal cord. Additionally, this portion of the brain regulates the cardiovascular + respiratory systems.
Going beyond the field of treatment, we see Kinisi looking into the diagnostics of SCIs. For instance, in Alzheimer’s Disease (AD), patients who have lost the ability to communicate through verbal communication can benefit from BCIs that allow them to convey binary thoughts (yes/no) in addition to emotions. Expanding into the more significant problem of SCIs with a combination of cutting-edge technology, we believe we can enable patients to regain their independence and life back.
