NeuroSole

NeuroSole is a portable, cheap, non-invasive, and easy-to-use early diagnostic system for Parkinson’s disease. 

Parkinson's disease (PD) has the fastest-rising prevalence of all neurological diseases, affecting up to 10 million individuals. In addition, in 2022 nearly 90,000 patients have been diagnosed with Parkinson’s, a steep 50% increase from preceding years, making it the second most common neurodegenerative disease in the United States. 

For diagnosis, patients have to go through numerous exhaustive, expensive, and time-consuming tests such as DaTscan, MRI, and PET scans to rule out other diseases. Patients also spend an average of $2,500 to obtain the necessary examinations before diagnosis. The symptoms of Parkinson’s are also similar to that of Lewy body dementia, and muscular dystrophy, making it even harder for doctors to diagnose Parkinson’s. 

Currently, Parkinson's is misdiagnosed up to 30% of the time and according to Parkinson’s UK 34% of people that were misdiagnosed reported a decline in health. This leads many patients to develop Parkinson's symptoms 1-2 years before their official diagnosis, leading to worsened conditions and chronic symptoms. Moreover, according to ParkinsonsDisease.net, “diagnosing PD early is important because treatments such as levodopa/carbidopa are more effective when administered early on in the disease.” Thus, the increase in  late or misdiagnosis leads to ineffectiveness of medications and ongoing progression of the disease. 

Altogether, due to a lack of an accessible, affordable, and accurate diagnostic system available for patients in the early stages prevents them from taking further action in order to slow the progression of the disease. 

NeuroSole is a novel diagnostic tool that utilizes a 1D convolutional neural networks model to analyze Parkinsonian gait patterns, which are one of the earliest symptoms in Parkinson’s. To obtain data points regarding one's gait, each pair of NeuroSole insoles are embedded with 12 pressure sensors to transmit the ground reaction forces to the 1D convolutional neural network model. The sensors placed in the insoles of the shoe record the pressure points during each phase of the gait cycle. To accurately assess the patient's gait, six sensors are positioned in each insole: one sensor near the toes, two in the metatarsal region, one in the midfoot, and two near the heel. The data points collected by the sensors are sent from the Arduino Uno to the Raspberry Pi which then sends the data to the cloud where the data is processed and used to detect the likelihood of developing Parkinson’s using the 1D-Convnet model.   

NeuroSole comes with the insoles and a free downloadable app. The app displays the results of the examination, allows users to monitor symptoms, and share results with caregivers. To use NeuroSole, the user has to download the app on their mobile, insert the insoles into their shoes and enable the app to start tracking. After a short walk of 2-3 minutes, the app immediately reports the likelihood of developing PD and if desired the results can be shared with caregivers. Patients can use our product on a regular basis to monitor the progression.  

To develop the model the 1D CNN algorithm was used as it can perform activity recognition tasks such as a person is standing, walking, or jumping. I also used software applications like the Raspberry Pi OS and the Arduino IDE to read in the sensor values and develop the model. 

The demographics of our primary target market encompass individuals who are 50+ across the globe. The psychographics are individuals looking for immediate, cheap, and non-invasive alternatives to diagnose Parkinson’s in its early stages. This also includes individuals who are cautious about their health, have poor health records, or could possibly inherit the disease from close family members. The secondary target market includes disorder movement specialists and neurologists who want to verify the diagnosis results and monitor their patient's symptoms and progression over long gaps between visits. Our very first customers include patients at local hospitals, neurologist's clinics, and the elderly individuals in our local community. 

Currently our target market doesn’t have other alternatives on the market to accurately diagnose Parkinson’s in its early stages in the comfort of the homes. Our product aims to provide an effective solution where users can diagnose in the comfort of the homes while intervening at the proper time to make treatments such as Carbidopa-levodopa more effective at managing symptoms. Through our product, we hope to increase the effectiveness of medication and decrease the number of patients that are misdiagnosed in the coming years. In the long run, NeuroSole can provide individuals with peace of mind and allow them to lead active and independent lives. 

My twin sister and I have decided to work together on creating a life-changing product that draws from our shared life experiences and knowledge. This problem matters to us because our grandfather was diagnosed with stage 3 Parkinson's, which made us recognize the invaluable need for tools that can help patients to diagnose neurological diseases like Parkinson’s in its early stages. Our experience motivates us to create a product that will assist millions of patients worldwide to take necessary medications at the appropriate time. Our product will aid neurologists by streamlining the diagnostic process and simultaneously helping patients take preventative care as needed. In addition, we have a strong background in computer science as we both advanced to the finals in ACSL, and are currently silver contestants in USACO. Moreover, we have strong fundamentals in business and have been competitively successful in DECA by placing in ICDC, which will allow us to build a strong business plan for our company. In addition, in our first two years of high school we performed extensive research on neurodegenerative disease similar to Parkinson’s and received statistically significant results and even published the results in the Harvard Journal of Emerging Investigators. We have also been recognized for our research on neurodegenerative disorders by the city council and mayor. All of this will allow us to develop a one-of-a-kind product that caters to our customers by solving the current drawbacks in the market. 

To test our product we reached out to a local neurologist clinic and podiatrist in Cupertino for feedback on the prototype's uses, comfort, and functionality. After the session, we refined our product by adjusting the specific orientation and placement of the sensors in the insole. This rearrangement helped gather a wider range of values and covered all the pressure points of the foot. 

In addition, we sent a feedback survey to the elderly members of the Cupertino Senior Center to gather their responses to our product. We provided the price, functionality, product demo, and design of NeuroSole. In total, we received over 50+ responses. The members indicated how NeuroSole would be greatly beneficial as they would be able to easily identify changes in their gait and use it whenever they wanted. Members also suggested displaying the output in words instead of displaying a percentage as it would be easier to comprehend the result. Taking this into consideration, we modified the app to report words like normal, mild, moderate, severe, and very severe. 

NeuroSole is a portable, cheap, non-invasive, and easy-to-use early diagnostic system for Parkinson’s, and is currently the only diagnostic tool on the market to utilize vertical ground reaction forces from insoles to analyze Parkinsonian gait. NeuroSole stands out because of three key values which are accessibility, affordability, and accuracy. NeuroSole can transform the diagnostic market by opening options for customers to perform at-home diagnosis with accurate results within minutes and also enable doctors to utilize NeuroSole in their diagnostic process.

In addition, other products on the market like the Parkinson’s Toolkit app provide medical professionals with clinical practice guides to treat PD, but they do not provide quantitative real-time data to diagnose Parkinson’s like NeuroSole. Other diagnostic options like brain scans and neurological tests can be harmful, expensive, and time-consuming. In contrast, our product is non-invasive and can display results within minutes of walking. In addition, our product has myriad features to set us apart from emerging competitors such as reporting results to caregivers and regularly monitoring their symptoms. Altogether, NeuroSole provides our customers with a one-of-a-kind product that is accurate, affordable, and easy-to-use. Neurosole is a product that once placed in the hands of patients and neurologists can revolutionize the industry of diagnostics for parkisons.

At NeuroSole, our goal is to reduce the number of patients who are misdiagnosed for Parkinson's and help neurologists diagnose Parkinson's at the early stages in a non-invasive, quick, and simple method. In the coming years we hope to decrease the number of patients who are misdiagnosed by 5% to 10%. Our product aims to increase the average lifetime of Parkinson's disease patients by intervening at the proper time and making treatments such as Carbidopa-levodopa more effective at managing symptoms. According to ParkinsonsDisease.net, “diagnosing PD early is important because treatments such as levodopa/carbidopa are more effective when administered early on in the disease.” Through our products we hope to instill hope and confidence in patients and allow them to lead active and independent lives, by addressing  the issues at an early stage and continuously monitor their progression.

NeuroSole is a novel diagnostic tool that finds abnormalities in the walk by utilizing gait patterns, one of the earliest symptoms in Parkinson’s. To develop the hardware of our product, we embedded 12 pressure sensors in a pair of insoles to transmit the ground reaction forces to the 1D convolutional neural network model. The sensors record the pressure points during each phase of the gait cycle. To accurately assess the patient's gait, six sensors are positioned in each insole: one sensor near the toes, two in the metatarsal region, one in the midfoot, and two near the heel. The data points collected by the sensors are sent to the Arduino Uno which serves as an analog-to-digital converter. The Arduino Uno digitizes the sensor signal and transmits it via a serial protocol to the Raspberry Pi. The Raspberry Pi then sends the data to the cloud where the data is processed and used to detect the likelihood of developing Parkinson’s using the 1D-Convnet model.   

In order to develop the AI model, the Gait in Parkinson's Disease dataset was obtained from PhysioNet. The dataset is a collection of multichannel recordings from force sensors beneath the feet of 93 patients with Parkinson's Disease and 73 healthy controls. The dataset contains the vertical component of the ground reaction forces from the sensors placed on the insole of each foot. For the data pre-processing, the first and last 15 seconds of the walking were removed to ignore the starting and stopping disturbances. Furthermore, for the feature extraction, the most prominent features were analyzed such as swing time, stride time, gait asymmetry, and double stance time for both feet. In order to build and train the model we utilized the 1D convolutional neural networks algorithm. For the hidden layers in the model, a scaled exponential linear unit activation function was utilized, and a sigmoid activation was used for the final layer. The algorithm was trained using Adam optimizer with a batch size of 850.  

The AI model reports the likelihood of developing Parkinson’s on the NeuroSole app after the user walks for a few minutes. The app is currently being developed where the users can automatically share results to caregivers and monitor their past test results. 

Next year, our first year of launching into the market, we hope to serve an average of 800 users which will be a combination of patients and doctors, neurologists, physiotherapists who want to utilize NeuroSole in the diagnostic and recovery process. The customers will utilize the diagnostic tool, the insoles, and the free downloadable app to monitor, share and record their results. We hope to provide these individuals with an early diagnosis allowing them to obtain necessary medications to prevent worsening of conditions. 

A barrier that we might face next year is the relatively high cost when compared to the revenue, because of initial start-up costs and fixed costs such as machinery, workspace and more. However, in the coming years the costs will gradually decrease with mass production. Another problem that might initially exist is educating the customers through marketing on the multitude of benefits one can receive and all the functionalities of our product. 

We plan on partnering with hospitals in order to connect with physicians who can recommend NeuroSole to their patients. Developing relationships with medical centers, and gaining the endorsement of medical professionals will significantly help to improve our brand image and credibility. In the future, we aim on partnering with the Valley Parkinson’s Clinic and well-known orthotic brands  like OrthoFeet and Ossur to gain more customer attraction. In addition, we hope to partner with Polaris and Aberdare venture capitalists in the healthcare sector to gain further funding, expand our reach, and gain insightful feedback.

Our business plans to effectively utilize a mix of direct-to-consumer models through our eCommerce website and partner with retail stores to attract the most number of customers. The customers can purchase the insoles and download the app for free. 

Growth Model: Phase 1 - Product Development & Refinement: In this phase, we will focus on prototyping and refining our product with the feedback we will receive from members in the California Parkinson Foundation, and local neurologists. We will also work on optimizing the website for SEO and integrating calls to action. 

Phase 2 - Marketing & Brand Awareness: Initially, to educate and attract customers we plan to utilize social media platforms with our focus on impression marketing. Throughout the initial stages, we hope to create flyers to post outside of clinics and connect with physicians who can recommend NeuroSole to their patients. Gaining the endorsement of medical professionals will significantly help to improve our brand image and increase credibility.  

Phase 3 - Scaling & Market Expansion: In the coming years we look forward to expanding our market to neurodegenerative diseases similar to PD like Amyotrophic Lateral Sclerosis, Charcot-Marie-Tooth, and Huntington’s disease. One of the most prominent symptoms of these diseases is also gait, allowing us to utilize the same technology of inbuilt sensors in insoles to detect symptoms. This will drastically increase our customer base, and make our products more inclusive to all neurodegenerative diseases.

Our four main sources of funding include angel investors, SBIR and STTR funds, venture capitalists, and crowdfunding. To procure the initial funds, we plan to launch a crowdfunding campaign with a goal of $30,000. SBIR and STTR grants are a great fit as they will not dilute our capital and will allow us to gain guidance from experts. In addition, we plan to partner with Polaris and Aberdare venture capitalists in the healthcare sector to gain further funding, expand our reach, and gain insightful feedback.