Wave Care

My grandpa calls mental illness a western hoax. 

This perception is precisely why I have been forced to watch my mom struggle and be afraid to seek help for years. The same is true for millions in my South Asian community: mental illness isn't considered real. 

With a misdiagnosis rate of 65%, it’s no surprise that underserved populations, which already have fragile trust with healthcare systems, have a hard time believing depression is a real problem. The lack of innovation towards mental health diagnostics only further deters people from seeking help, and this presents a vicious cycle in the psychiatric industry. 

For the majority of cases, depression is inaccurately diagnosed and this leads to over 182 million people around the world not receiving the help they need and others being prescribed antidepressant medication they don’t require.

Current depression diagnostic procedures rely solely on the subjective discretion of doctors, where personal experiences heavily influence decisions. The DSM-5, the core tool in this process, was last revised 10 years ago and has not been updated to meet the current mental health needs. 

Change is needed to actualize holistic depression diagnostics, and Wave Care strives to do exactly this through implementing a layer of objectivity. 

I have been experimenting with BCI kits for the past two years, and one experiment I did in particular took my interest. I was visually analyzing the brainwaves of my friend when she watched a sad video, versus a happy one. I recognized there were indicators of a pattern and wondered if it would persist when these emotions are extrapolated over a long period of time, as depression is. This trivial experiment set me off on my journey building Wave Care. 

Wave Care leverages AI and BCIs to analyze patterns in brainwave (EEG) data to deliver depression susceptibility. For more technical detail, the algorithm uses a convolutional neural network to differentiate between thousands of EEG image data frames, and compiles this to determine whether the patient likely is clinically depressed or not. The algorithm is distributed through application programming interface (API) licensing with hospitals and clinics. This way, hospital systems are able to easily integrate the algorithm with their existing EEG technology using the purchased Wave Care digital key. 

Initially, Wave Care was developed to replace the DSM-5; however as user research was conducted, the product was shaped into a diagnostic aid to ensure a holistic process.

Wave Care helps individuals struggling with mental health from stigmatized communities as it decreases the misdiagnosis rate so they can receive accurate medication and therapy. In turn, these people will not be victims of incorrect prescriptions and the risk of overdose. Once people witness actual positive change from correct treatment, trust will build between these communities and mental health care. Overtime, Wave Care will reduce misconceptions for depression and more people suffering in silence will have the courage to reach out and ask for help, breaking the psychiatric vicious cycle.

The problem Wave Care is solving has been validated countless times in 36+ research interviews conducted with psychiatrists working in the diagnostic field, individuals struggling with mental health, and members from communities where mental health is heavily stigmatized. 

Key quotes from interviews: “The biggest problem with mental health is that there is no scan, test or blood work. We don’t have anything for mental health.”

“If a doctor can show me data about depression, I’ll feel more comfortable accepting it.”

I have also read dozens of research papers detailing the implementation of EEG-based solutions for diagnosing PTSD, and iterated on my own design with Wave Care based on my findings. I published an initial research proposal online and sent it to mentors working in healthtech from Google’s Moonshot Factory and The Neuro, where I was given feedback on my algorithm’s approach. 

I also spoke directly with higher ups in hospital systems such as SickKids to validate the idea’s business model. I learned that EEG technology is accessible for most hospitals and clinics, which aided me in solidifying a software-based idea. 

Revisions are in the works for the DSM-5, the depression checklist of symptoms for diagnosis; however, each of the factors remain qualitative. Wave Care differs as it provides a quantitative lens to depression which can be used in conjunction with the DSM-5. AI isn't currently being used in health for mental health conditions specifically. The industry is currently satisfied and complacent with the status quo, despite the numbers being extremely negative. 

Thymia, a UK-based company, is also approaching mental health from a data-driven perspective by collecting insights from games for children to diagnose anxiety. It is offered directly to consumers and this is a key differentiator for Wave Care as we approach hospitals, which receive all kinds of patients so our procedure can be recommended from a credible source. 

Wave Care's disruption in the AI diagnostics and mental health space will create a ripple effect and expand to other conditions including PTSD, bi=polar disorder, epilepsy and many more. It will mark the start of data-driven diagnostics being carried over to the brain as well.

Mental health and depression transcends all ages, and impacts people from all walks of life. Through integrating Wave Care as an API in hospitals, the solution addresses the UN SDG #3 by positioning itself as an accessible aid. 

Every human being deserves care. However, mental health care is often neglected and this is largely due to stigma, where conditions including depression and PTSD aren’t considered real. I am innovating for good because I’m breaking the stigma to ensure more people receive accurate care. Additionally, Wave Care is completely non-invasive and will not inflict any physical or mental harm on patients. For hospitals as well, Wave Care reduces the time required for depression diagnosis and saves millions of dollars. I hope to build trust between healthcare, mental health and local communities through relieving a huge concern: depression’s “realness.”

Wave Care is currently in the implementation stage, while testing perpetually iterates from the tech standpoint. 

I pitched Wave Care’s idea and design at the Shad National Design Challenge where I won and have taken the project forward to now collaborate with research institutes for implementation.

To date, Wave Care has been tested on a dataset of 1000+ patients and boasts an 89.5% accuracy, which is 24.5% higher than the industry’s status quo. Currently, the API infrastructure is being refined so Wave Care can be implemented through deals with hospitals.

As the innovation is brought to real-world implementation, a few key collaborators will include Sick Kids hospital, the Toronto General Hospital as well as Columbia University. With the two hospitals, I will be exploring the correct onboarding processes and training for integrating the API. Alongside Columbia University’s EEG lab, I hope to continue researching and testing the algorithm to ensure it is continuously being iterated and improved. 

Through these collaborators, I plan to acquire relevant data by working with psychiatrists and having them do EEG data collection with willing patients as voluntary participation in a research study. 

Healthcare systems are very legacy-based and adverse to change, so introducing new technology to diagnostics might be a lengthy onboarding process for many hospitals. The fact that there have been no changes to the DSM-5 in the last decade is only a testament to this challenge. To combat this, I propose beginning implementation with private clinics to demonstrate a proof of model to larger, public hospital systems. I also plan to invest in education and training for the technology as many of the concepts may be foreign to physicians that have not explored emerging tech before. 

Of course, there are always ethical and security concerns in regards to health data. After consulting with 10+ hospital representatives and psychiatrists, I learned that this EEG data would be kept equally private to the patient as the DSM-5 data. EEG data isn't conclusive for depression diagnostics, thus why Wave Care is positioned as an aid in order to make a clinician's diagnostic procedure more holistic.

In the next 5 years, I hope for EEG testing and Wave Care's algorithm to be standard practice in partnered hospitals. In these hospitals, I'm aiming for the misdiagnosis rate to be reduced from 65% to 10%. My facilitating bi-annual re-diagnosing procedures, we will be able to validate the accuracy of Wave Care and its impact. 

From a non-technical standpoint, I hope to spark technological development across other conditions as well. With a proof of model and concept from depression, I hope to expand to epilepsy and/or PTSD.

1 full-time staff

2 contractors

2 years

Mental health is extremely environmental, thus why data diversity is extremely crucial. This is why we have been extremely intentional in our data collection as well as user research, speaking with indigenous doctors, patients from all sorts of backgrounds and getting in contact with hospitals ranging from public to private.

- Service contracts to governments (public hospitals)

- Selling products (API tokens) to private hospitals & sole-clinicians 

Current

- R&D: $2000

- Data Acquisition: $1000

Next Year:

- Human capital: $50,000 (bringing on a full-time co-founder + more contractors)

- R&D: $10,000 (in-person, hospital pilots)

- Development: $40,000 (full deployment)

I am seeking $100,000 in order to meet my goals for impact and deployment next year. 

- Human capital: $50,000 (bringing on a full-time co-founder + more contractors)

- R&D: $10,000 (in-person, hospital pilots)

- Development: $40,000 (full deployment)

I am seeking $100,000 in order to meet my goals for impact and deployment next year. 

- Human capital: $50,000 (bringing on a full-time co-founder + more contractors)

- R&D: $10,000 (in-person, hospital pilots --> this also include data acquisition)

- Development: $40,000 (full deployment)