Proposal: A platform that scales brain data collection for neurotech ML

Proposals Round 7
#1

Name of Project: Mindland

Proposal in one sentence: Build an app incentivizing the contribution of HRV/Brain data to train AI models.

Description of the project and what problem it is solving:

Our mission is to make a system where people will upload their valuable biometric data such as brain data on chain and monetize them on their will. We will use tokens and gamification to scale the collection of these data for accelerating neuroscience and ML research. We start off with Meditate-and-Earn, where people are incentivized to upload their brainwave and HRV data during meditation to receive a fair share of the economic value created there, but we will apply this mechanisms to collect brain data of other tasks too.

Phase 1: Meditate-and-Earn

  • Incentivize users to practice meditation and help them make it into a habit with crypto and NFT rewards, as well as gamification. To track users’ meditative state, we monitor brain waves using BCI or HRV using fitness tracker / smartphone

Phase 2: Evidence-based meditation

  • Using users’ brainwave / heart rate data during meditation collected in Meditate-and-Earn, the AI in Mindland will be able to make suggestions on how to improve users’ meditation. The subscription revenue from this content will be distributed to users as well.

Phase 3: Brainwave Hub

  • Become a hub for brain data not limited to meditation where data providers will be rewarded too while maintaining the ownership of their brain data. The biggest possible use for this hub is neuromarketing, which is a marketing tactic that uses brainwave data. The market for neuromarketing is estimated to reach $21 billion by 2030.

We have 1.4K+ discord members and ran our closed alpha test with 150 people chosen from the community. The BCI model we are using is the Muse headband, and 20 people out of 150 testers having already bought one to test the closed alpha.

15 Relive
15 Relive1920Ă—1080 142 KB

Grant Deliverables:

  • Grant Deliverable 1: Launch our beta-test with three part-time developers and one UI/UX designer.
  • Grant Deliverable 2: Reach 5k total beta test users

Spread the Love

We are looking for a data scientist who previously worked with EEG data.

Squad

Squad Lead:

  • Twitter handle: @nyneurotech
  • discord handle: Nao | Mindland#7618
  • USDC ETH mainnet wallet address for potential funds
    • 0x8e5c585e373A743DcdD6deDaD3f803dAb0d4B7D8

Squad members:

  • Twitter handle: @royang_c
  • discord handle: reian#4434
2 Likes
#2

Who is going to pay for the “brain data” ?

Where will the monetary incentives come from? sounds like a recipe for insolvency.

I think thousands of years of meditation have it already down to a science.

Estimated by whom? And what information can you gleam from an EEG, which might affect purchasing decisions?

This sounds incredibly … hairbrained, at least openBCI i can understand, because they are reading motor signals and general arousal, and I can see it being good for VR.

#3

→ Various companies and research institutions. The most obvious use case is neuromarketing, which already has more than $2B market size and big companies like Pepsi Co, PayPal, and Hyundai are interested in.

Another type of brain data consumer is content creators, and this can be us or any other company. With brain data collected from users, content creators could create content such as personalized evidence-based meditation analysis or personalized music that enhances focus. Considering the amount of money people spend on mental health or productivity content, the market is pretty big.

→ Ads, revenue from subscription content created with the data collected from users, partnerships with above mentioned companies / research institutions, partnerships with insurance companies and so on. I kind of get that these sources of money are difficult to get from the very beginning, but that is exactly why we are using crypto. Some people will play our app expecting future returns, and we saw that in our alpha testing.

→ You may want to look at existing meditation apps and research papers. If you study more, it will become obvious that those results are only based on a limited number of subjects and bigger data is needed to train ML / DL models and make more detailed analysis.

→ This was an estimate by Straits. You may also want to read papers like this (Recognition of Consumer Preference by Analysis and Classification EEG Signals - PMC) to learn about what can be read from EEG. I picked an easy one to read. It may sound hairbrained having a limited amount of knowledge, but I think it is pretty reasonable to bet on this field considering the upside.

#4

Love this.

We are looking for a data scientist who previously worked with EEG data.

There are a number of people in the community who have experience with this, we ran a BCI hackathon earlier this year with LYNX with an EEG dataset. Perhaps we can also chat about doing sth similar? I think this could be a great way to interface between our two communities.

Question from me would be where is the EEG data stored? How do you handle ownership & data anonymization? There are some cool Web3 solutions for that like Ocean Protocol, Streamr, etc. (and privacy-preserving software from OpenMined) Also, are you planning to open source your AI model?

#5

What I struggle to figure out, is what is the utility to the user, from their perspective.

The value add is, you pay us for a device, which will tell you to breathe deep breaths, and will psychically probe your emotional state, so that we can sell you stuff better, and in exchange we will pay you for wearing it?

I just want you to think about the optics for a second. “it felt like they were controlling us”

#6

We agree this wouuld be a great idea and would def be interested to chat about it :slight_smile:

Ocean Protocol would be one of the solutions. We are also looking into IoTex MachineFi as well, but we are still in the process of deciding and talking with different teams about this.

#8

Hey, this is a cool idea. I had some thoughts/questions:

There could potentially be some very sensitive information in the brain data people will be putting on-chain (e.g. early indicators of disease or predisposition/risk). How will you manage this? Will it be anon and encrypted? Even if anon, could it still be possible to trace back to individuals in some cases (e.g. through wallet data)

With “x-to-earn”, how will you prevent people spoofing data or lying? e.g. just leaving a device on and logging it as meditation.

What model will you use to tell meditation quality? Are you using the Muse metrics, or building your own?

1 Like
#9

You might want to also/instead think about attention-to-earn (BAT is proof of concept there’s value in this). Could even develop something and potentially sell to BAT. “Attention” is one of the easiest and most reliable measures with BCIs

#10

Thank you for the comment!
How we are managing the data → We’re looking at various possibilities such as Ocean Protocol’s compute-to-data. The easiest option would be to make them anon with zkp because we still need some demographics.
To prevent people spoofing data, we are looking at impedance of each electrodes and able to tell if the headset is on or not. Also, people earn more tokens if they get higher scores, and we feel that this encourages them to play the game rather than do other things.
We are looking at theta waves, which is often used as a good indicator of meditative states, and we are also planning to brush up the algorithm using the data collected.

We really love the idea of attention-to-earn. We were just discussing it with our advisors the other day, and we’re definitely planning on looking into it more deeply.