Moose powered Brain Mapping Demo
(i.e., a data analytics demo for your brain)
This is a demo of a data intensive Brain Mapping application that uses the Moose and the Muse Headband products.
Moose is an application / infrastructure tool that simplifies the creation of data intensive applications. It's a product that a few of us over at 514 Labs have been building as an open source tool product.
The Muse Headband is an industry leading wearable device that measures brainwave activity. It is a small, comfortable band that you wear on your head and it has 4 built-in sensors that measure the electrical activity of your brain.
We also support the use of OpenAI's gpt-4o model to analyze your brain data. This is optional, but it's a great way to obtain deeper insights into your brain data.
To enable OpenAI gpt-4o data analysis, you'll need to set the OPENAI_API_KEY environment variable in the moose project's .env.local file (which you'll need to create) located in the apps/brainmoose directory. This is optional, but it's a great way to see deeper insights into your brain data.
The Muse device also has accelerometer and gyroscope sensors in order to capture 3D head movement.
Those are the same sensors used in smart watches and fitness trackers. Using these sensors, we calculate movement scores for a particular session.
The Consumption API is a Moose API that allows you to query the movement scores for a particular session.
http://localhost:4000/consumption/sessionInsights?sessions=1735784964|Meditation,1735785243|Coding
The sessions parameter is a comma separated list of sessionId and sessionLabel pairs.
See a sample response in the SampleAnalysis/session01.md file.
To allow Moose and the Muse Headband to communicate, we use a UDP Datagram Protocol (UDP) server called DAS (Data Acquisition Service) to receive UDP packets from an app that connects to the Muse Headband.
The DAS service (apps/das) is a Node.js application that uses the osc-min library to parse incoming UDP packets and send them via HTTP to the Moose server.
Moose in-turn ingests the data and uses RedPanda and the Clickhouse database to store in-flight and resting data.
timestamp DateTime('UTC')
bandOn Bool
acc Nested(x Float64, y Float64, z Float64)
gyro Nested(x Float64, y Float64, z Float64)
alpha Float64
beta Float64
delta Float64
theta Float64
gamma Float64
sessionId String
We evaluate head movement during a session by calculating a movement score. We do this by aggregating the magnitude of the vectors accelerometer (acc) and gyroscope (gyro) data for a specific sessionId.
The magnitude of a vector is calculated using the Euclidean norm (i.e., the square root of the sum of the squares of its components).
Here's a Clickhouse SQL query to calculate the movement scores:
SELECT
sessionId,
SUM(sqrt((arrayElement(acc, 1) * arrayElement(acc, 1)) +
(arrayElement(acc, 2) * arrayElement(acc, 2)) +
(arrayElement(acc, 3) * arrayElement(acc, 3))))
AS acc_movement_score,
SUM(sqrt((arrayElement(gyro, 1) * arrayElement(gyro, 1)) +
(arrayElement(gyro, 2) * arrayElement(gyro, 2)) +
(arrayElement(gyro, 3) * arrayElement(gyro, 3))))
AS gyro_movement_score,
(SUM(sqrt((arrayElement(acc, 1) * arrayElement(acc, 1)) +
(arrayElement(acc, 2) * arrayElement(acc, 2)) +
(arrayElement(acc, 3) * arrayElement(acc, 3)))) +
SUM(sqrt((arrayElement(gyro, 1) * arrayElement(gyro, 1)) +
(arrayElement(gyro, 2) * arrayElement(gyro, 2)) +
(arrayElement(gyro, 3) * arrayElement(gyro, 3)))))
AS total_movement_score
FROM
Brain_0_0
WHERE
sessionId = '1735785243'
GROUP BY
sessionId;
Meditation Movement Scores
| sessionId | acc_movement_score | gyro_movement_score | total_movement_score |
|---|---|---|---|
| 1735784964 | 96119.6529543984 | 773537.9565430604 | 869657.6094974588 |
Coding Movement Scores
| sessionId | acc_movement_score | gyro_movement_score | total_movement_score |
|---|---|---|---|
| 1735785243 | 91714.62041451405 | 801765.3217066663 | 893479.9421211804 |