Skip to main content

Tools to losslessly compress and decompress raw neural data.

Project description

Neuralink Compression Challenge

Summary

content.neuralink.com/compression-challenge/data.zip is one hour of raw electrode recordings from a Neuralink implant.

This Neuralink is implanted in the motor cortex of a non-human primate, and recordings were made while playing a video game, like this.

Compression is essential: N1 implant generates ~200Mbps of eletrode data (1024 electrodes @ 20kHz, 10b resolution) and can transmit ~1Mbps wirelessly. So > 200x compression is needed. Compression must run in real time (< 1ms) at low power (< 10mW, including radio).

Neuralink is looking for new approaches to this compression problem, and exceptional engineers to work on it.

Task

Build executables ./encode and ./decode which pass eval.sh. This verifies compression is lossless and measures compression ratio.

Your submission will be scored on the compression ratio it achieves on a different set of electrode recordings. Bonus points for optimizing latency and power efficiency

Submit with source code and build script. Should at least build on Linux.

Pypi Brainwire Package

Encode and Decode are now available under the brainwire pypi package. The modules will detect neural spikes, compress the data, and decompress the data to present a sample rate and array of amplitudes that only contain detected spike information. They can be downloaded using pip install brainwire and implemented with brainwire.encode.compress(file_path) or brainwire.decode.decompress(compressed_brainwire_data_format) where 'compressed_brainwire_data_format' is the compressed data returned from 'brainwire.encode.compress()' and the 'file_path' is the path to raw neural data in '.wav' format.

Sample Compression Results:

Encoding Time:

  • Time Δ Microseconds: 3686937.0 μs
  • Time Δ Milliseconds: 3686.937 ms
  • Time Δ Seconds: 3.686937 s

Decoding Time:

  • Time Δ Microseconds: 3871124.0 μs
  • Time Δ Milliseconds: 3871.124 ms
  • Time Δ Seconds: 3.871124 s

Total Time (encode and decode):

  • Time Δ Microseconds: 7558100.0 μs
  • Time Δ Milliseconds: 7558.1 ms
  • Time Δ Seconds: 7.5581 s

Compression Ratio

  • Original File Size: 197526
  • Compressed File Size: 92251
  • Percent of Compression: 53.30%

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

brainwire-0.0.8.tar.gz (15.7 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

brainwire-0.0.8-py3-none-any.whl (11.5 kB view details)

Uploaded Python 3

File details

Details for the file brainwire-0.0.8.tar.gz.

File metadata

  • Download URL: brainwire-0.0.8.tar.gz
  • Upload date:
  • Size: 15.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.1.1 CPython/3.11.5

File hashes

Hashes for brainwire-0.0.8.tar.gz
Algorithm Hash digest
SHA256 62be67eec34919f7fbe3293717c1b569f2561e463295bd43feaacd674a4afa17
MD5 260e2373a8185a7f44d18c5566564fef
BLAKE2b-256 ecd612e87911a73da25c34e7bc1d97062bea6ccb7d0d8f8588247285208f3317

See more details on using hashes here.

File details

Details for the file brainwire-0.0.8-py3-none-any.whl.

File metadata

  • Download URL: brainwire-0.0.8-py3-none-any.whl
  • Upload date:
  • Size: 11.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.1.1 CPython/3.11.5

File hashes

Hashes for brainwire-0.0.8-py3-none-any.whl
Algorithm Hash digest
SHA256 e79204c24d788fb3cbe97809bf12fa673570476bf7d71d489fa7a89739aacce3
MD5 4a6118a63e03a743d96f1ed8a3926b19
BLAKE2b-256 c68859dfcc77046f22e91bffaf919e2fe750b06dc3a999a86b6da96dee20df5f

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page