The Child Mind Institute is developing mhealthx as a general-purpose, open source software package to automate feature extraction from sensor data. Arno Klein originally created mhealthx at Sage Bionetworks to extract voice, accelerometry, and touchscreen tapping features from mobile health research apps such as mPower, the Parkinson disease symptom tracking app built on top of Apple’s ResearchKit. Behind the scenes, open source Python 3 and other languages run within a modular Nipype pipeline framework on Linux (tested with Python 3.5 on Ubuntu 14.04).

Release:1.0.0
Date:May 17, 2017

Links:

Inputs

All data are optionally accessed from Synapse tables in a project on synapse.org:

  • Voice: WAV files
  • Tapping: JSON files
  • Accelerometry: JSON files

Processing

  • Run different feature extraction software packages on the input data.
  • Output features to new tables.

Outputs

  • Tables