establish a working manual for accelerometer analyses with GGIR

[mwvoss]

Task: Draft a document walking through analysis of accelerometer data with GGIR. The document should outline steps for analysis, include example code, and include rationale for parameter decisions based on the literature or knowledge of the population.

For the text, suggest to first write into a google doc for easy shared editing and commenting, and then we can put the finished first draft in a markdown file (ends in .md) on the GitHub repository for accelerometer analyses (https://github.com/HBClab/accelerometerAnalyses).

We'll name the markdown accelerometerAnalysesGGIR.md, and then we'll modify the accelerometerProcessingManual.md later to reflect only the setup and data transfer.

[mwvoss]

Include our rationale and decisions about the following questions from Jairo Hidalgo Migueles working with IGNITE team:

• What time the accelerometers started and finished the recordings?

• How many days?

• Do you plan to remove the first and last day of the measurement? Or do you want to use everything?

• Do you want to split days at midnight? It would result in 24-h blocks; or would you rather use the waking-up time of every participant to split the days? It would result in inconsistent block lengths.

[mwvoss]

Link to working google document to flesh out our rationale and mechanics for usage of GGIR with our older adult samples

[mwvoss]

We will consider the code used for the Cabanas-Sanchez et al 2019/2020 paper as our new code-base to start from since it's been used in a published paper that does the same analysis we'd like to do.

code

However, before we run it, we should figure out if anything needs to be changed to better match our accelerometer methods.

google doc of code annotated for deciding on modifications

The main methods to consider would be that our participants wore their device for 7 days on their non-dominant hand, with the device set to record at 60 Hz from midnight-to-midnight.