The GGIR 3.3-0 release is now on CRAN. In this blog post I will talk you through the main updates since the 3.2-6 CRAN release from April 2025. For a full overview see the GGIR changelog, where # numbers refer to specific issues in the GGIR GitHub repository.
General updates
- The guidance on citing GGIR has been revised. The 2019 journal article was a good description of GGIR at that point in time, but is now outdated. Further, a general problem with journal papers about research software is that they fuel the belief that only papers matter in science taking away credit from software development work. Therefore, we now recommend citing GGIR as software via its own DOI: https://wadpac.github.io/GGIR/authors.html#citation
- In the documentation we now promote the GitHub discussion page as the central place to ask questions and report problems. The GGIR Google group has been closed for new members and new messages. References to the Google group have been removed from the documentation.
New functionality
- As you may know, GGIR offers the option to estimate the main sleep window when no sleep diary is available. This approach sometimes fails to handle severely fragmented sleep. To address this an optional algorithm has been added that identifies and corrects such misclassified nights. For more details see the corresponding new documentation.
- Previously, working with accelerometer data collected in several timezones forced us to process each time zone subset separately. This limitation has now been partly addressed. In the new situation you only need to process data from different timezones separately in GGIR part 1. Next, you can move all part 1 milestone files to one output folder and run all subsequent GGIR parts (2-6) based on that folder.
- Visual reports (pdf) via parameter `visualreport` are now also generated in parallel by default, just like the rest of GGIR.
- Specifying unknown or mistyped parameter names to GGIR functions now triggers an error message. The message informs the user about the unknown parameter name and urges them to fix it.
Bugs fixed in part 4
- Bug #1376 caused more nights to be included in the analyses than desirable if all dates in a sleeplog for at least one recording did not overlap with the dates in an accelerometer recording.
- Bug #1366 caused GGIR part 4 to not load the sleeplog data from new data added to an existing analysis. This has now been fixed. The sleeplog is now reloaded when data in part 3 is newer than the timestamp of the diary or if the diary itself has been changed since the last time it was loaded.
- Bug #1359 made it impossible to use basic format sleeplog in combination with sleepwindowType “TimeInBed”, this has now been fixed.
Deprecated functionalities
- The experimental nap-detection functionality that was added in GGIR release 2.5-2 has been deprecated. It has been superseded by a much more simplified and better integrated nap detection. The new nap detection functionality has so far only been evaluated in pre-schoolers is now discussed in documentation Annex NapDetection.