Actigraphy - Science method

Dear Hao-Yu Shih,

If You are interested in the patient's overall activity during the improvement of his/her ADLs, use the unaffected wrist. Use the affected wrist, if You want to know how much the patient uses his/her paretic arm in everyday life, f.i. to validate a forced use paradigm. If You are interested in the change of activity ratio between both arms, well You should have 2 actiwatches or 1 actiwatch on the affected arm and software, which can separate the data from the watch in two parallel data streams for near field activity (= affected arm) and far field activity form the whole rest of the body. - it might be possible by a wavelet analysis of the signal.

Yours

Ulrich Jobst

Jawac device to measure mandibular movements. Read the Bassam Chakar's article.

For the record, I'd like to add the following package; https://github.com/ghammad/pyActigraphy (online documentation available at https://ghammad.github.io/pyActigraphy). The package allows you to:

read native actigraphy data files with various formats:

clean the raw data and mask spurious periods of inactivity;

produce activity profile plots;

visualize sleep agendas and compute summary statistics;

calculate typical wake/sleep cycle-related variables;

automatically detect rest periods using various algorithms (Cole-Kripke, Sadeh, ..., Crespo, Roenneberg);

perform complex analyses:

Disclaimer: I am the developer of that package.

As well as Actiwatch, Actigraph GTX3 has been used in many large research studies and is well validated. I agree adding video monitoring will improve the rigor of your study as really the best way to monitor sleep in a sleep deprivation study is by EEG as this picks up microsleep.

The following review article should be helpful:

If the data is deidentified, feel free to email the file to me, and I can try to change it for you. 

Hi Stefan

I wonder if I can use MESA data with GGIR  R package ,if someone used it , please give me more info

It should be worn in the non-dominant wrist. Normally dominant hand is used for writing, eating and many other day-to-day activities those may likely to exaggerate activities of the subject.  

I'd probably use a common statistical approach:  take an average (mean, median or mode) from a period immediately prior to the nurse intervention, and an average from a period of equal length immediately after the intervention and combine the two to replace the missing data.  I'd keep track of how much of the data is replaced and report that info, and keep track of the time of day at which it occurs, for circadian rhythm analyses.

I work on the National Sleep Research Resource (NSRR -- https://sleepdata.org/). We have published many large, de-identified sleep databases (EDFs and covariates).

None of our current databases match the specifications laid out above (i.e. missing actigraphy), though we hope to soon post data collected as part of the Multi-Ethnic Study of Atherosclerosis Sleep ancillary study (MESA Sleep -- http://www.mesa-nhlbi.org/). As part of that project, participants (2,000+) wore an actigraphy device for 7 days and went through a single night of full PSG. Most subjects have a single night of concurrent PSG and actigraphy. Combining data from different devices may pose new challenges, but it might be fruitful.

Here's a recent publication that describes the MESA Sleep project: https://www.researchgate.net/publication/268791362_RacialEthnic_Differences_in_Sleep_Disturbances_The_Multi-Ethnic_Study_of_Atherosclerosis_(MESA)

i agree that it would make sense to use a 30 sec epoch to align with your direct observation, Actilife will allow you to look at epoch by epoch (based on what you selected) which would likely allow you to match them up fairly easily.

thanks every one for suggestions 

Thank you, Kevin! I will search through the CRAN.

Ian, NPCRA is not very widely used, I don't think. But here is a paper that has used it to characterize rest-activity circadian patterns. Good luck! Please feel free to get in touch by email jaw@bu.edu if you want to chat further--I would also love to hear if you find anything useful. 

There is an free app for that...

Really there is a sleep app for your iphone, developed by some sleep researchers I think from UBC that is pretty good. They have all the validation studies but you will have to bug them for it. Its called "sleep time" by Azumio.

Hello!

it depends on the content of the tool that you have selected. Basically the tool is made considering the target population. If the illiterate population has to be included, a picture based tool could be useful. The numbers of hours of sleep one had had could be marked in the specified box or space denoting the 

dear Pri, this system is very widely used, but depending on the question you want to answer with your measurement, you have to take into account that there is a strong influence of the frequencies of the movements to the amplitude of the measurement. There are also some publications about this.

The correlation with questionaires is very low, also shown in different publications, e.g. Ebner-Priemer

The sturdyness of the unit, reliability and validity of actigraphy would have come into consideration when researchers at Portsmouth University commissioned the purchase of several Phillips Respironics II and Spectrum actiwatch watches. We will be using these in our upcoming pilot and feasibility study on testing the effects of Respiratory Muscle Training on Quality of Sleep and Snoring.

Thanks for your comments and the paper Matthew - I much appreciate your input!

Actigaphy is not a tool used in the diagnosis or treatment of restless leg syndrome. However, some advanced level III home sleep test devices such as the NATUS MPR with its ExG channel, can measure limb movements found in periodic limb movement disorder. As Alain Nicholas above said, RLS is a clinical disorder - and apart from checking the ferritin level and making sure it is in the 100-150 mcg/L range, use of the 4 tiers of pharmacological medications including antiparkinson drugs, benzodiazepines such as clonazepam, anticonvulsant agents such as gabapentin, and in extreme cases, opiates such as tramadol are useful treatment adjuncts. I will start by using oral iron to boost the ferritin level first.

With regard to Actigraph raw data: I developed R-package GGIR for multi-day raw data collected with Genea and GENEActiv, http://cran.r-project.org/web/packages/GGIR/index.html

I could expand this package with Actigraph raw data functionality if people are interested....