biopsykit.sleep.sleep_wake_detection.algorithms.cole_kripke module

Sleep/Wake detection using the Cole/Kripke Algorithm.

class biopsykit.sleep.sleep_wake_detection.algorithms.cole_kripke.ColeKripke(**kwargs)[source]

Bases: biopsykit.sleep.sleep_wake_detection.algorithms._base._SleepWakeBase

Class representing the Cole/Kripke Algorithm for sleep/wake detection.

The Cole/Kripke Algorithm runs sleep wake detection on epoch level activity data. Epochs are 1 minute long and activity is represented by an activity index which comes from Actigraph data or from raw acceleration data converted into activity index data.

Parameters

  • epoch_length (int) – epoch length in seconds. Epoch lengths are usually 10, 30, or 60 seconds.

  • scale_factor (float) – scale factor to use for the predictions (default corresponds to scale factor optimized for use with the activity index, if other activity measures are desired the scale factor can be modified or optimized.) The recommended range for the scale factor is between 0.1 and 0.25 depending on the sensitivity to activity desired, and possibly the population being observed. According to the paper by Cole and Kripke, the scale factor depends on the epoch length. See the paper for more details.

References

Cole, R. J., Kripke, D. F., Gruen, W., Mullaney, D. J., & Gillin, J. C. (1992). Automatic Sleep/Wake Identification From Wrist Activity. Sleep, 15(5), 461-469. https://doi.org/10.1093/sleep/15.5.461

fit(data, **kwargs)[source]

Fit sleep/wake detection algorithm to input data.

Note

Algorithms that do not have to (re)fit a ML model before sleep/wake prediction, such as rule-based algorithms, will internally bypass this method as the fit step is not needed.

Parameters

data (array_like) – input data

predict(data, **kwargs)[source]

Apply sleep/wake prediction on activity index values.

Parameters

  • data (array_like) – array with activity index values

  • **kwargs –

    additional arguments to be passed to the algorithm for prediction, such as:

    • rescore_data (bool): True to apply Webster’s rescoring rules to the sleep/wake predictions, False otherwise. Default: True

    • epoch_length (int): activity data epoch lengths in seconds, i.e. Epoch lengths are usually 30 or 60 seconds. Default: 60

Returns

dataframe with sleep/wake predictions

Return type

SleepWakeDataFrame

biopsykit.sleep.sleep_wake_detection.algorithms package biopsykit.sleep.sleep_wake_detection.algorithms.cole_kripke_old module