yasa.moving_transform¶

yasa.moving_transform(x, y=None, sf=100, window=0.3, step=0.1, method='corr', interp=False)[source]¶

Moving transformation of one or two time-series.

Parameters

xarray_like

Single-channel data

yarray_like, optional

Second single-channel data (only used if method in [‘corr’, ‘covar’]).

sffloat

Sampling frequency.

windowint

Window size in seconds.

stepint

Step in seconds. A step of 0.1 second (100 ms) is usually a good default. If step == 0, overlap at every sample (slowest) If step == nperseg, no overlap (fastest) Higher values = higher precision = slower computation.

methodstr

Transformation to use. Available methods are:

'mean' : arithmetic mean of x
'min' : minimum value of x
'max' : maximum value of x
'ptp' : peak-to-peak amplitude of x
'prop_above_zero' : proportion of values of x that are above zero
'rms' : root mean square of x
'slope' : slope of the least-square regression of x (in a.u / sec)
'corr' : Correlation between x and y
'covar' : Covariance between x and y

interpboolean

If True, a cubic interpolation is performed to ensure that the output has the same size as the input.

Returns

tnp.array: Time vector, in seconds, corresponding to the MIDDLE of each epoch.
outnp.array: Transformed signal

Notes

This function was inspired by the transform_signal function of the Wonambi package (https://github.com/wonambi-python/wonambi).