Linear trend removal is automatically included when applying high pass filtering to the data via TemporalHighPassFilter(). However, it can also be applied separately, via the command <FMRname>.LinearTrendRemoval().
The high pass filter is applied in the frequency domain.
The temporal high-pass filter can be specified in cycles or in Hz.
When it is specified in cycles, is this relative to the number of time points specified in the functional data, i.e. length of the time series.
When the units are specified in "Hz", it is in signal intensity changes per Hz, f.e:
succeeded = docFMR.TemporalHighPassFilter(0.115385, "Hz");
The units in Hertz can be computed from the cycles per time course via: number of cycles / number of data points.
Temporal Gaussian smoothing in FMR projects can be applied by specifying the width of the smoothing kernel in seconds ("s") or in TR ("TR").