Tweet
In reply to my previous posting, Lambert Schomaker says:
> o Why sampling the EMG envelope, which is low-pass filtered at 30 Hz,
> with such a high frequency (333.33Hz), more than 10 times f_Nyquist?
>
> o Why does the band-pass filter start with a high cut-off frequency
> of 30 Hz? In human surface EMG recordings, the band between 4 and 30 Hz
> contains a significant amount of (real myoelectric) power. In this
One minor correction: the Nyquist frequency is not equal to the highest
signal frequency, but twice the signal frequency. As a rule-of-thumb,
one usually samples at at least three times the signal frequency, which
in my case would be 90 Hz. Why then sample at 333 Hz? On reason is
simple: to be on the safe side. A filter with a cut-off frequency of
30 Hz can actually pass higher frequencies, but very much attenuated.
Another reason is, that we simultaneously sample the hoof acceleration,
which is not low-pass filtered. The hoof acceleration is used as an
alternative to a footswitch, and this signal contains a nice discontinuity
at the time of hoof impact. This discontinuity is used for triggering
the ensemble averaging. I suppose it should be possible to filter this
signal as well, while retaining enough of this sharp discontinuity, but
I haven't tried. But on the other hand, an old Chinese proverb says:
"If it works, leave it alone".
Lambert's second point is very interesting, I had always thought that
variations in EMG intensity would be the same throughout the spectrum.
I could probably find some literature references that make this
statement. If, as he suggests, the part below 30 Hz reflects a different
aspect of the muscle activity, we should carefully think what we want
to measure: the firing rate or the amplitude of the action potentials.
In practice however, I think the problem will not be so dramatic. I'll
check the paper by van Boxtel and Schomaker anyway. A serious disadvantage
of extending the passband of the filter to 4 Hz would be that movement
artifacts are much less suppressed.
--------------------------------------
Ton van den Bogert
Dept. of Veterinary Anatomy
University of Utrecht, Netherlands.
> o Why sampling the EMG envelope, which is low-pass filtered at 30 Hz,
> with such a high frequency (333.33Hz), more than 10 times f_Nyquist?
>
> o Why does the band-pass filter start with a high cut-off frequency
> of 30 Hz? In human surface EMG recordings, the band between 4 and 30 Hz
> contains a significant amount of (real myoelectric) power. In this
One minor correction: the Nyquist frequency is not equal to the highest
signal frequency, but twice the signal frequency. As a rule-of-thumb,
one usually samples at at least three times the signal frequency, which
in my case would be 90 Hz. Why then sample at 333 Hz? On reason is
simple: to be on the safe side. A filter with a cut-off frequency of
30 Hz can actually pass higher frequencies, but very much attenuated.
Another reason is, that we simultaneously sample the hoof acceleration,
which is not low-pass filtered. The hoof acceleration is used as an
alternative to a footswitch, and this signal contains a nice discontinuity
at the time of hoof impact. This discontinuity is used for triggering
the ensemble averaging. I suppose it should be possible to filter this
signal as well, while retaining enough of this sharp discontinuity, but
I haven't tried. But on the other hand, an old Chinese proverb says:
"If it works, leave it alone".
Lambert's second point is very interesting, I had always thought that
variations in EMG intensity would be the same throughout the spectrum.
I could probably find some literature references that make this
statement. If, as he suggests, the part below 30 Hz reflects a different
aspect of the muscle activity, we should carefully think what we want
to measure: the firing rate or the amplitude of the action potentials.
In practice however, I think the problem will not be so dramatic. I'll
check the paper by van Boxtel and Schomaker anyway. A serious disadvantage
of extending the passband of the filter to 4 Hz would be that movement
artifacts are much less suppressed.
--------------------------------------
Ton van den Bogert
Dept. of Veterinary Anatomy
University of Utrecht, Netherlands.