Nyquist sampling theorem (ie maths) - if sampling at a rate of f, you will get frequency-aliasing unless the input is filtered to remove everything at or above f/2.
It might be better to sample much more often that 0.1s if you want more information.
Nyquist is why common sampling frequencies for quality audio are 44.1kHz or 48kHz, as the human hearing range has a maximum of about 20kHz.