Different results when downsampling activation data

[blake harper]

Hello,

I am using measured muscle activation (calculated from EMG) and kinematics to calculate torques using the analyze tool. The original EMG data was measured at a sampling frequency of 200 Hz. I have run the analyze tool with activation at the full sampling frequency as well as downsampled versions of the same activation data. The torque results from the analyze tool vary greatly depending on the sampling frequency (see solid lines on attached torque plots) even though I can't see a noticeable difference in the actual activation when I downsample (see attached downsampled activation).

Has anyone ever encountered this change in results when downsampling data? Is there a minimum sampling frequency that should be used for activation or another reason why this might be happening?