The advantage of using Mean over Non-Pulsatile Mean (NPMN) is that the Mean derived data honors the placement of bad data marks; meaning it takes into account the user defined settings within the Noise tab, if enabled. This is useful when railed data or telemetry dropout is detected by the Noise filter and Bad Data Marks are automatically placed around these impacted data section. 

The NPMN will average all waveform data points, including those within Bad Data Marks.

Here you can see a section of data where the animal moved into a dead spot. The NPMN shows varying degrees of inaccurate data depending on the amount of dropout within the logging period. The Mean parameter however replaces the data with an x if there has been any dropout during the logged period. It is preferable to report no data as opposed to incorrect data.

Additional Resource

T_Mean displays "x" for values, while values show under T_NPMN