Duplicated sampling_rate Attribute on `waveform_mean` and `waveform_sd`

[h-mayorquin]

The sampling_rate attribute is defined identically on both waveform_mean and waveform_sd datasets in the Units table schema. This duplication is redundant since both datasets must share the same sampling rate to be meaningful, and it introduces the possibility of inconsistent values. PyNWB already treats this as a single value through the waveform_rate parameter, which gets mapped to both dataset attributes during serialization.

I think the schema should be refactored to have a single waveform_rate attribute at the Units group to prevent potential inconsistencies and align with the API.

c.c. @alejoe91