why does interpolation in high frequency time series create spurious correlation?

Hi all,

I am reading some papers on high frequency financial data analysis, by Engle.

Could anybody point me to some more indepth/tutorial treatment (such
as books), where it talks about why interpolation in high frequency
time series (resampling irregularly spaced transaction data into
regularly spaced usual time series) creates spurious correlation? (My
goal is to study the correlation of two high frequency time series,
and see if there could be pairs trading opportunities or other trading
opportunities.)

Thank you!