Note: Only the H1H2 times listed by Vuk were used for flagging
      intervals, since we have no a priori explanation for why
      the H1L1 intervals should have had correlated noise.

      Even the H1H2 times should be used with caution.