Our stimuli were large (approx 3 deg), and had low spatial frequencies (0.86 cpd). The gratings in figure 1 are much smaller than the gratings used in the real experiment. Changes in contrast polarity with our stimuli were easily observed.
However, when short temporal intervals separated multiple contrast reversals from each other, the reversals were less conspicuous, and possibly as a result thereof rivalry set in.

I know of three papers that have looked at the influence of polarity reversals on conventional binocular rivalry.

You have investigated it in 1986 (O’Shea & Blake, 1986, perception and psychophysics), using sinewave gratings that were counter-phased. One grating was modulated at 4Hz, the frequency of the other grating changed over trials.
In the methods section of O’Shea & Blake it is mentioned that this procedure yielded no normal rivalry for 0.5 Hz and 1Hz modulations, but it did at higher temporal frequencies. This suggests that normal rivalry would set in at >2Hz modulations (not far from the limit we reported). In O’Shea & Blake the gratings were small (1.2 deg) and had a higher spatial frequency (4 cpd), compared to our study (~3deg and 0.87 cpd). In our study the contrast reversals should therefore have been easier to not notice, and the reversals were more likely to disrupt the rivalry process.

The other two studies looked at the influence of sudden contrast reversals on the continuation of a dominance period during rivalry (Blake & Fox, 1974, Vision Res; Walker & Powell, 1979, Vision Res). In some sense this more similar to our stimuli which consisted of sudden contrast reversals in one eye, while the other eye temporarily received a non-changing stimulus.
These studies found that the dominance duration was shortened when a contrast reversal was executed in the suppressed eye. (FYI: both studies used 4cpd squarewave gratings, size 1x1 deg (Blake & Fox), and a circular 1 deg patch, and 4.3 deg patch for suppressed and dominance stimuli respectively.)

It therefore seems that there is evidence that contrast reversals do influence conventional binocular rivalry, as well as evidence that contrast reversals do not influence conventional binocular rivalry.