Interesting study which may contribute to our understanding. Three comments though:
1) Why expect cognitive inflexibility on a performance task in case of OC patients whose specialty is accuracy and main concern is avoidance of performance errors? In terms of errors the performance of OC patients on switch events compared with repeat events decreased from excellent to good, whereas that of MD patients decreased for mediocre to rather poor, while that of controls decreased from mediocre to very poor. Inflexibility of this type of OC patients is to be expected on learning tasks, because in contrast to controls they are reluctant to rely on trial and error learning, preferring to rely on a conscious, controlled way of executing new tasks. They will start better than controls, but fail to improve very much, unlike controls. Long ago I demonstrated such an effect in mirror drawing (not published).
2) The investigators do not seem to anticipate heterogeneity of OCD. However, in many studies such heterogeneity has been demonstrated or at least been made likely. The OCD group (N=18) of the investigators consisted of 10 people scoring high on checking and one on precision, one scored high on washing. I guess the present results with respect to error rates are mainly due to OC patients who are mainly characterized by (performance) checking and precision. In follow-up research a big difference in error scores may be found between typical washers and typical checkers.
3) Finding a neurological correlate of behavioral abnormalities does not make the first a better marker (biological) of the disorder in question than the second. First, there will always be much "noise", attenuating such correlations, and much is still unclear about these correlates. Second, this seems to reverse the usual procedures: the behavioral abnormality is the thing of interest, the thing to correct, and it is probably much better and more directly to demonstrate than a neurological abnormality. Third, a possible neurological correlate of OCD may be different for washers than for checkers.