Jiang et al (2006) tested the effect of "invisible" erotic pictures on the attention of gay/bisexual and heterosexual males and females. After comparing attraction/repulsion to male or female erotic stumuli, Jiang et al found that while gay men responded like heterosexual women to the suppressed images, gay/bisexual women responded in between this group and the heterosexual men.
Lykins et al (2006) found that both men and women focused less on faces and more on bodies in response to normally presented erotic (vs. non-erotic) stimuli, and the results suggested that women's attentional patterns may be more affected than men's by the erotica context. However, Lykins et al point out that because their study showed different pictures to the male and female groups (heterosexually delineated) comparison between groups is not valid.
Neither of these studies report the menstrual cycle stage of female participants, which may be an issue.
Gizewski et al (2006) studied fMRI results of 25 women during mid-luteal phase (post ovulation) and during menses. Self-reported arousal in women was similar to men's at the mid-luteal test, but significantly lower at menses. Some activation differences between mid-luteal and menstrual phases were reported (e.g. reduced in the left thalamus at menses). However, between-sex differences were more robust than menstrual phase differences.
One big concern I have with studies using "erotic stimuli" is the source of the images. Lykins et al wrote that they got their images of females from Playboy's website, and images of males from Falcon Studio's website. Both of these websites design their erotic content for men. Unfortunately, I can't think of any source for erotic images of men by women that would match Playboy on production values. Obviously, more women-driven erotica needs to be produced, in the name of science!