Abstract
This study examined women with premenstrual syndrome (PMS) during the late luteal phase, focusing on cognitive function under stress. Additionally, we investigated the association between cognitive performance and interoceptive processing at baseline. Acute stress was experimentally induced in women using the Trier Social Stress Test (TSST). We evaluated performance on the Emotional Face-Word Stroop task and the 3-Back task, comparing PMS and without PMS groups at pre-stress, post-stress, and recovery phases. Furthermore, we analyzed the association between baseline Emotional Interference and the Interoceptive Mismatch, which was defined as the discrepancy between interoceptive accuracy and sensibility. Mixed ANOVA revealed that, in the Emotional Face-Word Stroop task, the PMS group exhibited increased error rate without group differences in reaction time. Linear mixed-effects modeling (LMM) indicated subjective performance interference positively predicted incongruent reaction time, capturing individual variability beyond group effects. Conversely, in the 3-Back task, no group differences were observed, but LMM showed increased sensitivity from signal detection theory at post-stress and recovery phases, reflecting individual variability. Moreover, regression analysis revealed that greater Interoceptive Mismatch was significantly associated with lower Emotional Interference. This observation suggests that women with PMS exhibit a processing bias toward bodily sensations and emotions, reflecting a bottom-up style of cognitive processing. These results highlight the importance of understanding cognitive vulnerability in PMS from a multidimensional perspective, including interoception and subjective experiences. This understanding may contribute to the development of individualized support and preventive interventions for women with PMS.