Impulsivity is a clinical hallmark of borderline personality disorder (BPD).1 Motor response inhibition and interference control are aspects of behavioral impulse control, often used to assess impulsivity. 2 As a key region in fronto-striatal circuits, the anterior cingulate cortex (ACC) is crucial for conflict monitoring and rich in γ-Aminobutyric acid (GABA) innervation, but also important for BPD pathology. 3, 4 In this study, we tested whether ACC GABA levels would predict the activity and connectivity in the fronto-striatal network during interference inhibition in BPD patients.26 female BPD patients (26.77±6.55 years) and age-matched 25 female healthy controls (HCs) (27.53±6.59 years) were included. In vivo single voxel 1H-MRS using a MEGA-PRESS editing sequence was applied to acquire GABA data in the bilateral ACC voxel (40x30x20mm3) during the resting state (Fig.1). Afterwards, a Simon task was presented by an event-related design consisting of incongruent trails and congruent trials, in order to evaluate the interference inhibition. All images were acquired in a 3T scanner (MAGNETOM TimTrio; Siemens, Erlangen, Germany). GABA spectra were quantified with the jMRUI software. The fMRI data was carried out using statistical parametric mapping (SPM8). The generalized psycophysiogical interaction (gPPI) was performed for the task-modulated functional connectivity analysis, and bilateral ACC were chosen as a seed ROI based on the automated atlas labeling (AAL) parcellation. To examine the associations between ACC GABA levels and task-related BOLD response as well as, connectivity with ACC, a series of whole-brain voxel wise regression analysis were performed via SPM for both individual groups. The ACC GABA levels for each participant served as the regressor of interest for all analyses. The significance levels were set to P < 0.001(uncorrected) on a cluster level of ≥ 10 voxels.No group difference between BPD patients and HCs was observed at the behavioral level. Compared to HCs, BPD patients showed no difference in BOLD response but lower ACC-caudate connectivity in response to incongruent vs. congruent trials (Fig. 2(a)). This altered connectivity effect was driven by the decrease in ACC-caudate connectivity during the incongruent condition (Fig. 2(b)), and no group difference in connectivity with bilateral ACC was observed during the congruent condition. ACC GABA levels in BPD patients did not differ from HCs. BPD patients showed a significant positive correlation of the GABA levels and the BOLD response in fronto-striatal regions (including ACC, putamen, frontal areas and caudate) when reacting to incongruent vs. congruent trials (Fig. 3). Additionally, GABA levels in the ACC in BPD patients were positively related to ACC-caudate functional connectivity during the incongruent condition (Fig. 4). Those significant correlation effects were not found in the HCs group.Our results suggested disturbed ACC–caudate functional integration in BPD patients although they performed similarly compared to healthy subjects during interference inhibition in an emotionally neutral state. Moreover, this is the first evidence of a modulatory effect of ACC GABA levels on task-related recruitment and integration of fronto-striatal networks during interference inhibition in BPD patients. Our finding highlights that the GABAergic system in the ACC might play an important role in the modulation of impulsivity via regulating the local neural activity and remote connectivity between key regions.