ImportancePrimary insomnia (PI) has a high global incidence, and effective treatments with fewer side effects are needed. Acupuncture, a treatment used in traditional Chinese medicine, has become increasingly established as a treatment method for PI and is recognized by many physicians and patients. Some evidence has suggested that acupuncture was associated with improvements in objective sleep parameters and might induce changes in some brain regions. Individual studies with limited sample size and low detection thresholds may lead to false positives, and no systematic review of the effects of acupuncture has been conducted in PI.ObjectiveThe aim of this systematic review and coordinate-based meta-analysis was to summarize the literature on fMRI evaluation of patients with PI treated with acupuncture.DesignWe performed a methodical and comprehensive search of multiple publication databases (from inception to December 2022): Web of Science, PubMed, ScienceDirect, Embase, Wan Fang, China National Knowledge Infrastructure, and Chinese Scientific Journal Database. Bias and quality of studies were evaluated by three researchers. Furthermore, a seed-based D-mapping meta-analysis with permutation of subject images (SDM-PSI) was applied to investigate the central mechanisms behind acupuncture treatment at PI. The International Prospective Registry of Systematic Reviews received the protocol for this study. (PROSPERO: CRD42023400086).ResultsThe analysis included 305 patients with PI and 116 healthy controls from 11 studies. SDM-PSI analysis showed that patients with PI exhibited increased amplitudes of regional homogeneity and low-frequency fluctuations in the left superior frontal gyrus (1352 voxels, p = 0.0028), right angular gyrus (14 voxels, p = 0.0457), and cerebellum (12 voxels, p = 0.0446). Acupuncture improved the function of right superior frontal gyrus (1, 404 voxels, p = 0.0123), left inferior frontal gyrus (1068 voxels, p = 0.0088), left inferior temporal gyrus (903 voxels, p = 0.0074), left supramarginal gyrus (888 voxels, p = 0.0113), left precuneus (457 voxels, p = 0.0247), right precuneus (302 voxels, p = 0.0191), left supplementary motor area (82 voxels, p = 0.0354), and right parahippocampal gyrus (28 voxels, p = 0.0379). The brain regions affected by non-acupoint acupuncture were all located in the frontal lobe. The Cochrane risk-of bias tool and MINORS5 were used for quality assessment and the included articles had high performance bias and attrition bias.ConclusionThis coordinate-based meta-analysis found that acupuncture in patients with PI had significant effects on the default mode network, particularly on the frontal lobe and precuneus, and that non-acupoint acupuncture may provide some benefit to frontal brain region function.Systematic review registrationPROSPERO: CRD42023400086.