The most serious complication of neurovascular interventions is distal cerebral embolism. Diffusion-weighted MR imaging (DWI) appears to be the most sensitive technique for detecting early and small ischemic lesions. To evaluate the incidence and radiological features of embolic events associated with neuro-intervention, we applied DWI to screening for procedure-related ischemic lesions including silent embolisms. One hundred and thirty-seven patients who have received 154 neuro-interventional procedures were studied with DWI before and within 5 days after treatment. Imaging was performed, using single-shot echo-planar imaging with b value of 1000-1100 sec/mm2. DWI findings were classified into 5 groups by size and location of lesions: type 0 (n = 71), no lesions; I (n = 33), lesions in border-zone regions only; II (n = 9), lesions at perforator territories mainly; III (n = 29), small territorial lesions (< 5 mm); IV (n = 12), large territorial lesions (> or = 5 mm). DWI detected procedure-related lesions in 83 of 154 procedures (53.9%), 36 of which demonstrated new neurological symptoms during and/or after procedures. The parent artery occlusion for cerebral aneurysms had a higher incidence of symptomatic embolisms than other procedures. In 71 of 154 procedures (46.1%), DWI detected no lesions (type 0). Although type I was the most frequent pattern presented, it included few neurological symptoms. Type III often resulted in transient symptoms, and type II and IV tended to induce strokes. Because ischemic lesions detected by DWI were likely to arise in border-zone territories by parent artery occlusions, we considered that hypoperfusion as well as emboli were involved in the evolution of cerebral infarctions. Thus, DWI is a useful method to detect silent embolisms, and to determine the safety of neuro-intervention and the mechanism of embolic ischemia.