Recently, neuronal correlates of acupuncture stimulation in human brain have been investigated by functional neuroimaging. The preliminary findings suggest that acupuncture at analgesic points involves the pain-related neuromatrix and may have acupoint-brain correlation. Although multiple models of control stimulations have been applied to address the specificity of the needling effect clinically, their impacts have not been evaluated by functional neuroimaging. With the advantage of objective parameter setting, electroacupuncture (EA) was used in this study to devise three distinct controls for real EA, i.e., mock EA (no stimulation), minimal EA (superficial and light stimulation), and sham EA (same stimulation as real EA) applied at nonmeridian points. Fifteen healthy volunteers received real EA at analgesic point Gallbladder 34 (Yanglinquan), sham EA, and one of either mock EA or minimal EA over the left leg in counter-balanced orders. Multisubject analysis showed that sham EA and real EA both activated the reported distributed pain neuromatrix. However, real EA elicited significantly higher activation than sham EA over the hypothalamus and primary somatosensory-motor cortex and deactivation over the rostral segment of anterior cingulate cortex. In the comparison of minimal EA versus mock EA, minimal EA elicited significantly higher activation over the medial occipital cortex. Single-subject analysis showed that superior temporal gyrus (encompassing the auditory cortex) and medial occipital cortex (encompassing the visual cortex) frequently respond to minimal EA, sham EA, or real EA. We concluded that the hypothalamus-limbic system was significantly modulated by EA at acupoints rather than at nonmeridian points, while visual and auditory cortical activation was not a specific effect of treatment-relevant acupoints and required further investigation of the underlying neurophysiological mechanisms.