Within the field of neuroimaging, the discovery of a constellation of brain regions silently active when we are Bresting^has provided a new view into the elusive effects of meditative practice. This network, called the default mode network (DMN), has been shown by functional neuroimaging to be active when an individual is at rest. Meta-analyses of the fMRI neurocorrelates of meditation have shown that across diverse practices, the most common general effect appears to be modulation of regions within the DMN. The specific practice of mantra meditation is a form of task-positive concentration. These kinds of task-positive activities are regarded by many to be an anti-correlate of DMN activity, thereby in theory, reducing activations within the DMN as concentration on a particular task increases. Yet previous studies on mantra meditation using language tasks as controls have reported only activations, not deactivations within the DMN. A study by Berkovich-Ohana showed that word repetition as control task could mask some of the effects of mantra, which may explain why previous studies using language repetition as control task failed to find significant DMN deactivations during mantra meditation. To investigate these differences, we analyzed mantra meditation using a finger-tapping control task which required a minimal amount of focused attention, yet enough to reduce blood flow to the major hubs of the DMN and minimally reduce network activity. Using a finger-tapping control thus modifies the research question: Can a brief course of this kind be shown to modulate DMN activations during mantra practice beyond that of a minimally active/nonlanguage finger-tapping control condition; and if so, can we see a training effect over time? Our results show that over a 2-week period of Kundalini yoga/meditation, participants successfully decreased activations within subregions of the DMN, namely the pregenual anterior cingulate cortex (PACC) and the precuneus, during mantra meditation with trends toward decreased activation in the posterior cingulate cortex (PCC) as well. These decreased activations were significant despite use of a finger-tapping control task known to deactivate these regions, indicating that mantra practice suppresses activation in these DMN regions beyond the active control task. These results emphasize the importance of the choice of control task and provide insight into the Bmantra effect.^Here, we show that training in mantra meditation, like other practices such as focused attention and open monitoring, also has a suppressive effect on activity within the DMN.