The Effect of Meditation on Brain Structure

Cortical Thickness Mapping and Diffusion Tensor Imaging

Do-Hyung Kang; Hang Joon Jo; Wi Hoon Jung; Sun Hyung Kim; Ye-Ha Jung; Chi-Hoon Choi; Ul Soon Lee; Seung Chan An; Joon Hwan Jang and Jun Soo Kwon


Soc Cogn Affect Neurosci. 2013;8(1):27-33. 

In This Article


To our knowledge, this is the first study to investigate structural brain differences in experienced meditators compared with meditation-naïve controls by combining surface-based cortical thickness and DTI analysis. The major findings of the present study were as follows: (i) meditators had thicker cortex in the anterior portions of the brain, located in the frontal–temporal region, including bilateral ventromedial PFC, superior frontal cortex and middle and interior temporal cortices than controls and (ii) meditators compared with controls had thinner cortex in the posterior portions of the brain, located in parietal–occipital region, including bilateral postcentral and inferior parietal cortices and left PCC. Additionally, the left superior frontal cortical thickness adjacent to the primary motor cortex in meditators showed a positive trend correlation with the duration of meditation practice. Regarding WM, DTI results showed higher FA values in the anterior part of the brain in meditators, namely in the MPFC as in the region with the greater cortical thickness and reduced FA values in some of MPFC, PCC and the occipital cortex in meditators compared with controls.

The most striking findings of the present study were greater cortical thickness in the MPFC/OFC regions and superior frontal cortex and lesser cortical thickness in the inferior parietal cortex and PCC. A possible explanation for present findings is the usage-dependent selective elimination of synapses (Huttenlocher and Dabholkar, 1997; Takeuchi et al., 2011a). Previous studies have repeatedly reported that increased cognitive functions after training are associated with decreased GM volumes (Kanai and Rees, 2011; Takeuchi et al., 2011a). Meditation training can enhance various cognitive processes, such as emotional regulation, executive control and attention, particularly sustained attention (Zeidan et al., 2010; Jung et al., 2010). Therefore, thinner cortical thickness of brain regions in meditators, including the lateral and medial parietal areas, may be associated with their enhanced cognitive functions through meditation training, such as improved attention and self-perception. Recent studies revealed the nonlinearity of training-induced GM changes, showing that an initial increase in regional GM volumes followed by a decrease in regional GM volumes (Driemeyer et al., 2008; Boyke et al., 2008), and suggested that these changes are affected by training length and intensity (Takeuchi et al., 2011a,b). In this context, although all brain regions showing thicker and thinner cortices are involved in meditation training we used (i.e. BWV), brain regions showing thinner cortices may be more strongly involved in meditation training used. In the present study, lesser cortical thickness in the meditators was observed in the regions related to self-referential processing. The inferior parietal region such as the angular gyrus is known to be involved in the integration of sensory information from different modalities and the regulation of attention. The inferior parietal cortex and the PCC are key regions involved in egocentric processing of one's own body's in spatial context (Vogeley and Fink, 2003). In contrast, meditators compared with controls showed thicker cortical thickness in lateral PFC, MPFC and temporal areas. Most of regions showing thicker cortical thickness in meditators are related with emotional processing. The MPFC is involved in emotional regulation through the coordination of behavioral and autonomic responses, which is critical for social adaptation. A number of behavioral and functional studies of meditation have previously reported the numerous beneficial effects of meditation on emotional regulation (Lutz et al., 2008; Jung et al., 2010). Interestingly, the observed structural differences of the MPFC in meditators are consistent with our previous findings that demonstrated greater MPFC functional connectivity in the DMN during a resting-state (Jang et al., 2011). Thus, these results indicate that the MPFC in mediators shows both structural and functional differences. Meditation practitioners showed thicker cortical thickness in the region adjacent to the primary motor area as well as a positive trend between cortical thickness in this region and practice duration. This region is involved in the control and execution of voluntary motor functions (Toma et al., 1999). It can be thus speculated that the rhythmic actions of the BWV protocol as well as the focusing of attention on the actions may be associated with the correlation that was observed in this study. We also found greater cortical thickness in the superior frontal cortex. The superior frontal cortices are regions of the brain that are typically involved in the regulation and monitoring of attention (Tang et al., 2007). Recent functional MRI studies have reported greater activation in these regions in experienced meditators during meditation (Brefczynski-Lewis et al., 2007; Hölzel et al., 2007). Taken together, it can be suggested that the meditation training we used, BWV, is involved primarily in attention processing and self-perception and secondarily in emotional processing.

Another alternative explanation for present findings is that each component involved in meditation training may affect a different part of the brain in different ways. A recent study demonstrated specific correlations between the executive and alerting components of attention and cortical thickness in anatomically relevant regions, showing positive correlations between the executive control component of attention and thickness in the ACC extending into the frontal pole and dorsolateral PFC, lateral PFC, inferior frontal cortex and temporal areas as well as negative correlations between the alerting component of attention and thickness in the lateral and medial aspects of the parietal areas (Westlye et al., 2011). Based on the findings from previous studies, we speculate that brain regions showing significantly thicker cortices in meditators than controls, such as the PFC and temporal areas, may be associated with their improved executive control. Meanwhile, brain regions showing significantly thinner cortices in meditators than controls, including the lateral parietal cortex and precuneus, may be associated with their increased vigilance. Further studies examining direct relationship between neuropsychological test performance and GM structures in meditators are warranted to clarify these hypotheses.

Interestingly, the majority of brain regions showing significant group differences in cortical thickness correspond to the region known as the DMN. The DMN is known to be involved in internal mentation or attention that is detached from the external world (Buckner et al., 2008), which is consistent with the meditative state utilized in the present study. The DMN consists of two subunits: the anterior DMN (aDMN) and the posterior DMN (pDMN) (Damoiseaux et al., 2008; Otti et al., 2010; Kim and Lee, 2011). The aDMN comprises primarily the MPFC, anterior cingulate and the precuneus, whereas the pDMN mainly includes the precuneus and the inferior parietal cortex. The area that showed greater cortical thickness in meditators mostly coincided with the aDMN, while the area with lesser cortical thickness was consistent with parts of the pDMN. As mentioned above, greater cortical thickness in the aDMN could be interpreted as enhanced regulation of emotional state, while thinner cortical thickness in the pDMN could be interpreted as improved self-referential processing. Given that the regulation of attention and emotion and self-referential processing is the central commonality across various meditation methods, our results suggest that meditation practice is associated with structural differences in regions that are typically activated during meditation and in regions that are relevant to the practice of meditation. However, as an adjunct to the cross-sectional design of this study, our findings should be confirmed using a longitudinal design with measures taken before and after meditation. Additionally, future studies using diverse data including neurohormonal and genetic factors and multimodal brain images would be helpful to clarify the mechanism underlying structural changes associated with meditation practices.

The effects of meditation observed in the present study may be associated with the plastic nature of the central nervous system, which allows the adaptation of existing neural connections and consistent neurogenesis in order to accommodate new information and experiences. FA values from the DTI may be considered the capacity of information flow within and between the brain regions. The results in the present study support this idea, as our results showed higher FA values in distributed regions in meditators, including the lateral PFC, MPFC, temporal pole, middle and inferior temporal areas and part of the parietal lobe, which is consistent with the results of other recent DTI studies. Tang et al. (2010) reported significant FA changes in the anterior corona radiate, and Luders et al. (2011) reported pronounced structural connectivity in the uncinate fasciculus as well as in the corticospinal tract and the temporal component of the superior longitudinal fasciculus. The uncinate fasciculus is a major WM tract connecting the anterior temporal lobe with the MPFC/OFC area. In the present study, higher FA values in the MPFC and temporal pole in meditators may be linked to their thicker cortical thickness in those regions. It is, thus, conceivable that meditation training might mediate information flow between these structures, which is implicated in emotional processing. Increased FA values in the dorsolateral PFC and part of the parietal lobe (Supplementary Figure S5), the regions involved in the so-called task-positive network (TPN), in the meditators is consistent with previous findings showing increased FA values in the superior longitudinal fasciculus, the main fronto-parietal fiber tract (Luders et al., 2011). Such increases in FA values may be associated with enhanced executive attention in meditators. Therefore, we suggest that meditation training can affect in both the DMN and TPN. The mechanism underlying the structural differences observed in the present study still remains largely unknown. However, our group recently found higher plasma dopamine levels in BWV practitioners compared with controls (Jung et al., 2010). The increase of dopamine associated with meditation training, through improved antioxidant status (Sharma et al., 2008), may contribute to dopaminergic neurogenesis. Interestingly, we also found reduced FA values in the anterior medial part, the ventromedial OFC and the posterior medial part, the PCC, in meditators compared with controls. These two regions are known as the key nodes of the DMN and are considered to act as important connector hubs in the brain (Hagmann et al., 2008). A possible explanation is that the decreased FA value may be due to increases in the number of crossing fibers in these regions as the hub. In this study, meditators compared with controls had increased FA values in most of brain regions. Based on our results, we speculate that meditators have increased crossing fibers in the hub regions and it may result in decreased FA values in the regions. Future studies are needed to reveal the underlying mechanism of such FA changes.

There were no significant correlations between brain structure and practice duration, except for the region adjacent to the primary motor area. Previous studies have reported inconsistent findings, showing positive correlation between GM volumes and training duration (Hölzel et al., 2008; Grant et al., 2010) or no correlations (Luders et al., 2009; Vestergaard-Poulsen et al., 2009). Luders et al. (2011) described that the lack of any significant correlations between brain structure and the amount of meditation experience might result from the confound effect of age and the inaccuracy of the indicators chosen for determining the actual extent/intensity of the individual training. That is, extrapolations of training duration are subjective rather than precise. In addition, even if all meditators practiced in one particular style, they are engaged differently in their mental exercises. Based on previous studies, training intensity may more heavily affect GM structures than the total number of years of practice. Clearly, longitudinal studies will be necessary to determine whether the differences of FA values and cortical thickness in meditators we found were actually induced by meditation, or whether they are inherent prerequisites for the beginning and continuation of meditation (Luders et al., 2009, 2011).

The present study has several limitations. First, this study was not a longitudinal design but a cross-sectional design study. As a result of the cross-sectional nature of this study, it is unclear that the brain structural differences we found were directly caused by meditation training. Second, we used DTI images of 12 diffusion directions. Recent studies suggested that larger number of diffusion gradient directions and using isotropic voxels may improve the estimation of the diffusion tensor (Ni et al., 2006). Future directions should include longitudinal studies with more directions of the diffusion gradient.

In summary, this study used cortical thickness and DTI to demonstrate structural differences in long-term meditators. Meditators showed thicker cortex in the anterior parts of the brain such as the lateral PFC, MPFC and temporal areas, and thinner cortex in the posterior parts of the brain, including the lateral and medial parietal regions. Additionally, DTI results showed greater WM integrity in the MPFC. Our findings demonstrate meditators, compared to controls, have structural brain differences in both GM and WM. These brain structural differences, particularly in the frontal cortex, may be associated with repeated practice of attentional and emotional regulations.