- The paper performed a meta-analysis of 31 functional neuroimaging studies involving 527 participants using Activation Likelihood Estimation to identify consistent brain activity patterns during meditation.
- Key results show distinct brain activity patterns for four main meditation styles: Focused Attention, Mantra Recitation, Open Monitoring, and Loving-Kindness/Compassion meditation.
- These distinct neural patterns suggest different underlying psychological mechanisms for various meditation practices, supported by medium effect sizes across activations and deactivations.
This paper explores the functional neuroanatomy of meditation by reviewing and analyzing functional neuroimaging studies. Meditation is seen as a diverse set of mental practices with distinct techniques and goals. The paper seeks to understand if these different meditation styles have unique patterns of brain activity.
The authors reviewed 78 functional neuroimaging studies on meditation. They used a method called activation likelihood estimation to analyze data from 31 experiments involving 527 participants. The goal was to find consistent patterns of brain activation and deactivation associated with different meditation techniques.
The study focused on four common meditation styles:
- Focused attention
- Mantra recitation
- Open monitoring
- Compassion/loving-kindness
The central hypothesis was that meditation practices that are different at the psychological level would also show different patterns of brain activity.
The authors also looked at three other meditation styles: visualization, sense-withdrawal, and non-dual awareness practices, but there wasn't enough data to do a full analysis.
Methods
The authors searched databases like MEDLINE, Google Scholar, and PsycINFO for papers containing the word "meditation". They refined the search to include studies that used functional neuroimaging methods like functional magnetic resonance imaging (fMRI) or positron emission tomography (PET).
Studies were included if they:
- Reported specific peak locations of brain activity in Talairach or Montreal Neurological Institute (MNI) space. MNI space is a standardized coordinate system used to locate brain structures.
- Had a reasonable sample size.
- Involved participants who actually practiced meditation.
- Involved meditation during the brain scanning session.
- Included healthy, non-clinical participants.
- Were published in peer-reviewed scientific journals.
The included studies were then categorized by two independent coders based on the type of meditation practiced: focused attention, open monitoring, compassion/loving-kindness, or mantra recitation.
The meta-analysis used a method called activation likelihood estimation (ALE) to find areas of the brain that consistently showed activation or deactivation during each type of meditation. ALE is a statistical technique used to identify areas of the brain that show consistent activity across multiple studies. The ALE method identifies areas where there is a higher-than-chance clustering of activation foci, suggesting a reliable effect.
Here's how ALE works:
- For each study, the locations of peak brain activity are identified. These locations are represented as coordinates in a standard brain space (like MNI space).
- Around each coordinate, a 3D Gaussian distribution is created. This distribution represents the uncertainty in the exact location of the activation. Studies with larger sample sizes have narrower Gaussian distributions, reflecting more precise localization.
- The union of these 3D Gaussian distributions is computed to create a statistical map. Each voxel in the map represents the likelihood that activation occurs at that location.
- The resulting statistical map is then compared to a null distribution that includes the same number of peak foci distributed randomly throughout the brain. This comparison identifies clusters where convergence between foci of activation or deactivation is greater than would be expected by chance.
The resulting statistical maps were thresholded using a false discovery rate (FDR) of q = 0.05 and a cluster threshold of k = 100 mm3. The FDR is a statistical method used to correct for multiple comparisons and reduce the number of false positive results. The cluster threshold is a minimum size requirement for a group of activated voxels to be considered statistically significant.
In addition to the ALE meta-analysis, the authors also performed an effect size meta-analysis. Effect size is a measure of the strength of a relationship between two variables. In this case, the variables are meditation practice and brain activity. Effect sizes were calculated using Cohen's d, which represents the difference between two group means in standard deviation units.
To address potential publication bias, the authors constructed a funnel plot, which is a scatterplot of effect size against sample size.
Results
The meta-analysis found different patterns of brain activity for each of the four meditation styles.
- Focused Attention Meditation: This was associated with activation in the premotor cortex and dorsal anterior cingulate cortex, areas involved in attention and cognitive control. Deactivations were observed in the posterior cingulate cortex and inferior parietal lobule, regions of the default mode network associated with mind-wandering.
- Mantra Recitation Meditation: This showed activations in the premotor cortex, supplementary motor area, putamen/lateral globus pallidus, fusiform gyrus, and cuneus. A deactivation was observed in the anterior insula.
- Open Monitoring Meditation: This was associated with activation in the insula, inferior frontal gyrus, pre-supplementary motor area, and supplementary motor area. A deactivation was observed in the thalamus.
- Loving-Kindness and Compassion Meditation: This showed activations in the anterior insula, parieto-occipital sulcus, and somatosensory cortices/inferior parietal lobule.
The effect size meta-analysis found medium effect sizes for both activations and deactivations across all meditation practices, suggesting practical significance.
Discussion
The results suggest that different meditation practices have different effects on brain activity.
- Focused attention meditation seems to involve brain regions related to cognitive control and reducing mind-wandering.
- Mantra recitation meditation may rely on motor control networks and reduce sensory processing.
- Open monitoring meditation appears to engage areas involved in sensory awareness and cognitive control.
- Loving-kindness and compassion meditation may activate areas related to empathy and emotional processing.
The authors note that there was convergence across multiple techniques in brain areas such as the insula, pre/supplementary motor cortices, dorsal anterior cingulate cortex, and frontopolar cortex.
The effect size meta-analysis showed medium effects for both activations and deactivations, suggesting that these changes in brain activity could be meaningful.
The authors also discussed several limitations of their meta-analysis:
- The classification of meditation practices into distinct categories may not capture the full complexity of these practices.
- The number of studies included in the meta-analysis was relatively small, especially for some meditation categories.
- The study methods and participant samples were highly variable across studies.
- There are challenges in calculating and interpreting effect sizes in neuroimaging data.
