Mindfulness meditation: This ancient practice, now backed by contemporary science, offers benefits for mental and physical health. By directing attention to the present moment without judgment, mindfulness meditation helps cultivate a state of deep relaxation and heightened awareness.
In this study, authors delve into how even brief sessions of focused mindfulness meditation can alter brain activity, as evidenced through EEG and machine learning analysis. We’ll unpack the neural mechanisms behind mindfulness, revealing how short meditative practices can reduce stress and enhance cognitive function among young adults.
Mindfulness Meditation: Experiment Procedure
The meditation exercise lasted for 10 minutes total. Before doing the training, participants rested for around 10 minutes as well. In total, they collected data from 24 volunteers.
To avoid bias, they played the same audio recording of a qualified mindfulness instructor each time. You can read more about the exercise details in the original paper!
Before and after mindfulness training, experimenters collected pressure, hear rate and respiratory rate measures. During meditation, they used Smarting PRO – a mobile EEG device to collect brain activity.
Applying Deep Learning to Understand Effects of Mindfulness Training
Authors compared performance of three different learning models:
Multilayer Perceptron – MLP,
Long Short-Term Memory – LSTM and
Convolutional Neural Network – CNN.
The classification task involved determining from EEG whether a segment of data corresponded to a state of meditation or rest. The data was analyzed across various EEG frequency bands, such as delta, theta, alpha, beta, and gamma. To be specific, researchers derived power spectral density – PSD in these bands.
Here’s the overlook of the models and steps in training:
To further test the deep learning models, authors defined an ensemble where they combined different capabilities. To train and test the models, dataset was split into: training, validation and test sets.Let’s take a look at the accuracy results for all three models:
Performance of deep learning models in distinguishing rest vs. mindfulness meditation state.The ensemble model demonstrated a prediction accuracy of 77.0% on the test set. Meaning the approach of integrating spatial and temporal processing capabilities of CNN, LSTM, and MLP seems to show promise. Potentially this could offer enhanced EEG signal representation through combined model architectures.
Sensitive EEG Bands for Focused Mindfulness Meditation
Authors noted significant increases in alpha, beta, and gamma PSD during meditation compared to resting states. This can indicate a state of relaxed alertness beneficial for mental health.
The elevated beta activity appears in the prefrontal cortex, which is associated with heightened attention and relaxation. We can see the PSD topography for all frequency bands in mindfulness vs. resting states:
Brain topography of PSD across 8 bands in mindfulness vs. resting is depicted using a color-coded t-test scheme.Physiological Responses Associated with Mindfulness Meditation
Differences in blood pressure, heart rate and respiratory rate before and after focused mindfulness meditation.Post-meditation, participants showed significant reductions in heart and respiratory rates, suggesting a relaxation response. However, no significant changes were observed in blood pressure before and after meditation.
Another analysis authors did is investigate correlation between EEG and the rest of the physiological data. For this, they used Pearson correlation between channel PSD and changes in heart/respiration rate.
The study found positive correlations between the EEG PSD and changes in heart rate during the mindfulness meditation. The correlations were particularly notable in the frontal and right temporal regions, which are often associated with emotional regulation and stress response. This suggests that activity in these areas may be closely linked to physiological changes that occur during relaxation and meditative states.
Similar to heart rate, there was a positive correlation between changes in PSD and reductions in respiratory rate during meditation. This indicates that as participants entered deeper states of meditation, their breathing slowed. Which is a common indicator of deep relaxation and parasympathetic activation.
Final Thoughts and Future Work
Future steps might expand sample sizes by also incorporating control groups. Exploring meditation depths and practices can also deepen the understanding of mindfulness impact on EEG and physiological responses.
This research illuminates the potential of mindfulness meditation to foster a state of deep relaxation and improved emotional regulation. It also supports its use in therapeutic and stress management applications.
Original paper can be found here – https://www.researchsquare.com/article/rs-4962094/v1
Curious if this setup is right for your study?
Our team works with researchers across neuroscience, sports science and clinical applications - we'll help you figure out if mobile EEG fits your study design.
