Focus on the breath: Brain decoding reveals internal states of attention during meditation

Evidence suggests meditation may improve health and well-being. However, understanding how meditation practices impact therapeutic outcomes is poorly characterized, in part because existing measures cannot track internal attentional states during meditation. To address this, we applied machine learning to track fMRI brain activity patterns associated with distinct mental states during meditation. Individualized brain patterns were distinguished for different forms of internal attention (breath attention, mind wandering, and self-referential processing) during a directed internal attention task. Next, these brain patterns were used to track the internal focus of attention, from moment to moment, for meditators and matched controls during breath-focused meditation. We observed that while all participants spent the majority of time attending to breath (vs. mind wandering or self-referential processing), meditators with more lifetime practice demonstrated greater overall breath attention. This new framework holds promise for elucidating therapeutic mechanisms of meditation and furthering precision medicine approaches to health.