In this proof-of-concept study involving expert Vipassana practitioners (n=34), we successfully decoded self-reported meditative depth using source-localized EEG activity in the theta, alpha, and gamma bands, revealing a remarkable accuracy in predicting these gradations in unseen sessions across two separate visits. Our finding suggests that neither conventional EEG channel-level methods nor an a priori chosen set of default mode network regions adequately captured the complex, non-linear neural dynamics associated with varying meditation depths, echoing the nuanced phenomenology characteristic of discrete meditative states. Additionally, we introduce "spontaneous emergence" as an effective experiential sampling method, where participants naturally report their meditative depth. This approach, more ecologically valid and less intrusive than traditional probing, yielded comparable decoding performance. These results hold implications for advancing neurofeedback techniques in meditation, indicating new directions for more precise neural patterning and computational methods, and enhancing our understanding and facilitation of meditative practices.