Weekly Digest — Aug 16, 2026

The 2 big things in Spirit Tech this week: Two peer-reviewed papers, days apart, reached the same two-sided verdict on EEG meditation and neurofeedback technology. A Frontiers in Artificial Intelligence study showed that machine learning classifiers can distinguish Shamatha, Vipassana, and Metta meditation from mind-wandering with high accuracy when trained on a single meditator's EEG data. Across different meditators, accuracy dropped sharply. The same week, a Frontiers in Neuroscience systematic review took stock of the wider field: 25 studies since 2021 of closed-loop EEG neurofeedback, applied to ADHD, PTSD, chronic pain, and anxiety. Its verdict was two-sided. The systems technically work, decoding the target brain rhythms and delivering live feedback, and small trials reported real symptom drops, pain relief in most patients in one study and reductions in PTSD and ADHD scores in others. But almost every study was a pilot of ten to thirty people, few used sham controls, and only two followed patients past twelve weeks, so durable clinical benefit is unproven. The frontier for consumer neurofeedback is clear: it works and the early therapeutic signals are encouraging, but the technology cannot yet generalize across users or show its effects last. The endorsement came from Dr. Rachel Yehuda, who directs Mount Sinai's Parsons Research Center for Psychedelic Healing, ran the first MDMA-assisted therapy trial inside the VA system and trained its therapists by hand, on top of thirty years establishing that trauma leaves biological marks that pass to the next generation. She is as authoritative as the field gets. Her group is now working with Fireside on Lucy, an AI that plays a patient taking MDMA or psilocybin so trainees can rehearse the hardest moments and be scored, against scales like the Therapist Empathy Scale, before they ever sit with a real person. It learns from Fireside Project's 7,000-plus real conversations with people in the middle of psychedelic experiences, which is why her one caveat is about the data: "what Lucy gets fed determines how well Lucy is going to perform." The bottleneck in psychedelic therapy is no longer the drugs, it is the people, and therapists are still trained the slow way, cohort by cohort, by the few qualified to teach it. That makes the training layer a clear opening in psychedelic therapy for Spirit Tech, with a catch Dr. Yehuda names herself: the tools are arriving before the field has agreed what good practice is, and "we're not even sure what we want to scale yet."