We derive seven falsifiable predictions for the transferred potential (correlated brain activity between isolated individuals) from a fractional non-local field model with α = 1.2. The model is calibrated against existing replication data (Radin 2004, r = 0.20 at 10m, p = 0.0005) and predicts: (1) power-law distance decay with exponent −1.8; (2) preferential low-frequency transfer (δ/γ ≈ 7:1); (3) sigmoidal meditation onset with sharp threshold; (4) electromagnetic shielding transparency; (5) bidirectional symmetry (novel, never tested); (6) distance-independent transfer latency (novel, never tested); (7) null effect for non-bonded pairs. A two-regime distance model predicts that correlations decay locally but persist at a constant floor beyond a coherence length of 1,000–10,000 km — explaining Grinberg-Zylberbaum's planned intercontinental experiment. Connections to recent demonstrations of room-temperature quantum coherence in fractal gels and to orchestrated objective reduction (Orch-OR) are discussed. Pre-registered predictions deposited at DOI: 10.5281/zenodo.18823295.