GRRT (left panels) and convolved images (middle and right panels) for our best-fit models, using either the phenomenological prescription Θ_e = Θ_e(R_low, R_high, β) (first row) or the self-consistent prescription Θ_e = Θ_e(σ, β) (bottom row). The panels in the central column correspond to synthetic images at 230 GHz convolved with a beam with radius of 20 μas, mimicking the observational resolution for M87*, while in the right column, are reported the images at 345 GHz, with a smaller beam of 10 μas.