Plaza scale is a critical determinant of human comfort in urban environments. However, traditional design guidelines often prescribe fixed Length-to-Height (L/H) ratios (e.g., 2:1 or 3:1) without accounting for the dramatic vertical variations in modern high-density cities. This study bridges neuroscience and urban physics to investigate how building height modulates the optimal plaza scale for psychological adaptation. Using Virtual Reality integrated with neurophysiological metrics (Electroencephalography and Galvanic Skin Response), we measured the immediate responses of 42 participants across 26 spatial configurations with building heights ranging from 3 m to 100 m. Contrary to conventional theories, our results reveal that the optimal L/H ratio is nonlinear and strictly height-dependent. Quantitative analysis derived an adaptive scale model, \(\frac{L}{H} \approx 1.43 \times e^{-0.057H} + 1.92 \qquad (R^2 = 0.92),\) demonstrating a decaying requirement for openness as height increases. While low-rise environments (\(H \leq 9\,\mathrm{m}\)) favor larger ratios (approximately 3.0) for relaxation, high-rise environments (\(H \geq 54\,\mathrm{m}\)) require significantly tighter enclosures (approximately 1.9) to mitigate spatial stress. Neurophysiological data confirmed this divergence: in high-rise settings with excessive openings, participants exhibited increased Beta wave activity (cognitive load) and elevated physiological arousal. These findings challenge the “one-size-fits-all” approach, providing a geometry-based, data-driven framework for designing resilient public spaces in vertically growing cities.