(a) Schematic of the proposed model. (b) Example of the decision stage for N₀S_π stimuli at different signal levels. The top left shows the absolute value of the instantaneous IPD over time. Increasing the target level (with an IPD of π) increases fluctuations and thus the mean absolute IPD (y-axis top right). Taking the cosine of the mean IPD results in smaller output values with increasing IPD fluctuations. The decision variable D results directly from the cosine, subject to a Fisher’s-Z transform, so that lower values indicate a stronger signal prevalence (bottom right).