To investigate the validity of the four PAC approaches, we constructed 1.2 s of simulated data with known alpha-gamma PAC [fp = 10 Hz; fa = 50–70 Hz; code adapted from Kramer et al. (2008) and Özkurt and Schnitzler (2011)] and added a random level of noise (signal-to-noise ratio >−11.5 dB). Comodulograms were produced using the four PAC algorithms on 64 trials of simulated data. Using the same code, we also investigated how the four algorithms were affected by trial length (0.1–10 s in 0.1 s steps, script: 7_simulated_PAC_analysis.m).
