Best fits and standard errors of drift rate scale factor a for Av, using sliding four-day windows to extract trends. Spikes in lower two panels correspond to isolated session blocks that give very poor fits, cf. Figure 12. Excluding three such points for OU3, fits for ddm1, ddPow2, ddPow3, ddExp3 and OU3 (top to bottom) have average normalized standard errors 0.0393, 0.0601, 0.0946, 0.0752 and 0.1646, illustrating that fitting errors rise as parameters are added.