Fig. 7
From: Responses of Leaky Integrate-and-Fire Neurons to a Plurality of Stimuli in Their Receptive Fields
Model control. (a) QQ plots of the uniform residuals calculated using the transformation in Eq. (23) for the four methods fitted on single stimulus data and a grid size of \(\Delta t=0.002~\mbox{s}\) and \(\Delta x = 0.02\). The uniform residuals are pooled together from all 100 repetitions of the simulations. The bias is different for PDE methods and for IE methods, seen from how the points deviate from the identity line. (b) QQ plots of the uniform residuals of the probability-mixing (PM) model and the response-averaging (RA) model fitted on data simulated from both models responding to a stimulus mixture. For example, RA fitted on PM data means fitting the response-averaging model to data simulated from the probability-mixing model. From the QQ-plots a wrong model can be rejected