Preference for small objects with the same type of synaptic transmission function for the excitatory and inhibitory inputs of the FD-cell.

<p>A FD-cell response of the DDI model. In this model variant both the inhibitory and excitatory signal are transformed by an expansive synaptic transmission function. For both signals this function is the same (except for a constant scaling factor). The major difference is that the inhibitory signal is spatially integrated <i>before</i> applying the synaptic function, whereas the excitatory signal is integrated <i>afterwards</i>. The figure is to read from bottom to top. For illustration, imagine the integrating operation as an averaging of the input values x_i (in B,C,D top shown as spatial activity profile, i indicates the position) and the synaptic function as an expansive nonlinearity <i>f</i>() (grey curve, B,C,D top). Let the output be the difference between the excitatory signal and the inhibitory signal <i>f</i>(<i>x̅</i>). For a homogeneous activity profile without an object (B bottom) the output is zero because (B top). If some points peak out in the activity profile (C bottom), indicating a small object, the excitatory signal is bigger than the inhibitory signal <i>f</i>(<i>x̅</i>) (C top), the difference (grey area) is large. When the object covers almost the whole receptive field (D), the output (grey area) is larger than it is without an object, but is reduced compared with the smaller object. (A) Resulting response of the FD-cell as a function of object size.