Abstract

In their response to our article (Haferlach, Wessnitzer,
Mangan, & Webb, 2007), Vickerstaff and Di Paolo
correctly note that the response function of input units
used to evolve our network was the same cos(ha – hp)
as that used by Vickerstaff and Di Paolo (2005), and
resembles the POL neuron arrangement in insects
(Labhart & Meyer, 2002) only in the use of three
instead of two such units. However it is important to
note that our evolved network structure—which maintains
a population encoding of the home vector over a
set of memory neurons that integrate the input coming
from the direction cells—is in fact generalizable to a
wide range of direction cell response functions. To
demonstrate this point, we here show the results of
integrating this network with a very recent model of
e-Vector orientation coding in the central complex of
the insect brain (Sakura, Lambrinos, & Labhart,
2008)