<html><div style='background-color:'><DIV class=RTE>
<P>First of all thanks a lot for the fast and clear reply!</P>
<P>I just looked at the short discription on the layer type you mention and I think I will be able to figure it out. It looks like I can actually use the same layer type for the input, because it is exactly what I do while generating the patterns. I'll try and let you know whether I succeed or not.</P>
<P>If I do so, I would probably also have to do it for incoming signals I use (other than the retinal input) which are scalars at the moment, won't I?</P>
<P>Okay, if the wizard indeed sets it up correctly, the non-learning in the recurrent setting was indeed because of the first problem.</P>
<P>With kind regards,</P>
<P>Frank<BR><BR></P></DIV></div><br clear=all><hr>Live Search, for accurate results! <a href="http://g.msn.com/8HMBENNL/2728??PS=47575" target="_top">Click here</a> </html>