g&@ &&#TNPP2OMi & TNPP &&TNPP   @ - "-- !@-- "-& 4& - @Times New Roman- . 2 1.̙- "-NF----A-- @Times New Roman- .2 3 Introduction  .-- s -- @Times New Roman- . 2  .@Times New Roman- .d2 &>Previous work has shown that action potential generation is a                   . .'2 reliable phenomenon (     . .2 Mainen   . . 2 P& . .2 k Sejnowski   . .2 1995). Repeated   . .g2 @presentations of the same fluctuating input yield patterns of sp                   . .2 iking   . .32 that vary by less than 2 ms.        .@Times New Roman- . 2  .@Times New Roman- .f2 &?Here we use the in vitro slice preparation of rat prefrontal co                   . .2 rtex to   . .d2 )>study the dependence of this reliability on the amplitude and                . .g2 H@frequency content of the inputs, and the eventual differences in              . .=2 f$reliability between pyramidal cells          . .2 f} interneurons  . . 2 f. .@Times New Roman- . 2  .@Times New Roman- .f2 &?We also construct a compartmental model of pyramidal cells and                   . .2  interneurons   . .`2 ;with intrinsic noise in order to reproduce and explain the                 . .2 data.  .@Times New Roman- . 2  .@Times New Roman- .f2 &?Information transmission was then analyzed using optimal linear                . .02 reconstruction techniques.     .-- "System-&TNPP &