MODULATION OF THE INPUT OUTPUT FUNCTION OF RAT PIRIFORM CORTEX PYRAMIDAL CELLS/

1. In transverse brain slice preparations of rat piriform cortex, we c haracterized the repetitive firing properties of layer II pyramidal ce lls in control conditions (n = 78) and during perfusion of the choline rgic agonist carbachol (n = 26), with the ultimate goal of developing realistic computational simulations of the cholinergic modulation of t he input/output function of these neurons. The response of neurons to prolonged (1 s) intracellular current injections was examined at a ful l range of current injection amplitudes, providing three-dimensional p lots of firing frequency versus current amplitude versus time. 2. All neurons showed adaptation in response to intracellular current injecti on, with repetitive generation of action potentials at frequencies tha t were highest at the onset of the pulse and that decreased considerab ly thereafter. Substantial differences were observed between cells wit h regard to their rates of adaptation and the maximal number of action potentials they could generate during the current pulse. 3. The adapt ation characteristics of each neuron were quantified by plotting the n umber of action potentials generated in 1 s as a function of the norma lized current injection amplitude and measuring the area beneath this plot of the number of spikes versus current injection amplitude (S-I p lot). This value was termed S-I value and allowed neurons to be plotte d on a continuum including neurons showing strong adaptation (S-I valu e <8.0) and neurons showing weak adaptation (S-I value <8.0). The grou p showing weak adaptation contained 36% of the cells in control soluti on and 93.8% of the cells in 20 mu M carbachol. 4. Neurons showing str ong adaptation did not differ significantly from neurons showing weak adaptation in control conditions in measurements of resting potential, input resistance, threshold, and spike amplitude. Only a small differ ence was found in frequencies of firing measured soon after pulse onse t (after 100 ms). This implies that differences in S-I values are prim arily due to different rates of adaptation in later parts of the respo nse. 5. Perfusion with solution containing the cholinergic agonist car bachol (2-100 mu M) or 0 Ca2+ and 200 mu M cadmium resulted in a subst antial increase in the S-I values of neurons showing strong adaptation but had only a small effect on their initial firing rates. The effect on weakly adapting cells was smaller. In the presence of 20 mu M carb achol, neurons showed a distribution shifted predominantly toward weak adaptation (n = 26). 6. On the basis of these findings we constructed two computational models of pyramidal neurons: model 1, showing stron g adaptation; and model 2, showing weak adaptation. These models diffe r in the maximal conductance of the M current and the afterhyperpolari zation current. The S-I values of these two models provided a represen tation of typical values in the population of neurons in control solut ion and the population of neurons recorded during cholinergic modulati on. Use of these single-cell simulations allows a more realistic repre sentation of the effect of acetylcholine on the input/output function of neurons in a network biophysical simulation. This allows exploratio n of the role of acetylcholine in associative memory function, as desc ribed in the accompanying paper.