SINGLE NEURON LOCAL RATIONAL ARITHMETIC REVEALED IN PHASE-SPACE OF INPUT CONDUCTANCES

We present a phase space analysis to explore the potential of single n euron local arithmetic operations on its input conductances. This anal ysis was conducted first by deriving a rational function model of loca l spatial summation by using the equivalent circuits for steady-state membrane potentials. it is shown that developed functional phases exis t in the space of input conductances, where a single neuron's local op eration on input conductances can be described in terms of a set of we ll-defined arithmetic functions. It is further suggested that this sin gle neuron local rational arithmetic is programmable, in the sense tha t the selection of these functional phases can be effectively instruct ed by presynaptic activities. This programmability adds the aegree of freedom in a single neuron's ability to process the input information.