## Figure P47

Here, qLi is the ipsp for the left spike source, a is the filter natural frequency in radians/ms, yL are the actual unit input impulses to the synaptic filter on the left input axon, and KLi adjusts the "gain" of the synaptic LPF. To make two equal, synchronous input spike sources for the system, will require two IPFM SGLs with dc inputs, one of which is shown below:

" LEFT IPFM Freq. source: dvL = eL - zL zL = yL*phi/dT

wL = IF vL > phi THEN 1 ELSE 0 sL = DELAY (wL, dT) xL = wL - sL

The firing threshold is set to phi = 1, and dT = 0.001, the Euler integrator delT. eL is a dc level so that the unit impulses, yL(t), will have constant frequency. The right-hand IPFM frequency source is made to increase slightly in frequency by adding a pulse of height dE to the dc input; thus, eR = E0 + dE[U(t - ti) - U(t - t2)].

Simulate the phase/frequency difference detector system. The model will have two IPFM frequency sources, two RPFM SGLs, two excitatory and two inhibitory synapses. The synaptic dynamics are all equal, as are all the SGLs. Let K = 600, phi = 1, Eo = 0.333, dE = 3.33E - 3, a = 0.2 r/ms (synaptic poles), c = 0.5 r/ms (RPFM poles). Observe over what steady-state range of fR = fL the system will reliably detect a 1% increase in fR (Eo adjusts the SS frequency).

4.8. Section 4.4.2 we examined the architecture of a BDHS. Figure 4.4-4 illustrates the response of a BDHS to a swept frequency input. Figure P4.8 illustrates a BDHS system.

a. This problem will determine this system's steady-state band-pass characteristic, i.e., fmax, fmin, and Q of the passband. Let Vin = Vo, (no ramp). Use the Simnon program, BDsupr1.t in the text with the following parameters: SD = 0, Kr = 0 (no noise, no swept frequency), phi1 = 1, phi3 = 0.30, c3 = 7 r/ms, a1 = 1, a2 = 2, a3 = 1, g1 = 0.8, g2 = 1, g3 = 1.50, D1 = 0.333 ms, D2 = 2 ms, Do1 = 1, Do2 = 1, Do3 = 1.2.

b. By manipulating D1, D3, and other system parameters, see how narrow a passband can be created. The center of the passband should be at about 30 pps.

4.9. This problem will examine the properties of a hypothetical model neural notch filter that blocks the transmission of incoming spikes having a certain range of frequency. The system is based on Reiss's (1964) "band suppressor" architecture. Figure P4.9 illustrates the neural notch system. The program is listed below with desired parameters.