In this paper we present a 2-D silicon cochlea which includes an automatic quality factor control (AQC) loop. This control-loop is an improved version of that presented in [1]where the control-loop imposes both a ceiling and a threshold level on the output amplitude of the basilar membrane (BM) resonators. In this improved version we include only a single set-point in our control-loop. This allows us to tune the BM resonators close to a Hopf bifurcation. We present test results from a fabricated integrated circuit which, when compared with biological data, demonstrates the feasibility of our active 2-D cochlea model.
I. BACKGROUNDThe outer hair cells (OHCs) are instrumental in the active behaviour of the biological cochlea [2]. Many researchers have shown that the OHCs poise themselves on or close to a dynamical nonlinearity known as a Hopf bifurcation and that this is the source of the nonlinear amplification observed in the biological cochlea ([3], [4], [5] and [6]).In [1] we attempted to replicate the system-level behaviour of the OHCs by means of AQC.This system was implemented using a threshold level (with hysteresis) and a ceiling level to set the maximum output amplitude of the BM resonators. The control-loop increases the Q-value of the BM resonator when its output amplitude is below the threshold level and decreases the Q-value when the output amplitude is above the ceiling level. The Q-value remains unchanged when the output amplitude of the BM resonator lies between the threshold and the ceiling. A simplified graph showing the operation of AQC is shown in Figure 1.The AQC operation in [1] is similar to that found in cochlear implants and hearing aids where the loudness of the output signal is set to be within a comfortable range (Target Amplitude in Figure 1) for the user. While this is acceptable for a hearing prosthesis it does not allow us to easily set the AQC loop at a Hopf bifurcation point as the amplitude of the output from the BM resonators does not increase once it is above the threshold level. Hence, we are unable to obtain, to the same extent, the large-signal compression that can be obtained by the biological cochlea. It also appears that this solution is not biologically plausible. The nonlinear, active effects of the OHCs do not have much influence on the operation of the cochlea until the pressure-wave input signal is below some set-point. This would suggest that the mechanical response of the basilar membrane inhibits the action of the OHCs at amplitudes above the set-point [7]. The set-point can be altered by higher levels of the auditory brainstem by-wayof efferent fibres allowing hearing to adapt to both noisy and quiet listening environments [4].
Figure 1 Original AQC operationIn this paper we have improved our AQC model so that it matches biology more closely. Specifically we have removed the threshold and ceiling levels from AQC operation and replaced them with a single set-point level that is adjustable. BM resonators including the improved AQC circuitry were connected together ...