Boltzmann analysis of CM waveforms using virtual instrument software

We describe a modification to our technique for the rapid analysis of low-frequency cochlear microphonic (CM) waveforms in the basal turn of the guinea pig cochlea (Patuzzi and Moleirinho, 1998). The transfer curve relating instantaneous sound pressure in the ear canal to instantaneous receptor current through the outer hair cells (OHCs) is determined from the distorted microphonic waveform generated in the extracellular fluid near the hair cells, assuming a first-order Boltzmann activation curve. Previously, the analysis was done in real time using custom-built electronic circuitry. Here, the same task is performed numerically using virtual instrument software (National Instruments LabVIEW 4.1) running on a personal computer. The assumed theoretical function describing the CM waveform is Vcm = Voff + Vsat/[1 + exp[(Eo+Z.Po.sin(2pi f + phi(tot)))/kT]], where the six parameters are (i) a DC offset voltage (Voff); (ii) the frequency of the sinusoidal stimulus (f); (iii) the phase of the sinusoidal stimulus (phi(tot)); (iv) the maximal amplitude of the distorted microphonic signal (Vsat); (v) the sensitivity of the transduction process (Z); and (vi) the operating point on the sigmoidal transfer curve (Eo). The software obtains the least-squares fit to the CM waveforms by continuously deriving the six parameters at a speed of about one determination per second. The independent fitting of the frequency and phase allows the data to be analysed off-line from data previously recorded to tape (i.e. the frequency and phase of the microphonic response need not be known accurately beforehand). We present here an outline of the software we have used, and give an example of the changes which can be monitored using the technique (transient asphyxia). The method's advantages and limitations have been discussed in our previous paper. The virtual instrument described here is available from the authors on request.