The Optical Headstage
The headstage design is proprietary, and we're not ready to say too much about its operation at the moment, but we would like to publish full technical details in an appropriate journal in due course. However, we do explain the basic operating principles here.
Figure 2 - Resistive Feedback Headstage.
The signal-to-noise problem with patch clamping is not just one of accurately passing very small currents - although standard feedback arrangements and components with very low leakage currents can look after that - but also of accurately measuring them. In the standard patch clamp headstage arrangement shown in Fig. 2, in which a feedback resistor is used to pass current, any difference between the command potential and the electrode potential causes a current to be passed through the resistor in such a direction as to oppose this difference. The current through the feedback resistor causes a potential difference to appear across it, which can be monitored by subtracting the command potential from the amplifier output. Unfortunately, resistors with values in the range normally used in electronic circuits, i.e. up to a few Megohms, generate levels of thermal noise (due to the random movement of electrons within them) that are large compared with the voltages that represent the patch currents. The noise voltage actually increases as the resistance increases, but only as the square root, whereas the voltages that represent the patch currents increase linearly with the resistance. The overall situation thus improves only with the square root of the resistance, and this favours the use of very high-value feedback resistors, of ten gigohms or more, but the effects of stray capacitance then become very important. Even minute amounts will significantly attenuate the high-frequency response, and practical circuits must be designed carefully so as to control and equalise these effects. For clarity, the equalisation circuitry is not shown in Fig. 2, because we're dealing with an ideal resistor in this model.
Figure 3 - Capacitive Feedback Headstage.
Although resistive headstages can be made to work well, the alternative method of using a capacitor to pass the patch currents has become a viable alternative. This method is illustrated in Fig 3. Capacitors don't generate thermal noise, so in principle capacitive (also called integrating) headstages are somewhat quieter than resistive ones, but other problems must be solved in order to achieve satisfactory overall performance. First, a steady current will cause the capacitor to charge, so some method of discharging it periodically must be included. Secondly, the output voltage is now the time integral of the current, and a differentiation must be performed in order to recover the actual current waveform. Such circuits necessarily have very high gain at high frequencies, so they must have both low noise and low susceptibility to interference pickup in order to realise the potential benefits of the capacitive technique. Finally, the performance of the capacitor itself must be exemplary. In view of the very high AC gain levels any less than ideal capacitor behaviour, such as non-linearity with voltage or microphonic pickup, become increasingly important. Although these problems have been overcome, we didn't relish the thought of going down that road ourselves.
This is where the optical technique comes in. We liked the idea not just because it was new and different - although that was certainly attractive - but because it seemed such a clean solution to the problem. This is not to say that the optical technique doesn't pose problems of its own, but rather that the nature of those problems doesn't threaten the main performance requirements of high bandwidth and low noise. The main practical problems to solve with the optical approach are those of calibration and linearity, which can be tackled successfully in ways that don't compromise the performance of the system in respect of noise and bandwidth.
Figure 4 - Optical Feedback Headstage.
In the Optopatch, currents are passed by shining light onto miniature photodiodes which are connected between the differential inputs of the headstage amplifier, as shown in Fig. 4. This method gives the advantage of steady current-passing that the resistive method provides, but in this case there is no resistance to contribute any (additional) thermal noise, so the theoretical performance is the same as the capacitive circuit. The light is produced by passing linearised control currents through light-emitting diodes, which are also incorporated within the headstage in order to achieve a compact and totally enclosed optical design, although the drive electronics are on the daughter board in the main enclosure in order to minimise the size of the headstage. By using different illumination pathways of different overall efficiency, we have been able to implement both the current-passing ranges of +/-1nA and +/-100nA full scale by the optical technique. Although the optical method doesn't offer any significant advantage over the resistive method in the higher current-passing range, it simplifies the system design to have both ranges operating in the same way, but we should emphasise again that the overall system design can in principle support other types of headstage as well.
In the circuit of Fig. 4, any difference between the command and electrode
potentials generates an error voltage at the output of the headstage amplifier.
This provides the input to the optical current-passing stage, causing a
photodiode current that tends to restore the input potential difference, and
the amplifier output voltage is a direct measure of the current. The operation
of the circuit is thus similar to a resistive headstage, except that the
command potential does not appear on the amplifier output, and so does not need
to be subtracted from it.
