INSTRUMENTATION OF MEASURING CELLULAR CAPACITANCE VIA SIGNAL PROCESSING TOOL

This master thesis pertains to a systematic testing and evaluation of three neuroscience instruments previously developed at the Biomedical Engineering Laboratory of the University of Rhode Island. The three instruments are the Universal Clamp (UC), the Neuron Emulator (NE), and the Cell Capacitance Emulator (CCE). The UC is an innovative instrument that employs a fast-digital signal processor to deliver and integrate various experimental tools for electrophysiology. The NE is an electronic device that uses a capacitance source to represent the passive and active electrical properties of a neuron. The CCE presents a dynamic resistance-capacitance model of a neuron with the capability of switching between two capacitances that have a small difference to represent the activity of a vesicle crossing the cell membrane. A platform for conducting a systematic testing on the functionality of the UC by use of the NE and the CCE was developed. The main objective was to explore the possibilities and drawbacks of using such a platform to test neuroscience instruments without the need for a wet experiment involving live neurons. The functions of the UC under evaluation included voltage clamp, current clamp, and cell capacitance measurement. The contributions of this research include the hardware and software improvements on the NE and CCE necessary for integrating them into the testing platform. An attempt to reduce of the rise time of the action potential spike by adding a low-pass filter at the output circuitry made the action potential waveform more realistic; however, the low-pass filter also reduced the speed for the feedback currents and resulted in an unsuccessful voltage clamp. The representation of the cell capacitance changes due to a vesicle activity on the order of 10 fF with a 2-ms pulse width was challenging. The parasitic capacitance from the wiring and the breadboard was often on the order of 10 pF. The existing algorithm based on short-time Fourier transform for detecting the vesicle capacitance was insufficient to detect such a small and fast capacitance change. In summary, this thesis research has demonstrated an instrumentation platform and testing methods for electrophysiological instruments using only electronic devices. While the system does not completely replace the need for using live neurons, many standard experimental protocols such as current clamp and voltage clamp can be tested in an efficient and effective way.

To make an intracellular account, the tip of a fine (sharp) microelectrode must be embedded inside the cell, with the goal that the film potential can be estimated.
Normally, the resting film capability of a sound cell will be -60 to -80 mV, and amid an activity potential the layer potential may achieve +40 mV. [3] Therefore, we are using the action potential simulation models in this study as it is the culture of the BME lab to consider environmental matters as major ethical manner.

Voltage Clamp
The voltage clamp strategy enables an experimenter to the cell voltage at a picked esteem. This makes it conceivable to gauge how much ionic current crosses a cell's layer at any given voltage. This is essential in light of the fact that huge numbers of the particle   which it gets from the signal generator. This signal is enhanced and returned into the cell through the chronicle terminal, which is known to be the recording electrode. This methodology is still being used with the exception that a small axon needs to be used, which cause major difficulty during experiments. The single electrode method of the Universal clamp solves this issue as it will be explained in detail in section 2.6.

Current Clamp
The current clamp method records the layer potential by infusing current into a cell through the account cathode. Not at all like in the voltage clamp method, where the film potential is held at a level dictated by the experimenter, in mode the layer potential is allowed to shift, and the enhancer records whatever voltage the cell produces without anyone else or because of incitement. This method is utilized to consider how a cell reacts when electric current enters a cell; this is imperative for example for seeing how neurons react to neurotransmitters that demonstration by opening layer particle channels.
Most current clamp amplifiers give practically no intensification of the voltage changes recorded from the phone. The amplifier is really an electrometer, once in a while alluded to as a "solidarity pick up enhancer"; its primary reason for existing is to diminish the electrical load on the little flags (in the mV go) delivered by cells with the goal that they can be precisely recorded by low-impedance hardware. The speaker builds the current behind the flag while diminishing the opposition over which that present passes.
Consider this case in view of Ohm's law: A voltage of 10 mV is produced by passing 10 nanoamperes of current crosswise over 1 MOhm of obstruction.

Dynamic Clamp
Dynamic clamp uses a real-time interface between one or several living cells and a computer or analog device to simulate dynamic processes such as membrane or synaptic currents in living cells [4,5].

Patch Clamp
The Patch clamp method was created by Erwin Neher and Bert Sakmann who got the Nobel Prize in 1991. Customary intracellular chronicle includes skewering a phone with a fine terminal; fix clip recording adopts an alternate strategy. A fix brace microelectrode is a micropipette with a generally substantial tip breadth. The microelectrode is put by a cell, and delicate suction is connected through the microelectrode to draw a bit of the cell layer into the microelectrode tip; the glass tip frames a high opposition 'seal' with the cell layer. This setup is the "cell-appended" mode, and it can be utilized for concentrate the action of the particle directs that are available in the fix of layer. On the off chance that more suction is currently connected, the little fix of film in the terminal tip can be uprooted, leaving the anode fixed to whatever remains of the cell. This "entire cell" mode permits exceptionally stable intracellular account. A detriment (contrasted with traditional intracellular account with sharp terminals) is that the intracellular liquid of the cell blends with the arrangement inside the chronicle anode, thus some imperative segments of the intracellular liquid can be weakened. A variation of this system endeavors to limit these issues. Refer to the figure to see a sample of the patch clamp 4 different methods of measuring.

Universal Clamp
The universal clamp has a faster loop which is not limited by PC ability. It also uses a standard software, hardware is standard as well for control, windows which is not optimized for operation. Therefore, it is more insufficient. It was more customized, monitoring and data acquisition only, the front end implements the control; while the old system has only one end. The new system has both which separate and minimize cost, easier to repair, and it has smaller customized processors. The NE had a current input port and a voltage output port. In response to the input current (stimulation) the NE was able to adjust the firing rate of the action potentials accordingly. While the first-generation NE was able to present action potentials for the purposes of teaching and testing the neuroscience instruments, it could not be voltage clamp. The first-generation NE was based on a voltage source, which always produced a sufficient current to maintain the wave shape of the action potential and could not be cancelled out by an external feedback current.
Thus, in 2011 the development of the 2nd-generation NE was started with the intention that the NE is responsive to voltage clamping [3]. The circuit design was completed redone and centered around the concept of a capacitance source. Instead of using a voltage source or a current source, a capacitor is pre-charged and switched on at appropriate times to produce the action potential. Because the charge on the capacitor is limited, the action potential can be clamped by use of an external feedback current to cancel out the discharge of the capacitor. The design was novel and resulted in a US patent [4].

Neuron Emulator Innovation
The patent illustrated the following, 'this invention was concerned with an electronic neuron emulator for a single-electrode setting that has both the passive properties (membrane resistance and capacitance) and the active properties (action potential) of a live neuron. A unique design feature is that the currents used to generate action potentials come from a pre-charged capacitor. Unlike a voltage source or a current source, the charge on the capacitor is limited, thereby providing a more realistic condition for testing neuroscience instruments such as the single-electrode voltage clamp and the patch clamp.
Currently the standard device for testing electronic neuroscience instruments such as a voltage clamp amplifier is a simple resistor-capacitor (RC) circuit. While the RC circuit can represent the passive electrical properties of a live neuron, it cannot generate action potentials to interact with the voltage clamp amplifier in a dynamic way.
Previously we developed a neuron emulator that used an oscillator to generate the action potentials [1]. The oscillator was a time-varying voltage source, which was too strong to overcome by use of a voltage clamp amplifier. Thus, it was not possible to voltage-clamp the action potentials generated by our previous device. The neuron emulator in this invention is based on a totally different concept. An action potential is generated by switching a pre-charged capacitor into the output circuitry. Once the capacitor is discharged, it is switched out of the output circuitry and charged up for the next firing of the action potential. A microprocessor is embedded in the device to provide controls of the firing of the action potentials. This concept was first published in [2] on April 3, 2011.

Theory of Operation of the Neuro Emulator
The patent illustrated the following, 'this invention was concerned with an Vest. The switching operation is accomplished by three switches S1, S2, and S3.
These switches are controlled by a microprocessor. The output of the Rm-Cm circuit represents the membrane potential of the neuron and is sent to the analog-to-digital converter via an amplifier. Thus, the microprocessor can constantly monitor the membrane potential and adjust the firing rate accordingly. The output of the Rm-Cm circuit can also be accessed externally via resistor Ra as the output voltage Vout, where Ra represents the resistance of the electrode. The timing of the switching signals is done is such a way that the action potential is generated by turning S1 and S2 off and S3 on.
The figure further describes the switching of the Rap-Cap circuit in two stages.
Stage 1 represents the resting potential as S1 and S2 are on and S3 off. The Rap-Cap circuit is disconnected from the Rm-Cm circuit and is charged by Vpeak. Stage 2 represents the action potential as S1 and S2 are off and S3 on. The Rap-Cap circuit is connected in series with the Rm-Cm circuit. The output Vout momentarily jumps to Vpeak and then discharges. The firing of consecutive action potentials is accomplished by alternating between stage 1 and stage 2.
The aforementioned neuron emulator produces an action potential that has the waveform of a simple exponential discharge. This waveform can be further improved and made closer to a real action potential.

Preliminary Testing Results
The preliminary testing results are summarized here.   To measure a small capacitance, change on the order of 10 fF with a duration on the order of 1 ms is very challenge but possible.

Introduction to the Cell Capacitance Emulator
The cell capacitance emulator (CCE) is a simulation tool for testing a The specific contribution of this thesis is to construct a CCE on a breadboard and to improve its functionality.

Instrumentation of the Cell Capacitance Emulator
The smallest capacitance value for the standard capacitors available from the electronic suppliers is 4.7 pF, which is about 3 orders of magnitude larger than that of the vesicle capacitance (~10 fF). The idea behind the cell capacitance emulator is to

Figure 14 The schematic diagram of the CCE and the instrumentation setup with the Universal Clamp and the capacitance meter
The CCE outputs a dynamic RC circuit. The resistor Ra (1 M ) represents the cell membrane resistance. The capacitor presenting the cell membrane capacitance is switched between C1 and C2. Both capacitors are 10 pF; however, they are chosen such they there is a small difference between the them. The differential capacitance represents the vesicle capacitance. This dynamic RC circuit is used as the input to the Universal Clamp. The purpose of the Universal Clamp is to detect the vesicle activities as discrete events, not to measure the exact value of the vesicle capacitance.
The CCE also has a capacitance-only output, which is sent to a precision capacitance meter for determining the exact capacitance values of C1 and C2.  The CCE has been implemented on a breadboard as shown in  Measuring this vesicle-induced capacitance change is possible but very challenging.

Existing Method
Recent studies have shown the possibility of measuring this capacitance signal.

Figure 16 High-resolution membrane capacitance measurements for the study of exocytosis and endocytosis, adapted from Rituper et al. [2]
The instrumentation system was rather complicated that required multiple instruments including a patch-clamp amplifier and a lock-in amplifier. Signal processing techniques such as signal averaging were used to improve the signal-tonoise ratio. The requirements of post (off-line) processing make this method difficult for real-time applications.

A Frequency-Domain Method
The current version of the Universal Clamp has a preliminary vesicle detection algorithm, which was implemented by Dr. Jiang Wu. The algorithm was based on the short -time fourier transform [3] as described below.

A Time-Domain Method
An innovative algorithm has been proposed to perform the vesicle capacitance detection by taking advantage of the Universal Clamp hardware and by optimizing the signal processing tactics [5]. The concept of the lock-in amplifier (using sinusoidal modulation and demodulation) is integrated into the detection algorithm. The Universal Clamp delivers a sinusoidal current excitation and performs a matched filter on the induced voltage, which has the same noise-reduction effect of an lock-in amplifier but without actually using one. The intended input and output waveforms are shown in Figure 5.4.1.

Figure 17 illustrates the concept of a switching excitation with sinusoidally amplitude-modulated current injection and an induced voltage that has a sinusoidal envelope with a phase shift
As shown in Figure 5       properly connected, where # is a unique ID number for the specific Universal Clamp controller board.

1.
Full screen -This is the Windows option for expanding the window to full screen. The ESC key reverses from the full-screen mode to the standard mode.

2.
Voltage window -This window is used to monitor the voltage measurement in real time.

3.
Current window -This window is used to monitor the injection current in real time.

4.
Network activities -This is the message window that shows the activities of the network communications between the Universal Clamp and the computer.

5.
Packet queue status -The data from the Universal Clamp are segmented into packets. A queue is used to buffer the packets. 18. Pause/play graphic display -This button is used to freeze/unfreeze the display.
19. Record/stop data to disk -This button controls the storage of data to a disk file called record.dat in the DCO_v104 folder.
desirable digital filter. See Section 7 Digital Filter Bank for more details.

Digital Filter Bank
Each filter was implemented as a 60 Hz notch filter combined with a low-pass filter. The choice of the cut-off frequency for the low-pass filter includes 1 KHz, 2 KHz, 3 KHz, 4 KHz, and 5 KHz, as shown. The filters were high-order finite-impulseresponse (FIR) filters. They don't have a sharp cut-off response but are always stable and linear phase filters. See Figure 6.5.1 below.

Figure 25 Digital Filter Bank
The example below in figure shows that the filter with the 1 KHz cut-off frequency reduces the noise level from 2 to less than 1 mV peak-to-peak.   Presently, there are 3 ranges for current injection: 6 nA, 60 nA, and 300 nA.
The default is 60 nA, which means the maximum allowable current injection is 60 nA.
To change the current range, not only the Current Range needs to be changed but also two specific jumpers on the circuit board needs to be changed. A attempt to change the current range results in a pop-up window as shown in Figure 6.7.2 below. This is to remind the testing user the jumpers on the circuit board inside Universal Clamp need to be set and match the current range setting.   Generally speaking, the P gain is the most sensitive parameter on the PID controller's performance. The integral gain can further control the error at the DC level. The derivative gain can further improve the response at higher frequencies.
The desirable PID gains can be entered and then downloaded to the Universal Clamp by clicking the Set PID Params button. For example, in order to suppress the initial oscillation of the voltage clamp shown above, the Proportion gain is changed to P = 5.0. The result in Figure 6.8.2 below shows a reduction of oscillation and a significant shorter settle time at the onset of the voltage clamp.   The rise and fall of the spikes from this neuron emulator are at least 10 times faster than those from the real neurons. Thus, the response time of the Universal Clamp system for voltage clamping is considered adequate.

Voltage Clamp With Modified Neuron Emulator
As described in Section 3.5.2, an attempt to slow down the initial rise time of the action potential generator by the Neuron Emulator was proposed by adding a lowpass filter at the output. The resulting action potential has a rise time on the order of 1 ms and better resembles a real action potential, as shown in Figure 6.10.1.  In summary, the attempt to reduce of the rise time of the action potential spike by adding a low-pass filter at the output circuitry does make the action potential waveform more realistic. However, the low-pass filter also reduces the speed for the feedback currents and results in an unsuccessful voltage clamp. The figure shows the block diagram of the instrumentation system for testing the cell capacitance measurements. The Cell Capacitance Emulator (CCE) has two outputs. The "RC output" is sent to the Universal Clamp. The "C output" is sent to the capacitance meter (Model 3000, GLK Instruments, San Diego, CA) for determining the precise capacitance values. The capacitance meter also has an analog output, which is displayed on an oscilloscope.

Figure 39 Block diagram of the instrumentation system for testing the cell capacitance measurements
The operation on the Universal Clamp is as follows. By electing the Vesicle Mode and then clicking the Set Mode button to operate in the capacitance measurement mode. A sinusoidal current waveform was used to excite the neuron for the capacitance measurement. The baseline-to-peak amplitude of the sinusoidal wave is set by use of the Vesicle Amplitude field and the baseline by the Current field. See Section 6.7 Current Clamp for more details. The current amplitude should be set at a large value for increasing the signal-to-noise ratio. But too large an amplitude may have undesirable effects such as triggering the action potentials. It is also possible to use the current pulse mode such that the baseline of the sinusoidal wave can change during the pulse duration.

Test Results
The voltage window now shows the magnitude of the frequency response, and the current window shows the phase of the frequency response. The detection of the vesicle activities can be based on either the magnitude and the phase. In general, the phase should be more sensitive than the magnitude. The result in Figure 7.2.1 below shows a simulation of the vesicle detection by touching the passive neuron model with fingers intermittently. The Vesicle Amplitude is set to 10 nA. The baseline of the sinusoidal wave is 10 nA, changes to -10 nA for 8 s, and returns to 10 nA.   the market. Thus, its performance can be considered as the state-of-the-art standard.
The 500 fA pulses are presented to the capacitance meter. As shown in Figure   7.2.4, the capacitance meter is capable of detecting the 500 fF pulses with the 200 ms pulse width (left), but not with the 50 ms pulse width (right). The specifications of GLK Model 3000 indicate that the analog bandwidth depends on the selected range [1], as shown in Table. The range selected for the above measurement is 200 pF, which has a time resolution of 92 ms. This explains why the capacitance meter can measure capacitance pulses with the 200 ms pulse width, but not the 50 ms pulse width.
In summary, the representation of the cell capacitance change due to a vesicle activity on the order of 10 fF with a 2-ms pulse width was challenging. The parasitic capacitance from the wiring and the breadboard was often on the order of 10 pF. The existing algorithm based on short-time Fourier transform for detecting the vesicle capacitance was insufficient to detect such a small and fast capacitance change.
Nevertheless, the proposed instrumentation system based on the cell capacitance