A Chain is Only as Strong as its Weakest Link: the Microelectrode

The targets for Deep Brain Stimulation (DBS) for movement disorders are not the subthalamic nucleus (STN), globus pallidus interna (GPi) or the ventral intermediate nucleus of the thalamus (Vim).  It is the sensori-motor region of the STN.  It is the motor homunculus appropriate to the patient’s symptomology in the GPi and Vim.  To date, the only way to confidentially know whether the target has been obtained is by microelectrode recordings of extracellular action potentials (spikes) generated in individual neurons with simultaneous sensori-motor activations of the periphery.

Critical to successful microelectrode recordings is the prevention and recognition and elimination of noise and artifact.  Noise and artifact are ubiquitous but nonetheless, can be dealt with successfully.  Modern electronic electrophysiological recording systems are remarkably robust at preventing or reducing noise and artifact from the external electrical environment.  However, a sound knowledge of electricity, electronics and electroneurophysiology is still necessary for successful intraoperative neurophysiological monitoring because an “ideal” recording environment cannot be guaranteed for each and every case.

Environment sources of noise and artifact generally are appreciated.  What may not be appreciated is that actual neuronal spikes are a predominant source of “noise” as systems become increasingly robust at preventing environmental noise.  This seems counterintuitive and hence a source of confusion.

The voice in the crowd

The challenge for the intraoperative neurophysiologist is to hear the seemingly one-sided conversations of individual neurons in a crowd of many neurons.  The situation can be likened to attempting to listen to the conversation of a single fan in a stadium of fans where everyone is talking.  The neurophysiologist is in the arena with a microphone trying to listen to the single fan.  If the neurophysiologist is using the wrong microphone, it is unlikely that she will be able to disambiguate the speech of the single fan of interest from the speech of the crowd.

More importantly, it is critical that the neurophysiologist in the arena  knows whether or not the voice she is listening to is that of the single fan of interest.  To do so, the speech volume must “stand out” from the volume of the other voices in the crowd.  If she would use a volume meter, the voices in the crowd would create amplitudes that vary by some small magnitude, while intermittently, the voice of the single fan of interest would be substantially larger than all the other voices.  Also, the frequency at which the single fan speaks must be much less than the voices of the crowd.  If the single fan of interest was talking incessantly, there would be no way to distinguish the voice of the single fan.

Another way the neurophysiologist in the arena can have confidence that she is listening to the voice of the single fan of interest is if what the neurophysiologist is hearing “makes sense.”    If what is heard appears to be strands of different conversations being superimposed, then the neurophysiologist cannot be sure that she is listening to the single fan.

The voice of neurons

The issues described above for recording the speech of a single fan in a stadium crowd hold true for analyzing and interpreting the train of neuronal spikes.  The amplitude of the spike of the neuron of interest must “stand out” from the amplitude of the spikes from all the other neurons (perhaps hundreds) that are in the vicinity of the microelectrode tip.  This is critical in determining whether the neuron of interest is related to sensori-motor activations of the periphery.  If the spikes of other neurons not related to sensori-motor activations are approximately the same size, then the pattern of spikes indicating relationship to the peripheral activations will not be seen over the unrelated spikes of other neurons.  The key then is to ensure that only a few neurons have relatively large spikes and all the other neurons have small spikes.

No electronic amplifiers, filters or computer systems can ensure that only a few neurons have relatively large spikes and all the others have small spikes.  This critical condition can only be assured by the proper microelectrode.  Thus, the appropriate use of the proper microelectrode is critical to identifying the actual target for DBS.  It is important that the intraoperative neurologist have a sophisticated understanding of the physics of microelectrodes.

Physics of microelectrode recording

As might be appreciated, one way to “zero in” on the voice of the single fan in the crowd is to use a microphone that is so small that it can only pick up one voice at a time and then move the microphone close to the person of interest.  In actuality, as one cannot know which voice is the one of interest, it is necessary to move the microphone from one person to another.

Similarly, one would like a microelectrode with a recording tip so small as to pick up the spikes from one or a few neurons in the immediate vicinity of the microelectrode tip and far enough away from the dendrites and cell body (soma) of other neurons so spikes from those neurons are much smaller in amplitude.  However, the small tip presents challenges.

The recording tip typically is made of a conductive metal such as an alloy of platinum-iridium.  As metal atoms loosely hold onto electrons, the electrical field generated by a neuronal spike “pushes” the loosely held electrons towards the electronic amplifier, filter and computer systems.  However, the metal conductor has an impedance to movement of the electrons, thereby reducing the signal that can be recorded.  The smaller electrode tip creates greater impedance and smaller the signal.  The higher impedance also increases the risk of artifact and noise.  Thus, one is confronted with wanting a recording tip that is very small, but at the same time, one that does not create too much impedance.

It is not the volume (size) of the microelectrode tip that is the major determinant of impedance, but rather the surface area.  Thus, an optimal microelectrode has the highest surface area to smallest volume possible.  The human brain is an example.  While the volume of the human brain is relatively small compared to some large animals, its wrinkled (walnut-like) surface greatly increases the surface area relative to the volume.  Likewise, the tip of the microelectrode can be wrinkled in manufacturing to increase the surface area relative to the volume.

Perhaps the best way to increase the surface area relative to the volume is to use very porous metals, in other words, metals that have the microscopic appearance of a sponge rather than a solid mass.  The surface area within the crevices of the sponge is far larger than the actual volume of the sponge.  Platinum-iridium alloys are very porous and thus, are most optimal for use in microelectrode records compared to the popular alternative, tungsten.

It’s the size, not the impedance

There is a common misconception that the impedance defines the optimal quality of the microelectrode.  This is incorrect.  Rather, it is the size of the microelectrode relative to the neuron that is critical.  The impedance is only secondarily related to the size of the microelectrode tip.  For example, a very small tip with a highly porous metal, such as platinum-iridium, may be significantly smaller than a microelectrode made with tungsten.  Thus, the platinum-iridium microelectrode, generally, will be more successful at isolating the spikes of the neurons of interest.

This author’s experience

This author has conducted microelectrode recordings of neuronal extracellular action potentials for over 40 years using a variety of microelectrode types.  Invariably, platinum-iridium microelectrodes with tip exposures on the order of 20 microns and impedances between 0.5-0.6 megaohms have provided the best isolation of spikes from individual neurons and the most stable recordings.  The efficacy of neuronal recording is far better than with tungsten microelectrodes and certainly better than electrodes with impedances less than 0.4 megaohms.


A careful analysis of the issues noted above invariably leads to the conclusion that platinum-iridium microelectrodes are superior.  Other applications may require somewhat different specifications.  To be sure, platinum-iridium microelectrodes may cost a bit more, but the additional costs likely pales in comparison to the added monitoring time and frustration from using suboptimal microelectrodes.