Choice of electrode

Stimulating electrodes require two poles, one 'working' and one 'counter', between which charge is applied. The working electrode is placed in the stimulation site, and the counter electrode close by to complete the circuit.

Single electrodes can be composed of a probe containing both poles, or there can be two individual probes for each. Monitoring electrodes can be mono-polar, although they are best used with a 'reference' electrode. Potentials of the working electrode and counter electrode can then be given with respect to the reference electrode.
[Merrill et al,2005]

Probe electrodes are shaped to ease insertion and can be quite long to penetrate deep into the brain. They consist of one or more conductors in an insulating sheath.

One can create fine arrays composed of many probes that interact with a wide group of neurons. These arrays are especially useful for monitoring neural activity, due to their large field of acquisition.

A cuff electrode is one wrapped around a nerve, usually made from coils of wire, as used in vagus nerve stimulation. These are excellent for prolonged use, but are hard to adjust once inserted.

A sieve electrode is one placed through a nerve, severing the neurons. As the neurons re-grow they pass between pores in the electrode, making a very intimate connection. These are mainly used in experimentally in animals, but have great potential for both monitoring and stimulating specific groups of neurons in a nerve.
[Gimsa et al,2005]

Materials used in the implant

Each electrode consists of a conducting element sheathed in an insulator. The electrode must be totally stable, with no deterioration or loss of electrical conductance throughout its use.
Platinum is often the first choice, as it is extremely inert and a good conductor, though rather expensive. Other materials include gold, tungsten, tantalum and nickel. Electrodes containing iron are a definitely unsuitable for brain stimulation, as iron oxidises readily (rusting), making it neurotoxic.

The electrode material affects how charge is transferred into the medium. Conductors that create a layer of ionised particles on their surface during a pulse are preferable as this surface layer is recovered by reversing the potential. Thus the surface area of the electrode is important in determining how much charge can be stored in this way, where the electrode-layer acts as a capacitor. An electrode witout much capitance will release more energy in irreversible 'Faradaic' reactions, compared with a high capitance electrode.

Insulators are necessary to isolate the electronics from the neural and bodily tissues. Inert compounds are used to cover the conductor and any electronics implanted with it. Teflon is used for flexible elctrodes and couplings, while silicon and plastics are to coat inflexible electrodes.

Covering the surface of the electrode with bio-materials can improve acceptance of the foreign object into the body, as inflammation and scarring adversely affect performance of electrode arrays. Bio-compatibility is improved by coatiing electrodes with nanoscale layers of proteins such as chitosan, gelatin and laminin. In the future these surface layers can be designed to sooth the insult to surrounding tissue as well.
[He & Bellamkonda,2005]

Insertion of electrodes

Putting arrays of electrodes into the brain requires accuracy and delicacy.
Single electrodes, like those used in deep brain stimulation, may have to pass through several layers of neural tissue before reaching the target. Choosing a route which passes through functionally robust areas is important to minimise damage the brain. Rupturing blood vessels is a major risk and the primary cause of failure when implanting electrodes.

Electrode arrays containing tens to thousands of individual probes cause significant compression of neural tissue during insertion, in a similar way to a bed of nails. It appears that inserting arrays mechanically is preferable, being faster less compression occurs and neurons are less traumatised, leading to longer lasting contacts.
[Rennakera et al,2005]

Stimulation profile

Designing a stimulation profile involves a compromise between what is effective and what is damaging to the tissue and the electrode. What an individual can tolerate depends on their age and species, as well as the site of stimulation.
The effects of electric pulses depend on current flow - the more current the greater the affect. Yet high currents will cause electrolysis (transfer of electrons to and from electrodes) creating reaction potentially toxic reaction products and pockets of gas. Current flow is proportional to voltage, so relatively low voltages are necessary for the longevity of the subject and the implants. Another way to make stimulation more tolerable to cells is to use an appropriate waveform.

Prolonged electrical pulses give rise to an increasing impedance between electrodes as they polarise; in other words less and less current can flow, as discussed in the mechanisms page. Therefore the longest pulse of practical use is about 0.25 ms, with many profiles using around 0.1 ms. Below 50µs the effect of the pulse becomes quite ineffective. To avoid electrode polarisation, raising impedance in the circuit, alternating current can be used, shifting the polarity between electrodes. Thus there is a stimulating pulse, and a reversal pulse, which can be individually coordinated into a purposeful waveform.

The activitation time constants of several key ion channels are hundreds of microseconds to milliseconds. Stimulating pulses of similar durations allow one to selectively manipulate the opening and closing of these ion channels. Also the myelination of fibres and their width impart specific characteristics that can be exploited. Waveforms selective for certain types of fibre require stimulation or reversal phases with long pulse widths, thus they may sacrifice safety for selectivity.

Typically current control is used to generate electrical pulses, where the ammount of current flow is regulated. A simpler option is voltage control, where the current depends on the voltage and resitance between electrodes - Current = Voltage/resitance. The disadvantage of voltage control is that only the net potential between the working and counter electrodes is controlled.

The phase of pulses is important. One can apply a current in one direction, usually using the working electrode as the cathode (negative pole), then 'opening' the circuit - essentially disconnecting the electrodes. Additionally the poles can be reversed after each pulse, which acts to reverse electrochemical changes occuring at the initial pulse, minimising unrecoverable charge in the form of reduced/oxidised products. However this may reverse some of the desired effect of the stimulation phase, i.e. it may suppress an action potential.

With these two concerns in mind, one is limited to short pulses of low voltage. What is left to adjust is the waveform of each pulse, the frequency at which these pulses are applied (number per second), and how often one applies trains of pulses (constantly on, or only on when needed).

In some cases safety must be compromised, for instance in the treatment of epilepsy where electrodes are implanted into the epicentre of the seizure. This works by forcefully hyperpolarising the neurons, preventing activity, and requires long (>1s) monophasic pulses to control epileptic activity.
[Merrill et al,2005]

Site of stimulation

The site of stimulation is vital; results entirely depend on which neurons are affected by the stimulation, and what function those neurons serve. Choosing a site relies on current knowledge of functional anatomy, while finding the site relies on imaging and monitoring during surgery. Once one has a site in mind, then consider how to best affect that site.

Myelinated neurons are hard to stimulate along their axon, compared with bare axons. The most common route of stimulation is through dendrites. Creating an electric field at a right angle to a neurite's length is the ideal way to stimulate it, achieved by placing electrodes on either side.

Stimulating a group of neurons is sometimes best done by targeting their dendrites, or even axons that synapse with the neuron. The threshold to activate an axon is usually lower than that to activate the cell body directly. Thus a small nucleus of neurons might be hard to stimulate selectively with pulses direct to it, instead one might look for an afferent pathway that synapses into the region of the nucleus you want to stimulate.