Surgery

The most crucial constraint of ESB is the need for invasive surgery to accurately target neurons, especially ones deep in the brain. There are many blood vessels in the brain, and pushing a probe through it's tissue must be done with utmost care to avoid rupturing these vessels. Most fatalities after neurosurgery are due to cerebral haemorrhage. Anchoring electrodes into the skull also has problems - knocks to the head could displace the probes, requiring furthur surgery to correct.
Hopefully as experience with ESB accumilates the surgery will become easier and safer. The hardware itself will also inevitably evolve. However at the present time invasive ESB is a tricky business.

Idiosyncratic brains

One major problem with electrical brain stimulation is that individual variations in brain structure disrupt the potential for accurate, systematic treatment. In animals with complex brains, who live for tens of years, there is massive scope for idiosyncratic brain organisation. The variation responsiveness to EBS therapies between subjects can be accounted for by individual phenotypes.

One intriguing insight into the possibilities of EBS come from stimulation of a patients exposed cortical surface during neuro-surgery. Stimulating the somatosensory cortex induced sensation in the respective body area for the patient. Other less reproducible findings include propagation of certain thoughts during stimulation of the parietal cortex. The cortex is perhaps the most individual part of the brain, showing least functional homology between individuals. Conversely the phylogenically older regions show more homology between individuals, and indeed greater preservation between species.

Taking this line of thought further, some individuals' brains will, potentially, have functional pathways that are not present in others, due to the synergy of genetics and lifelong interaction with the environment. Thus when systematically applying ESB to humans the possible sites for stimulation are limited to those gross ones which are homologous between individuals. There are therefore many other brain areas which can be stimulated, producing unique effects in each person. Whether these can be tapped depends largely on how imaging technology develops.

The problem of general trends vs specific variations in brain function is a form of the ancient problem with all scientific enquiry: science looks for order and form, by observing common traits, and then drawing relationships. This demands that anomalies, chaotic pieces of the puzzle, are neglected to sharpen focus on the more easily understood facets.

One advance which may accelerate design of individually tailored systems is that of more flexible stimulation devices. Trans-cranial magnetic stimulation, nano-robotics, and viral vectors could all allow targeting accuracy that can only be dreamed of now, as discussed in the future directions page.

ESB vs Pharmaceuticals

One might wonder what the point of going through massive surgery to stimulate the brain is exactly? Couldn't one just take a drug that would produce a similar effect?
Currently pharmaceuticals are a long way from being site specific, as they only select for molecular targets, which are usually present in many locations throughout an organism. In comparison ESB is able to target discrete physical areas, while missing others, where a drug might interact with them all. Indeed ESB can, theoretically, be applied to a single neuron.

Pharmaceuticals usually interact with cells to exert their effects. In between are long transduction mechanisms, and these can vary between species and individuals. Distribution of a drug is also variable, and a whole discipline - pharmacodynamics - attempts to understand and harness the often erratic nature of drug presence in vivo. ESB has the advantage of being instantaneous to apply and withdraw, with minute control over the voltage and current applied.

Placebos and shams

The ideal way to test any therapy's effectiveness is to use a double blind placebo (or sham) controlled study design. However such controls are hard to use with many ESB therapies, due to the invasive and blatant nature of implanting electrodes into the brain. It would be rather hard to persuade someone that they had a device implanted without giving them surgical scars. It would also be rather unethical to put someone through a massive surgical operation if they were only getting a placebo.
One possibility is through controlling the stimulator device, turning it on or off without the patient's knowledge. In this way one can tell a patient that their device in functioning when it is in fact inactive.

Unwanted interference

Having long electrodes is often necessary to penetrate deep into the brain, however the longer the conductor the more likely it is to experience interference. Effectively the electrode can act as an aerial, transducing electromagnetic waves into electrical charge. This is only going to be a problem when significant amounts of radiation are about.

The more sophisticated neural implants often make use of radio control to program the device, making post-surgery adjustments to the stimulation profile easy. This comes with a drawback - electromagnetic interference could disrupt the device's functioning, at best disabling it, at worst making it go haywire.
Luckily most devices are made with strict guidance to be resistant to such interference by using robust data exchange protocols. If the system detects an error, it is made to shutdown immediately rather than continue working in an unusual way.
Yet as neural implants become more sophisticated, perhaps even acting as two-way transceivers fro communication, the possibility of malfunction rises too. While computers can be turned off and sent back to the shop, having a neural chip crash could be somewhat more eventful.