Temporal lobe epilepsy (TLE) is the most common form of seizure activity in humans. It tends to arise predictably in discrete regions on the extreme poles of either side of the brain. The ability to detect signs that a seizure is about to occur, and short-circuit any undesirable neuronal activity with targeted laser light is arguably the most advanced technology we can imagine for our day. It is here — at least for genetically enhanced Mus musculus, the ordinary house mouse.
Optogenetics is the science of using genetically modified viruses to insert light-responsive channels into the neurons that they infect. If the virus is introduced early on in development, all the progeny (offspring) of that cell can potentially be light responsive. For neurons, this means one of three things depending on the kind of channels that are used. Either the light opens channels and the neurons fire an electric signal; the light closes the channel and they temporarily can’t fire; or the light opens, in effect, negative channels which makes it even more difficult for the neurons to become responsive in the near term.
A new study in Nature Communications has demonstrated exquisite control over TLE in mice pre-engineered to express these channel subtypes in certain subpopulations of neurons in the hippocampus, the key control organ in the heart of the temporal lobe. By rendering these mice susceptible to spontaneously generated seizure activity using kainate, an acid derived from seaweed, the researchers could detect signs that seizures were beginning with implanted EEG electrodes — and then shut them off with light. They used a feedback loop running through Matlab software to control the process.
The full diagnostic and therapeutic power of the technique was realized when the hippocampi on both sides of the animals were outfitted with this hardware. If a seizure was detected on one side, for example, the researchers successfully exercised several options. They could activate excitatory channels on the same (ipsilateral) side to fire all the cells the fiber beacon could reach in unison, and reset their activity just like a defibrillator squelches the chaotic quiver of an ailing heart. Alternatively they could activate inhibitory pathways, either by exciting inhibitory cells which in turn synapse onto excitatory cells, or through activating inhibitory channels directly on excitatory neurons. In addition to all that, they could stimulate neurons on the opposite, contralateral side, that span over to the side where the seizure activity was seen, and assert brain superiority from afar.
How about controlling epilepsy in humans?
What blockades stand in the way of human acquisition of this capability?
The mice in this study had these engineered channels in their neurons via their birthright — it was easier to in this case to precisely introduce them to the select target neurons. Recent work demonstrated transfection of adult animals to express these channels in neurons that have become quiescent, or cease to replicate and share the virus. Biocompatible polymer electrodes (pictured right) have been fabricated that provision for local delivery of the transformed channel genes through a central bore, light through integrated fiber, and recording of electrical activity through on-board electrodes.
To achieve the full power of the optical stimulation, multiple cells need to be selectively targeted by electrode arrays (pictured below). The most advanced devices to date have been developed by Ed Boyden and others and contain hundreds of discrete delivery points within a compact form factor.
With these new tools the treatment of undesirable brain activity is set to explode. It is a tricky affair though — seizures can present on the EEG without patient awareness and conversely the patient can feel a seizure that presents in the absence of any corresponding EEG signature. If we are to benefit from this technology the implant industry needs to play a bit of catch-up. Remote presence robots like the RP-7 slated to instruct nurses in the programming of neural implants are probably not going to cut it. The primary agent presiding over the feedback loop needs to be the patient. That means having an open and flexible platform that can be reconfigured on second timescales, not the week timescale of the scheduled appointment. We won’t have these types of implants tomorrow, but we can probably take comfort that we may have them before we are even ready for them.