The Blackout
By Siegfried Othmer
Shortly after the blackout on the East Coast where they have so much practice, an op ed appeared in the New York Times by Albert-László Barabási, the author of “Linked: The New Science of Networks.” The point of the editorial was to point out that interconnectivity can be both a boon and a bane. By moving to a distributed energy system with multiple sources of power, we have to provide an increase in interconnectivity to allow the shifting of power to meet fluctuating demands.
Such interconnectivity increases the availability of resources to meet contingencies, but this particular benefit is not cost-free. Interconnectivity also promotes interdependency. The network also propagates problems. Recently, we have seen a number of discussions about the possible cause of the breakdown. We are not thinking here about the ultimate political cause---whether it is in deregulation or elsewhere---but rather the mechanical cause. It was pointed out rather quickly that it is not enough to identify a single outage as the critical element. The system is supposed to be robust against the dropout of any particular element, or even of several such elements of the system, whether they be power sources or transmission lines. So wherever the finger gets pointed, we will already know ahead of time that that will not be the answer, but at most a part of the answer---one element in a series of cascading failures.
We have also heard that there was trouble on the lines in the form of huge power surges of mysterious causation hours before the system finally collapsed. I am speculating here---which I can safely do since this will have no policy consequence whatsoever---but I wonder if the system was not in fact working quite well, that is to say, “within specifications.” Perhaps there was no obvious failure at all. We may have seen here an instance of what all feedback systems are heir to, namely a runaway oscillatory response. There may have been a chain of events that led to a buildup of power fluctuations rather than a dampening of them, even as every constituent part played its assigned and intended role. Or perhaps somebody was indeed asleep at the switch somewhere, but under ordinary circumstances that would have been accommodated by the system.
In other words, we may have a system here that under certain rare circumstances can be provoked to go into a kind of seizure, even though all the parts are working pretty much as they were expected to. The Pennsylvania and New Jersey power authority can now pat itself on the back for staying out of the blackout by disconnecting itself in time. But this was likewise not cost-free. It is quite possible that that very act of disconnection doomed the rest of the system to failure because it removed some of the buffer that was needed. It is as if a climber drops the safety rope when he sees himself at risk. This way he saves himself, but dooms his compatriots. It’s not an easy call.
It is this intrinsic vulnerability of networks, their capricious tendency to migrate toward collective activity, that Barabási was writing about. We may have to go to another level of modeling to assure stability under a greater variety of inputs. We may have to run the system with computers the way we fly military aircraft with computers. Human reaction time may simply no longer be adequate to deal with a rapid escalation of events.
Why is this all relevant to issues of health? The fact is that the brain has the same problem of collective events, and that is why the problem of brain instabilities is so important. The laws that apply to both kinds of systems are essentially identical. Our brain needs to be rapidly responding to keep us alive, but yet it needs to be stable. Brain stability may be considered the “prime directive” as far as brain function is concerned.
Of course it may very well turn out that the blackout is attributable to some human blunder, or sequence of blunders. That would be almost a shame because we would miss out on the lesson that needs to be learned. It would be much more useful if this event just gave all of us instruction with respect to our vulnerability to oscillatory runaway in massively connected systems even when all the pieces are behaving more or less as they were designed.
The analogy to what is going on in the brain is possibly even closer than I suggested. We are, after all, dealing with a “signal” at 60 Hz, and for systems to cooperate, they all have to synchronize themselves not only in frequency but in phase. This means that our whole nation is in fact operating its power grid synchronously, which is quite a trick. Of course the speed of light comes into the picture also, but let’s not worry the details. One way to get huge shifts in the direction of power flows is to shift the phase of one element or another. Here we have an analogy to what is going on in the brain, where we believe synchrony at certain brain frequencies governs the communication between different cortical regions. Management of the power grid can be seen in terms of management of both phase and amplitude of the signal, and the same goes for neurofeedback.
The other analogy is to the “non-localizability” of the problem. The likelihood is that when the story of the blackout is finally told, there will be multiple locations that contributed to the ultimate breakdown of the system. Or, conversely, even if we do identify one salient event, most knowledgeable observers would be reluctant to say, “let’s fix this one problem and we will be fine.” Similarly, we have to deal with non-localizability in the brain. For example, a lot of things masquerade as frontal lobe problems simply because the frontal lobe projects to nearly everything, but in fact the frontal lobe may be innocent until proven guilty. In Parkinson’s, we may be dealing with a relatively localized problem, but our neurofeedback remedy for it is likely non-localized! We ask the relevant networks to make whatever adjustment is necessary to control a tremor, for example.
Another analogy is to seizures, where in some instances once can predict a seizure some hours ahead of time, either by the EEG or through behavior. In the case of the blackout, it was apparent hours before darkness fell that something unruly was afoot within the network, as power demands fluctuated wildly without any discernible cause. No single observer had the power to damp down the oscillations and restore the system to calm regulation. Resolution therefore lies in dampening down the oscillations before they become unmanageable, and the same goes for EEG neurofeedback. We have to build stability into the system so that it remains within bounds whenever it is challenged.
Another aspect of this analogy is that our model has to accommodate the rare and extreme event. So much of what we deal with clinically consists of the rare event: Panic attacks, seizures, migraines, and incidents of domestic violence or criminality. We have the problem of training the brain when the evidence for such instabilities may be slight or intermittent. Likewise, a major power outage is a rare event, and one may be forgiven for believing that stability has been achieved when there are no breakdowns over a period of more than twenty years.
A final lesson to be drawn from this experience is the outsized ratio between the cost of prevention and the cost of the cure. Our dependency on electricity is something that bears essentially no relationship to its cost. More specifically, our dependency on the reliability of electric service---its continuity---is much greater than the cost of ensuring it. Similarly, the value of brain stability is far greater than the cost of achieving it through neurofeedback.
We are seeing in our society the atomizing forces of competition contending with the civilizing forces of collaboration. These antithetical imperatives come to a head in something like our electric grid, which is an intrinsically collaborative construct, though managed to date largely by the invisible hand. So the electric grid is on the path toward becoming another “tragedy of the commons,” used and abused by all. Likewise, nothing costs our society as much as brain instabilities untended, yet we are unwilling as a society to pay the tab for dealing with them.
P.S. The reader of “Linked” may also enjoy Duncan J. Watts’
“Six Degrees: The Science of a Connected Age.”
For further reading, go to www.eeginfo.com