Not so long ago, I was traveling with a friend of mine who said, "some people are just lucky bastards." What he was alluding to was one of his graduate students (a mutual friend) who, it seemed, got a publication for free. The whole story revolves around an earlier publication in PLoS One that was trying to answer the question "is decapitation a humane way to kill research animals?" Their paper, based on EEG data, completely failed to answer that question, but found some very strange EEG signals long after neural activity had apparently ceased. And when faced with surprising data, they did what any good scientist would do and speculated wildly. Wild speculation being kind of exciting, this made the national newspapers (Dutch language).
Needless to say, this is red rag and bull territory. It just so happened our friend, Bas Jan, was building a better neuron model to help solve neural imaging problems. The model accurately described the observed EEG data—a remarkable achievement all by itself—which then allowed the researchers to riff on the wildly speculative interpretation of the data.
How to kill a rat?
One of the guiding principles of animal research is to minimize suffering. Care for your animals well, don't carry out needless procedures, and, when it is time, kill them as painlessly as possible. With the strict ethical oversight within the university system, these ideals are rigorously enforced. But some questions are difficult to answer. For instance, what is the best way to kill an animal? Is a death that is presumed to be quick and relatively painless really so?
A group of researchers from Radboud University of Nijmegen set out to answer this question for decapitation. Decapitation is sometimes preferred for euthanizing lab animals because the brain is not perfused with anesthetics during death, making it easier to analyze brain tissue post mortem.
But, if the death isn't quick or nearly painless, then it shouldn't be used. To get a bead on this question, the researchers observed the electro encephalograms (EEGs) from rats as they were decapitated. In line with previous research, they found that the amount of activity in the brain appeared to decrease rapidly. They reached this conclusion by analyzing the amount of power in the EEG signal, which decreased by half every 6 seconds after decapitation. After 30 seconds the EEG was recording nothing but noise, and you would think that nothing more interesting would happen.
But you would be wrong. At about the one minute mark, a single, low frequency pulse appeared. The Nijmegen group speculated that this is the time at which the neurons' membrane potential fails, blocking further transmission of sodium ions. The nature of the EEG pulse indicated that a large portion of all the neurons in the brain were failing at the same time, leading to what the researchers referred to as a "wave of death." They then went on to claim that this could be the basis for determining brain death, because, at that point, presumably, there is no return for the neurons.
I smell a rat
My friends at the University of Twente read this paper with a kind of horrified fascination. The results were unexpected and exciting, but the conclusions... Well, rampant speculation should come with health warnings.
Luckily, it also turned out that my friends were in a good position to contribute. They modified their existing model neuron to incorporate modified ion channel behavior. In particular, they wanted to understand the potential difference across the membrane when blood flow ceases and the cells start to run out of energy. Their model is purely based on the physical transport mechanisms of ions in the presence of voltages that are created by the ions themselves.
Decapitation is modeled by turning off transport mechanisms outside the cell and removing oxygen. The lack of oxygen causes the cell's activity to die off, while the lack of transport allows ions to build up inside and outside the cell, changing the charge balance.
To their absolute surprise, their model also showed a spike in activity, albeit at about half a minute after blood flow ceases—but, notably, after all apparent activity has ceased. Examining the ion concentrations around the activity spike, they found that, as the membrane potential slowly increased to zero, potassium ions were trapped in the cell. But at about 30 seconds, the membrane potential rose above a critical value, allowing potassium to be released. On release, the membrane potential dropped again, blocking potassium transport. The result was a series of short, strong spikes in membrane potential as potassium was released by the cell.
But that is a single cell, not an EEG. Normally, scaling the results from a single neuron to a complete EEG signal is wild speculation. The problem is that the brain is a series of interconnected neurons, where the behavior of one neuron changes the behavior of may others, making things incredibly complicated. Estimating the average of a population's activity (this is what an EEG measures) from the response of a single neuron is laughable. In this case, though, the researchers are on firmer ground. At the time of these spikes, the neurons are no longer talking to each other, implying that the EEG can then be recreated by summing up the response from individual neurons.
It seems that the Nijmegen group had intuited the correct physical cause of their EEG signal. But what of their interpretation? First, it's incorrect to assume that neurons cannot be revived after the membrane potential fails. The University of Twente researchers point to a body of research that suggests otherwise, including papers where brain tissues were removed from patients several hours after death and the cells revived. So membrane potential doesn't represent either cell death or brain tissue death.
It seems that previous research had also hinted that there might be an EEG signature to cell membrane potential failure. Experiments on brain slices had shown that the flux of potassium that occurs when one cell membrane fails can trigger the membrane potentials in neighboring neurons to fail. The result is a slow moving wave of membrane failure. This result, though, doesn't fully explain the Nijmegen group's results—their wave is far too strong and short. A single point of initiation would take more time to cover the entire brain than the signals they measured suggested.
At this point, the University of Twente researchers engage in their own bit of speculation. They believe that the results can be explained by assuming that the physiological environment is such that many neurons in widely spaced locations can fail at approximately the same time independently, leading to local waves of membrane potential failure that encompass the brain in a much shorter time.
A fun pair of studies, and certainly not the last word on the subject.
PLoS One, 2011, DOI: 10.1371/journal.pone.0016514
PLoS One, 2011, DOI: 10.1371/journal.pone.0022127
Listing image by Photograph by news.gmu.edu
reader comments
38