Mar 19 2007

Do Nerves Conduct by Sound?

Thomas Heimburg is a Danish biophysicist who believes our current model of our nerves propagate signals is wrong, and he has an alternative hypothesis. He believes they propagate through mechanical waves – sound waves he calls solitons. The idea is intriguing, but as a neuroscientist I don’t buy it.

There is nothing pseudoscientific or fantastical about Heimburg’s ideas. He is starting from the point of view of a physicist, asking – does the current model of axonal propagation make sense from the point of view of thermodynamics? No, he says. The now classic Hodgkin-Huxley model of nerve propagation involves an electrical current that moves down a nerve axon by the flux of ions across the nerve cell membrane. This flux of ions is controlled by the opening and closing of ion channels – proteins embedded in the cell membrane through which ions flow.

The problem with this model, says Heimburg, is that conducting electricity across a resistance should consume energy and therefore produce heat, but precise measurements fail to show heat production. Therefore, he concludes, axons must propagate their signal as a mechanical pulse that does not produce heat. It’s simple thermodynamics.

I wrote him an e-mail asking him about his recent publications and this is a portion of his response:

Dear Steve Novella

Our model makes the easy-to-understand statement that currents through resistors should produce heat. However, a number of very respected colleagues over the past 50 years have noted that there is no net heat released. The first person to point this out was Adrian V. Hill in 1958. He was advisor of Hodgkin in Cambridge and he won the Nobel prize for his heat measurements in 1922. He pointed out that during the action potential a phase of heat release is followed by heat reabsorption. Within the accuracy of the measurement this heat re-uptake is complete. This is typical for mechanical waves or pulses, but not for dissipative phenomena as proposed by Hodgkin and Huxley. Hodgkin himself took this point very seriously and dedicated a whole paragraph to this finding in his textbook on nerves.

Our proposal that nerve pulses are density solitons does not imply, however, that the pulse has no electrical component. Membranes are charged capacitors and we see the pulse as a piezo-electric pulse. Thus, like in Hodgkin-Huxley our pulse consists of a nerve segment of charged capacitor that travels along the axon. In contrast to Hodgkin- Huxley this is based on reversible physics that does not consume energy. The measured currents are capacitive currents instead. When it comes to EEGs I therefore do not expect any difference.

Further, like Hodgkin-Huxley our model is meant to explain the propagation of pulses in the nerve axon but not the processes within the synapse. Therefore, our model does not make any statements on what happens in the synapse. We do not dispute in our papers the pharmacological evidence for synaptic processes.

Our studies are based on very sound physics published in respected journals. They look at biological systems from a very different perspective – one that is rooted in the laws of thermodynamics. Interestingly, this approach may yield an explanation for the action of general anesthetics that is extremely simple. This principle is freezing-point depression. Important also: in contrast to existing models it can be proven wrong – this I find an important characteristic of a sober and sound theory. It should make testable predictions!

When it comes to the response from the science community: A large part of the physics, biophysics and physical chemistry community in my field strongly greets these developments. They find the whole idea attractive and convincing. In contrast to the more molecular-biology based models they find it intuitive and easy to understand. Typically I can easily convince my (interdisciplinary) audiences about the problems with the text-book models and our anesthesia story within 45 minutes. There is a huge interest and desire in the more physical disciplines to understand the underlying biology problems.

Neurobiologists and biologists are typically interested and open to such a proposal. While they typically don’t care what the physical origin of these pulses really is, they find it very attractive that one can explain the phenomenon of general anesthesia very easily with high predictive power.

The community that does not like our approach are ion-channel scientists. Our models do not make any explicit mention of ion channels and don’t need them. For obvious reasons this community finds that disturbing. I do not dispute any of the findings in this field. I come from the Max-Planck Institute in GÃttingen/Germany where patch-clamp was developed and was a group leader there. I question, however, a number of the interpretations of these findings. I have never ever met a serious scientist who disputed that fact that there is reversible heat in nerve and that this implies reversible physics. Even patch-clamp experts find this convincing (at least those with a basic understanding of thermodynamics).

Like all science our models could of course be wrong or incomplete. Scientific papers always represent the present level of discussion. Our studies are meant to trigger such a discussion. But to disprove the basic idea would require serious thinking and investigation because the underlying physical argument is quite strong.

I hope that helps.

Best regards,


I guess I find myself most sympathetic with the ion-channel neuroscientists. There are a great number of studies looking at the role of ion channels, which Heimburg admits do not fit into his model. I think this is creating a bigger problem than the one he is solving. There is, in fact, a category of neurological disorders called channelopathies – mutations in ion channels that produce abnormal conduction. You could argue that these have their effects only at the synapses, but patch-clamp studies of nerve conduction show that they dramatically alter the propagation of electrical signals through the axon.

Ion channel neuroscience is fairly mature and sophisticated. It seems unlikely to me that this large body of evidence will be slain by a single stroke. Historically, mature sciences (not pre-scientific notions) are rarely overturned by a single discovery.

Further, I searched for recent studies that looked at the issue of heat production with axonal propagation. This one found that heat reabsorption was 45-85% of heat production – which means that they did find excess heat production. So I am not sure if the premise of Heimburg’s hypothesis is adequately established.

Heimburg’s research is very well done and is very interesting. It also highlights the fact that as science becomes more and more complex it is also tending to fragment as researchers are forced to specialize in order to have sufficient master over their area of research. Yet I strongly believe in the consilience of science, which means as the practice of science fragments we should build bridges to keep the ties between the different fragments connected. What I like best about Heimburg’s research, therefore, is that it is interdisciplinary – bringing the focus of physics and thermodynamics to neuroscience.

It is likely that neuroscientists can learn a great deal from the attention, and perhaps there are some mysteries to be uncovered by asking questions about the thermodynamics of axonal propagation. But the communication needs to be two way – we can’t just throw out what the ion channel experts have learned over decades of research simply because it makes the physics more elegant.

While I don’t think Heimburg has overturned Hodgkin and Huxley, I will want to keep my eye on this one to see how it all turns out.

No responses yet