Thursday, 10 May 2012

Scanning The Acidic Brain

According to University of Iowa researchers Vincent A. Magnotta and colleagues, any neuroscientist with an MRI scanner could soon be able to measure the acidity (pH) of the human brain in great detail: Detecting activity-evoked pH changes in human brain.

If it works out, it would open up a whole new dimension of neuroimaging - and might be able to answer some of the biggest questions in the field.

The method relies on measuring T1 relaxation in the rotating frame (T1ρ). Essentially, it's about the rate at which protons are swapped between water molecules and proteins. That rate is known to depend on pH.

Anyway. It certainly looks impressive. Using a standard 3 Tesla MRI scanner, they were able to image the whole brain once every 6.6 seconds - only slightly slower than conventional fMRI measurements of brain activity, where 2 or 3 seconds is more usual. The spatial resolution was comparable to fMRI.

Here's how it did on some bottles of jelly -

Then they moved onto mouse brains (the differences are smaller here)...

And finally they scanned some people. They were able to detect the (very small) pH changes caused by hyperventilation, which raises pH, and breathing air enriched in carbon dioxide, which lowers it.

Lovely pictures I'm sure you agree, and it's a very clever methodology from a technical point of view. But what will it mean for neuroscience?

Well, for one thing, it might be able to help resolve some of the debates over what conventional fMRI is actually measuring. For example, some neuroscientists believe that many (seemingly) interesting fMRI results may actually be (at least partially) reflections of subtle changes in breathing rate. Measuring acidity, an indirect proxy for breathing, could start to answer such questions.

The main question though is, what are we going to call the new method? "T1ρ MRI"... not a terribly catchy name.

Maybe MRalkalI?

ResearchBlogging.orgMagnotta, V., Heo, H., Dlouhy, B., Dahdaleh, N., Follmer, R., Thedens, D., Welsh, M., and Wemmie, J. (2012). Detecting activity-evoked pH changes in human brain Proceedings of the National Academy of Sciences DOI: 10.1073/pnas.1205902109


practiCal fMRI said...

Good find. I don't see any major holes in the method; the use of spin echo EPI with short TE for the human brain scans should reduce a lot of (SE) BOLD, although there will still be some. Also some T2* weighting due to the EPI readout, which also generates some BOLD contamination. But I still wouldn't discount the finding. I'll be interested to hear the opinions of the ISMRM types when they return to their email (and regular sleep patterns) today!

SAR (heating) will likely be a problem for routine, whole brain use. That's because of the spin echoes as well as the spin lock pulses. So I don't anticipate this method displacing BOLD fMRI at 3 T. It could prove to be a useful tool to furthering understanding of neurovascular coupling though.

DS said...

Hmm. No mention of the receive coil/array used in this study. No fMRI article should ever be published without a specific description of the receive coil/array. If a surface coil were used then I would be very suspicious of these results!

petrossa said...

One fails to see the difference in relevance/quality of the data between this and fMRI. It's just replacing one marker with another whilst the basic flawed system of interpretation is basically the same.

Neuroskeptic said...

petrossa: Well, that's the point I think, on its own pH imaging may be no easier to interpret but together with fMRI it might be able to eliminate some possibilities.

I wonder if you could interleave fMRI and T1p scans, so you could (say) scan BOLD, pH, BOLD, pH and so on in turn to measure activation & pH over the course of the same task?

petrossa said...

Yes that could refine results if indeed done properly. But it still gives ONLY the oxygenation of the brain in relative terms.

RELATIVE terms being the operative word here. As long as no definitive resting state stable oxygenation pattern exists that can be applied as baseline across the human species the results will always be mere indicators open to interpretation.

Since it's impossible to acquire such a resting 'pattern' (no brain rests in the same way whilst alive) you have to mediate out the differences mathematically.

And that gives you a baseline equivalent to let's say the pre-industrial CO2 level.

It's just a number without meaning.

It's a different case if you have a model with very few and fixed parameters. I don't know, how a building stresses during an earthquake for example.

With such models you can reach a very high degree of accuracy. But the more parameters which are interdependent in their suppressing/activating effect the less accurate the model becomes.

Forgive me for harping on about it, but a climate model is a very good example of how you get absolutely nowhere with representing reality using even the largest most sophisticated computers on earth.

Your fMRI etc has to with a fraction of that computing power plus simple statistical algorithms.


Sure you get about the answer you'd expect if you put the results obtained of standards acquired via other means. The problem is those 'other' means are even more simplistic and crude.

Proving your model recreates what a another even less refined method shows is hardly proof your system can by extrapolation show higher degrees of complexity with any degree of accuracy.

Back to climate models, they hardly can predict the past if you put in the historical observations as parameters.

If you get my drift.

swaran said...
This comment has been removed by the author.
fluoresentric said...

Thanks to provide information about Scanning The Acidic Brain. Good work keep it up !!!

PCR Primers