Lightnews — Scholar-powered news

Ulrike Horn @ulrikehorn.bsky.social · Aug 27

Thank you! 😊 Without all the help from internal and external cooperations it wouldn't have been possible! Also mentioning Johanna Vannesjo here who was there from the start to establish these measurements!

Ulrike Horn @ulrikehorn.bsky.social · Aug 27

Thank you! 😊

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

Thanks to all the people involved – this has been quite a journey!

2

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

Thank you for reading this far! We have more details in the preprint – what we did in terms of processing of fMRI data, whether individual differences in activation patterns reflect stable idiosyncratic responses and how participants rate and experience the heat pain we give them.

1 2

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

Groups who used 3T spinal cord fMRI already suggested that responses to long heat pain may be located in deeper layers of the dorsal horn (e.g. @christianbuchel.bsky.social lab), but of course with better resolution at 7T we can now explore this in detail!

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

Is this known? There is lots of evidence that patterns like these exist – starting in the periphery (A delta fibers firing differently) up to the brain (onset responses to heat that decay shortly after) but not much evidence from spinal cord data!

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

Is this an artifact? Probably not – if it would merely be driven by BOLD signal getting drained somewhere due to vasculature, it would move over time from inner to outer regions of the cord as the veins in the spinal cord have a radial structure. But we see the reverse pattern!

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

In summary we saw activation in the superficial dorsal horn for the very first response to a heat stimulus and more widespread activation especially in the deep dorsal horn for the sustained response to the stimulus!

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

And importantly: for these phasic responses it is the superficial dorsal horn (orange) that seems activated!

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

We found promising results in our first dataset, then pre-registered this analysis for the second dataset. And if we look into these phasic response results, we see a more focused activation on the ipsilateral dorsal horn:

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

We wondered whether at least for a certain amount of time the superficial dorsal horn would be involved and analyzed only the first 3 s of each heat block. There is evidence from recordings of action potentials of A delta fibers that some of them may fire in the beginning of a stimulus only!

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

But then again most of this knowledge comes from animal research in which methods clearly differ: often stimuli are different - like short electrical stimulation, and some methods do not even allow for insights into deeper laminae function.

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

Result: it is especially the deep dorsal horn that shows the strongest activation (dark red)! The superficial dorsal horn is not much involved in our activation. This was a surprise as the superficial dorsal horn is supposed to be the major termination zone of nociceptive primary afferents!

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

We wanted to come closer to these laminae by separating the dorsal horn into superficial (laminae I/II), middle (laminae III/IV) and deep layers (laminae V/VI) and created a high-resolution layer-specific grey matter atlas in our template space based on an annotated spinal cord cross section.

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

Based on cell structure the spinal cord has traditionally been divided into smaller functional units – the laminae. The dorsal horn consists of 6 laminae I-VI with the superficial (laminae I/II) and deep (lamina V) being the main entry points for afferent fibers conducting heat pain information.

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

But let’s not stop at this proof of principle! Now that we have actually higher resolution than previously at 3T and especially 0.8 × 0.8 mm2 in plane resolution, we can go even further and look into functional units smaller than just dorsal vs ventral.

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

This is not new: compare for example the glucose utilization that Abram & Kostreva report in 1986(!) using noxious heat pain in cats. It also fits with all the 3T spinal cord fMRI literature about heat pain that report similar activation patterns with particular focus on the dorsal horn.

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

And when we specifically test for activation in the ipsilateral dorsal horn, we can report significant results. 🥳🥳

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

If we look into this without strict thresholding, indeed we see activations on the ipsilateral side – stronger than on the contralateral side! Overall lots of activation in dorsal and ventral horns!

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

To investigate this, we use a simple design with blocks of 30s each using heat pain on the forearm – we have to establish the basics, right? What would we expect? Hopefully activation on the ipsilateral side and in the dorsal horn, as this is where the textbooks say heat pain is processed.

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

Both approaches boost our signal immensely and now finally allow for some insights about the functional units that process heat pain in the spinal cord!

a woman in a pink sweater is smiling and saying it 's getting hot in here

ALT: a woman in a pink sweater is smiling and saying it 's getting hot in here

media.tenor.com

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

In a second approach, @mrfinsterbusch.bsky.social developed an EPI protocol for slice-specific z-shimming, which allowed for the acquisition of an independent dataset without the need to pre-select participants (N = 25). This technique is based on earlier work at 3T and can be completely automated.

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

In a first approach, we acquired a large screening dataset (N = 62), from which we selected participants with adequate data quality (i.e. low levels of distortions and signal loss in EPI data) for participation in the task-based fMRI study (N = 16).

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

So, what options do we have? Scanning only the participants that have better data? Or find some solution to improve the data acquisition? Actually – we did both 😅

a man in a suit and tie is looking to his left

Alt: Mr Bean asking 'why not both?' and smiling

media.tenor.com

1 1

Ulrike Horn @ulrikehorn.bsky.social · Aug 26

Unfortunately, these inhomogeneity artifacts increase with higher field strength, causing significant signal loss and distortions in typical EPI acquisitions! Here, you see EPI scans of participants with varying data quality. The green outline indicates the anatomical location of the spinal cord.

1 1