Dr. Patrick Stroman

Queen's University Queen's University

Stroman Lab Research

Research in the Stroman Lab has two main branches; 1) identify how pain processing is altered in fibromyalgia (FM), and 2) development of fMRI methods for pain research.  This research has been progressing for over 20 years.

Pain is complicated, and so is pain research.

Pain isn’t just a sensation, it’s our cognitive and emotional response too, and it serves an important purpose. When pain processes aren’t working properly, this can result in chronic pain, or can cause of to feel pain from stimuli that shouldn’t normally be painful (like touch). We need to understand how this pain processing system works in humans in order to understand how (and detect when and where) it is going wrong in chronic pain conditions such as Fibromyalgia.

Pain serves the vital purpose of motivating us to reduce or avoid injury, and to protect injuries while they heal. How we feel pain can change depending on if we are sick, injured, or upset. We can learn what causes pain and how to anticipate it, and we can form expectations on how it will feel. All of these factors change how we experience pain at any given circumstance.

Our current basic understanding of pain is that receptors in our skin that respond to noxious stimuli send neural signals to our spinal cord, where they are processed and relayed up to our brainstem and then our brain. This ultimately produces the pain that we feel. These signals, however, can change over time. They can change based on having a brief stimulus or a sustained one, or having several brief stimuli in a short time. They can be altered by injury or inflammation. These signals are also regulated (turned up or down) by other signals coming down from the brain and brainstem. This feed-back changes our pain depending on our situation (such as being afraid, angry, in danger, relaxed, safe, etc.). Problems with the pain neural system that produce excessive or unexpected pain can occur at any of the levels of this pain processing system (the receptors, the nerves carrying the signal, the spinal cord, the brainstem, the brain, or a combination of many)

Functional magnetic resonance imaging (fMRI) is also complicated.

FMRI has several important capabilities though. With it we can study neural processes in humans (albeit indirectly) and observe how pain processing works while people are awake, comfortable, and able to communicate with us throughout the study. However, these methods are still under development and always improving, so much of our focus is on using fMRI the most efficiently we can to study pain.

Functional MRI is a specialized way of using MRI to acquire images repeatedly in a sequence, to get a sort of low frame-rate ‘movie’ of what is happening in terms of tissue changes in our brain. We measure signal changes over time, and these changes are related to changes in blood-oxygenation levels, so we call them BOLD (blood oxygenation-level dependent). With it we are able to study the neural system involved with pain processing at all levels of the central nervous system: the brain, brainstem, and spinal cord.

FMRI is rarely done in the brainstem or spinal cord

The problem with pain research is that almost all of the fMRI studies since it was developed 30 years ago have been done in the brain. The brainstem and spinal cord are key components of the pain processing system, but early fMRI only allowed us to study the brain. Research in the Stroman Lab is quite unique, and for the past 20 years has focused on developing fMRI methods for the brainstem and spinal cord with a large degree of success. More than half of the research papers published on spinal cord fMRI have been from our lab. However, there is still more work to do in order to make fMRI really effective for pain research.

FMRI research began with studies of the visual and motor areas of the brain.  Not because these are the most interesting areas, but at the time, these were the only regions that could be imaged with MRI at high magnetic fields. (This gets complicated, but basically whole-brain receiver coils had not yet been developed for high magnetic field MRI systems). The methods were developed to be effective for regions that are essentially functioning during a task or stimulus, and go to a “baseline” level of activity otherwise. It didn’t take long for the methods to advance, and whole brain studies to be carried out, and practically every region of the brain has been studied in some way. But, generally speaking, the basic methods that assume brain activity is either “on” or “off” are still used. Also, the brainstem and spinal cord are largely left out. This “usual way” of fMRI does not work for studying pain.  As described above, pain is complicated.

Methods to acquire images for fMRI need to be adapted

The methods to acquire data that are effective in the brain are not adequate in the brainstem and spinal cord. In order to adapt fMRI for studying pain, methods needed to be selected for imaging the brainstem and spinal cord, because these regions are more challenging to image than the brain. These methods have their advantages (such as being equally sensitive to BOLD changes but with less distortions in the spinal cord) and their disadvantages (such as being slower). It is more important to acquire good data at the cost of doing it slightly slower, than to acquire a lot of poor data really quickly.

Figure(left):  These two images show the lower part of the head, and the neck, in the same person. Imagine you are looking at the person from the side, and they are facing toward the left. The image on the left was acquired with the standard fMRI acquisition method (gradient-echo EPI).  The image on the right was acquired with the HASTE method.  Note that these images are of the same person, in the same scanner, on the same day. Your choice of methods can make a big difference to your quality of data. The image on the left is severely distorted, and the brainstem and spinal cord are not well-represented.  These two imaging methods have approximately equal sensitivity for fMRI. The HASTE method is slower, but accurately represents the anatomy, which is essential for fMRI studies.

The differences in properties between tissues, bone, and air, cause distortions in MR images in the region of the brainstem and spinal cord. However, these distortions can be avoided by choosing a different imaging method (HASTE) instead of the usual method for fMRI (gradient-echo EPI). Without going into detail, the underlying theory shows that these two imaging methods can have equal sensitivity to BOLD signal changes, but the HASTE method is slower. The advantage of HASTE is that it provides accurate images of the brainstem and spinal cord.

Specialized MRI-compatible stimulation equipment is also needed

To do our studies, we also had to develop an MRI-compatible device that could apply a brief stimulus with a specific, controllable, temperature, at the exact timing that we needed.  This stimulus is set to be hot enough to be painful, but not too hot as to cause injury. We developed the “MRI-compatible robotic contact heat thermal stimulator”, called RTS-2.

The RTS-2 is a robotic device that consists of a small square ‘thermode’ that is moved up and down in a casing. This thermode is heated up to the point where it is painful to the touch, but not so hot that it could cause injury. A participant lays their hand on the casing and, with the help of custom-written software, we can have the thermode rise up and make contact with the skin on their hand at precise times. We have the thermode positioned to make contact with the meaty part of the right thumb on the palm. This is because that area is part of the C6 dermatome, meaning that the signals from those receptors are carried to the 6th cervical segment of the spinal cord, which we then image.

Specialized analysis methods are needed for the brainstem and spinal cord

After acquiring our data, we need to analyze it to be able to interpret our results and draw conclusions from them. These analysis methods also need to be adapted for the brainstem and spinal cord regions as well. Custom-written analysis software has been developed by P. Stroman for over 20 years, in both MATLAB and python (programming languages). This process is still on-going as improvements continue to be made.

Figure (right):  Coordinated signaling between brainstem and spinal cord regions in relation to pain processing. Published in PAIN, volume 159, pages 2012–2020, 2018.

The results from spinal cord and brainstem fMRI studies over many years provided new information about human pain processing.  A key result showed the expected properties of pain, and also a mis-match between the usual way of doing fMRI analysis, and pain processing.

As described above, pain signaling from the spinal cord is up/down regulated by descending signals from the brain/brainstem to adjust pain based on our situation, mood, attention focus, etc. This signaling also varies over time after a stimulus is applied, or if two stimuli are applied in quick succession, etc.  The fMRI results showed this variation, and differences between people with different pain responses.  Results also showed that this regulation is continuous – that is, it occurs all the time, even when we are not feeling pain – it sets the state of pain sensitivity depending on our situation.

FMRI study methods need to be adapted for studying pain

In addition, methods also need to be adapted for the complex process of experiencing pain. Our results showed that pain regulation is continuous – that is, it occurs all the time, even when we are not feeling pain – and it can be affected by our situation, mood, attention focus, etc. We hypothesized that this continuous component of regulation is of key importance for understanding pain, and how chronic pain conditions can develop. Our methods need to account for the fact that pain regulation is continuous (and doesn’t just occur right as someone is feeling pain).

We needed to develop fMRI methods capable of probing this continuous component.  We therefore developed the “threat/safety” paradigm, in which participants were first trained to become familiar with the study procedures, the sensations they would feel, etc.  During fMRI acquisitions, initially participants would not know what to expect, but after one minute they were told if they would feel a painful hot stimulus on their hand a minute later. This gave them time to anticipate it.  If they were to feel the hot stimulus, it was applied for 30 seconds.  After that, the study participants knew that the pain was over, and they had a few minute to rest with the fMRI acquisition continued. If they were not to feel the stimulus, the fMRI acquisition just continued, without any stimulus. 

Figure (left):  Description of the timing of stimulation used for the “threat/safety” paradigm, for fMRI studies.

The results showed that we can see responses in the brainstem and spinal cord at the time when participants were told what to expect, and differences in the anticipation period depending on whether they would feel the stimulus, or not. We saw the responses to the stimulus when it was applied. After the stimulus the responses were quite similar. That is, we can observe the continuous regulation component.

New analysis methods are needed to investigate complex pain processes

One problem with our analysis approach though, is that it does not match conventional fMRI analysis methods. The most common fMRI analysis method is to predict the BOLD response, and to analyze the data to see where it occurs. The continuous component varies between people, and depends on emotional and cognitive factors, and, for now, is somewhat unpredictable (except for the consistent features we have observed).

We therefore turned to analysis methods that are “data-driven” and do not require a predicted model of the BOLD response. 

These methods are used for “resting-state” fMRI studies, and rely on the detection of coordinated signals between regions to show which signal variations are relevant to neural function during the fMRI acquisition.

We have developed analysis methods, based on connectivity, that are capable of identifying relevant regions of the brain, brainstem, and spinal cord, and we can then investigate the timing of BOLD signal variations in these regions, and how they vary across people, or different situations (“threat” condition or “safety” condition), and how they vary with different pain sensitivities.

Now the methods we have developed are being applied to study pain

Now, we are applying these methods to study how pain processing in fibromyalgia is different than in healthy people without fibromyalgia.  We are also continuing to develop the fMRI methods that are well-matched to studying pain.