I have not written for some time. This was partly due to a nice long holiday I spent with my wife and son in Southern France surfing and relaxing (as much as the latter is possible with a three-year-old child). After my return I have been busy managing the everyday madness of running a research group at a university that is almost broke. The good news is, my group is slowly growing. I have two new students, one who started yesterday and another one starting in November. One is a medical doctor from Syria, who comes on a full stipend from German Academic Exchange Service. Lucky me! Both will work with a new software application of ours that allows patients to make precise drawings of their bodily sensations. Such sensation maps can then be used to study phenomena like referred pain or the deqi sensation of acupuncture. But that is something for a future post. Today, I would like to talk instead about an important topic in neuroscientific research on acupuncture and related techniques, namely attention. You may ask, what such a seemingly unrelated concept might have to do with acupuncture or other somatosensory therapies. The answer is that attention is an important confounding factor in any experiment using functional brain imaging. To fully understand this, some basic neuroimaging facts may be helpful.
The basic principles of functional brain imaging
Let me quickly outline the basic setup of a functional MRI experiment. It always goes something like this: A subject is put in an MRI scanner and measured continuously while being stimulated or performing a task. The way this stimulus or task is presented follows a fixed time-course. After the experiment is finished, the researcher analyzes the MRI data looking for brain areas that exhibit a similar time-course. Such areas are thought to be involved in the processing of the stimulus or the performance of the task. The results are processed by statistical software that produces color-coded maps of the similarity between stimulus and brain activity time-courses. These maps are overlaid on a high-resolution structural brain image and voila – you get brains with glowing “blobs” (just like the ones below).
The brain does many things at once
Of course there are problems inherent to this approach. For example, you never know if the brain activity taking place during stimulation is truly relevant to the processes you are interested in. Brain imaging does not distinguish between brain events that are causally related or simply happen at the same moment by chance. But even if all measured brain activity is directly related to the experiment, the result may still be difficult to interpret. Think of an experiment, where you want to find out which brain areas mediate the analgesic effect of acupuncture. You stick a needle in your volunteer and twist it during pre-defined periods of time. Later you take the time-course of stimulation to analyze your data. So far so good. Now, let’s reflect on what is going on in the volunteer’s brain in that moment: He or she will surely experience some level of anxiousness that may improve or worsen upon needle insertion. There will be all sorts of expectations depending on prior experience and information of the volunteer. Stimulation of the needle will trigger a response in the somatosensory and in particular the pain system. At the same time, the body will react with changes in autonomic outflow depending on how much the needle is interpreted as a threat. Finally, the body will direct a considerable amount of its attention to the part, where the stimulation takes place. After all, the insertion of a needle is a rather unusual and potentially harmful event.
All these cognitive processes – anxiousness, expectation, attention, sensation, and autonomic reaction – happen at the same time. And all of them will contribute to the final image of brain activity that we get from our fMRI experiment. One could argue that to gain a detailed understanding of acupuncture’s underlying mechanisms, we must not neglect any of the above mentioned brain processes. However, it is questionable if all cognitive processes that are triggered by insertion of a needle are truly relevant for the observed clinical effects of acupuncture. Furthermore, brain images do not tell us, how much of the activity we see is related to each of the cognitive processes. Thus, it could be that unspecific effects, like attention dominate the results in which case most of the observed brain activity changes must be considered unspecific.
How to disentangle attention from other brain reactions to acupuncture
While there has been a considerable number of brain imaging studies trying to elucidate the underlying mechanisms of acupuncture, few have explicitly controlled for attention as a confounding variable. I have criticized this in one of my early papers, which evolved from my fruitless attempts to replicate acupuncture-fMRI results published by other groups. In particular, the group of Zang-Hee Cho had shown activations of the visual cortex upon stimulation of acupuncture points that are used to treat eye diseases. While this may sound like a plausible mechanism at first, I soon developed doubts, when I noticed that the visual cortex is frequently activated as a result of shifts in attention. Neither Cho’s nor any other study replicating his results had controlled for the effects of attention. Thus, instead of the noteworthy explanation “acupuncture affects the eye via the visual cortex”, I offered the more likely but less sexy alternative “acupuncture causes changes in attention that activate the visual cortex.
Recently, a Korean colleague of mine, Younbyoung Chae, published what I believe is the first study to clearly separate the effects of attention from other brain processes related to acupuncture. The title of his paper, which appeared in Nature‘s new open access journal Scientific Reports is Cortical Activation Patterns of Bodily Attention triggered by Acupuncture Stimulation.
In a very elegant study design he and his colleagues performed two fMRI scanning sessions: one with genuine acupuncture stimulation and the other one without. Their subjects, however, believed that they would receive acupuncture in both sessions. To make sure that subjects were really paying attention to the stimulation, Chae and colleagues used a trick. Instead of just stimulating a single point, they used two neighboring points on the inner side of the arm (PC6 and HT7) and asked subjects after each (pseudo-)stimulation to rate which of them had just been stimulated. Then they compared the brain activity changes caused by stimulation and pseudo-stimulation, respectively.
Brain processes underlying acupuncture with and without needle stimulation are largely overlapping
The results from their study showed that brain networks engaged in acupuncture and pseudo-stimulation are largely overlapping. To quantify the exact similarities and differences, Chae and colleagues calculated a so-called conjunction as well as a difference map for both kinds of stimulation. The analysis confirmed that the majority of brain areas activated by acupuncture stimulation is also activated if subjects expect acupuncture but never receive it.
The cortical areas showing this unspecific brain activity include the anterior cingulate cortex and the anterior insula, both of which belong to the so-called salience network. This is an ensemble of brain regions that encode the salience of sensory inputs, i.e. it helps selecting specific stimuli for additional processing by identifying those of potential relevance.
[…] we demonstrated a common brain network involved in the top-down modulation of acupuncture stimulation triggered by bodily attention.
But the authors also found differences between the two kinds of stimulation that – according to the authors – represent the “neural correlates of afferent signal processing”. That means, they reflect brain activity related to the processing of actual bodily sensations triggered by needle stimulation. These cortical regions included the primary and secondary somatosensory cortex as well as the midcingulate cortex and posterior insula. All these regions are known to be involved in processing of somatosensory stimuli.
A small weakness of the study, not discussed by the authors, is their rating system for stimulation intensity. It is unclear if subjects felt something under pseudo-stimulation, as the scale used for rating did not include a zero and the paper does not state if 1 means “no sensation” or “weak sensation”. Although you may think “Why should people have sensations without physical stimulation,” I can assure you from my own experiments that people report all sorts of sensations when expecting to feel something.
What is the importance of bodily attention in acupuncture?
In the end, we have to ask ourselves, to what extent this study can answer the question, what role attention plays in acupuncture. As the authors conclude,
These findings suggest that the components of enhanced bodily attention around the acupoint are significant in the neurophysiological response to acupuncture stimulation.
Put simply, this means than attention causes the major part of the changes in brain activation observed during acupuncture stimulation. The experiment was not designed to answer the question, if attention-related brain processes are part of the mechanism by which acupuncture achieves its clinical effects. So, as is often the case, more research is necessary.
From my own point of view, directed bodily attention plays a significant role in the underlying mechanisms not only of of acupuncture, but of most somatosensory therapies. It is my firm belief that the act of directing the patient’s attention to the body by means of stimulation is an important part of any such therapy. In fact, as a common theme of acupuncture, moxibustion, massage, cupping, manual therapy and even touch healing it may be a thread linking those seemingly unrelated therapies.