Mapping the Neural Pathways of Feelings


by Jim Schnabel

January 19, 2010

The realm of the senses extends far beyond the five basic modes of sight, hearing, touch, taste and smell. The mammalian brain also is wired for the senses of balance (equilibrioception), temperature (thermoception), and body position and orientation (proprioception). Particularly over the past decade, neuroscientists also have been mapping the vast territory of “interoception,” which gives a person a sense of the physiological condition of his body, and includes hunger, thirst, pain, and the need to breathe, among many other bodily feelings.

Researchers are keen to know more about interoception because it is widely believed to form the foundation of our emotions and behavioral motivations. “So for example, if I’m hungry, I want to eat. If I’m exhausted, I want to rest. If I feel excited, I want to feed that excitement,” explains Martin Paulus, a psychiatrist and neurobiologist whose work at the University of California at San Diego includes research on interoceptive brain networks. According to the influential “somatic marker” hypothesis, even relatively abstract decision-making makes use of an interoceptive sense of possible outcomes:  Someone about to place a large bet, for example, would be likely to experience, again, some of the physiological stress and despair he experienced the last time he gambled away his money.

An inability to “feel” potential outcomes, because interoceptive brain networks are not working properly, has been linked to poor decision-making skills.  But sometimes too much feeling can be a problem and interrupting interoceptive networks may be part of the solution.  A study by Antoine Bechara and colleagues at the University of Southern California, published in Science in 2007, found that some people with damage to their insular cortex, or insula, a key region in processing interoceptive signals, abruptly lost their cravings to smoke cigarettes and were able to kick their habits relatively easily.

Does interoception have its own pathways?

One unanswered question about interoception is whether it occurs solely via its own networks of nerves and brain regions or whether it also makes use of other sensory networks, such as those that convey the sense of touch.

A much-cited brain-imaging study in 2004, led by Hugo Critchley at University College, London, linked activity in the right front insula with the ability of people to judge the timing of their heartbeats, a classic interoceptive task. But a recent study by neuroscientist Daniel Tranel’s laboratory at the University of Iowa suggests that the boundaries between interoception and other sensory modes may be a bit blurred.

Published online Nov. 1 in Nature Neuroscience, the study focused on a patient named Roger who had severe right insula damage, but was nevertheless able to feel his heart beating via tactile nerves in his chest, albeit with a slight delay. When doctors put an anaesthetic cream, lidocaine, on his chest to block the sensitivity of those tactile nerves, Roger lost the ability to sense his heartbeat.

“This study shows that there are multiple independent pathways in the brain that provide self awareness of this sensation, and there is reason to suspect that a similar pattern exists for other body sensations as well,” says Sahib Khalsa, lead author on the study.

At least one prominent interoception researcher disagrees. A. D. “Bud” Craig, of the Barrow Neurological Institute in Phoenix, Arizona, argues among other things that “the test performed by Khalsa and colleagues was not a test of interoceptive awareness.” In a departure from the 2004 experiment, the Iowa researchers first gave Roger a heartbeat-amplifying drug, a chemical cousin of adrenaline known as isoproteronol. Without such a drug, they had noted, only a minority of ordinary subjects can reliably detect their own heartbeats. But at the relatively high doses of the drug used in their experiment, notes Craig, “it’s similar to letting somebody take his hand and beat on your chest wall from the outside; so it’s no surprise that it activates mechanoreceptors [tactile nerve endings] and proprioceptors [muscle stretch receptors] in the chest wall.”

In several high-profile papers starting in 2002, Craig has argued for a view of interoception as a fundamentally separate sense. “Our cortex has a different pathway for mechanoreceptors and proprioceptors that goes up to the somatosensory cortex, which is really involved in controlling skeletal muscle,” Craig says. “The interoceptive receptors, for pain and temperature and itch and sensual touch and so on, which represent the physiological condition of all tissues of the body, go to the insula, because it’s controlling smooth muscle” over which we do not have voluntary control. This anatomical separation of nerve pathways can be found in the spinal cord of all vertebrates, Craig notes, “so it is a 500 million year old division.”

Paulus, on the other hand, is more open to the possibility that interoception can involve some crossover from other sensory pathways: “Just because you have that one pathway doesn’t mean that it is the exclusive road. What [Khalsa and colleagues] are saying is that there might be some alternative routes, which I think makes a lot of sense, because it’s such an important process that you wouldn’t rely on just one pathway.”

Paulus emphasizes, though, that “we don’t fully understand the system yet” and that it is inherently difficult to separate these pathways in experiments. Patients with brain damage, for example, almost never have damage that is perfectly restricted to one structure or network. Moreover, damaged brains often adjust for lost functions by forming new pathways that would not be found in normal subjects. “Since Roger’s damage occurred many years ago, it is possible that his brain compensated by rearranging the functioning of intact brain areas,” Khalsa says.

Future experiments should resolve such questions.  In so doing they might also shed light on one of Craig’s boldest hypotheses, which is that interoception forms the basis for what we know as the self.  Interoceptive nerve pathways from the body end at the rear part of the insula, where body-like neuronal maps represent the various interoceptive sensations.  According to Craig, these “somatotopic” maps of feeling are represented and re-represented with greater complexity in the mid and frontal insula, as they gather other important sensory, emotional and cognitive inputs. Finally in the foremost part of the insula, he argues, they constitute the highly-integrated self, of which we are consciously aware when awake.

“People have trouble believing that there’s just one specific region in the brain that supports awareness,” Craig says. “It doesn’t fit with the ideas that some others have published, on the so-called neural correlates of consciousness. But we’ll see how things fall out.”