Perceiving Others’ Pain Grounded In Cognition, Not Sensation

Understanding and empathizing with others’ pain is rooted in cognitive neural processes rather than sensory ones, according to the results of a new study led by University of Colorado Boulder researchers.

The findings show that the act of perceiving others’ pain does not appear to involve the same neural circuitry as experiencing pain in one’s own body, suggesting that they are different interactions within the brain.

Senior author Tor Wager, director of the Cognitive and Affective Neuroscience Laboratory and Professor of Psychology and Neuroscience at CU-Boulder, said:

“The research suggests that empathy is a deliberative process that requires taking another person’s perspective rather than being an instinctive, automatic process.”

Empathy is a cornerstone of human social behavior, but the complex neural interactions underlying this behavior are not yet fully understood. Previous theories have suggested that the same brain regions that allow humans to feel pain in their own bodies might activate when perceiving the pain of others.

Overlapping Brain Patterns

To test this idea, the researchers compared patterns of brain activity in human volunteers as they experienced moderate pain directly (via heat, shock, or pressure) in one experimental session, and watched images of others’ hands or feet being injured in another experimental session.

When volunteers watched images, they were asked to try to imagine that the injuries were happening to their own bodies.

Thresholded LASSO-PCR

A) Thresholded LASSO-PCR (Least Absolute Shrinkage and Selection Operator-regularized Principal Components Regression) derived pattern for somatic pain; B Signature response computed as the dot product of the somatic pain pattern weights and estimated activation maps for each participant (including within-participant standard error of the mean; tUL(27) = 8.06, p<0.0001, tLL(27) = 9.20, p<0.0001, tUL-LL(27) = −0.17, n.s., for somatic pain; tUL(27) = −1.35, n.s., tLL(27) = 0.66, n.s.; tUL-LL(27) = −1.04, n.s., for vicarious pain); C Receiver Operator Characteristic (ROC) plot for two-choice forced-alternative accuracies for somatic and vicarious pain, high and low somatic pain, and high and low vicarious pain (accuracySom-Vic = 89%, p<0.0001, accuracyHsom-Lsom = 96%, p<0.0001; accuracyHvic-Lvic= 50% n.s.). Credit- DOI:

The researchers found that the brain patterns when the volunteers observed pain did not overlap with the brain patterns when the volunteers experienced pain themselves.

Instead, while observing pain, the volunteers showed brain patterns consistent with mentalizing, which involves imagining another person’s thoughts and intentions.

The results suggest that within the brain, the experience of observing someone else in pain is neurologically distinct from that of experiencing physical pain oneself.

“Most previous studies focused only on the points of similarity between these two distinct experiences in a few isolated brain regions while ignoring dissimilarities. Our new study used a more granular analysis method,”

said Anjali Krishnan, the lead author of the study and a post-doctoral research associate in the Institute of Cognitive Science at CU-Boulder while the research was conducted. She is currently an assistant professor at Brooklyn College of the City University of New York.

A Departure From Shared Representation

This new analysis method identified an empathy-predictive brain pattern that can be applied to new individuals to obtain a brain-related ‘vicarious pain score,’ opening new possibilities for measuring the strength of activity in brain systems that contribute to empathy.

The results may open new avenues of inquiry into how the brain regions involved in empathy help humans to relate to others when they experience different types of pain. Future studies may also explore the factors that influence one’s ability to adopt others’ perspectives and whether it might be possible to improve this ability.

Writing in response the paper’s submission, the journal editors commented:

This is a very important paper. It tackles a finding that in no short order became dogma and even slipped into popular ‘fact.’ The authors clearly demonstrate that the conflation of the brain representations of self and other pain is not warranted and that has been ‘verified’ without the simple but telling tests that the authors use in this paper. In particular, the authors compare 3 different intensities and 2 different locations and then toss in two additional modalities at single intensity and location. The work is compelling and shows that the somatic pain signature is different from the vicarious pain one.

They also requested the authors to expand on why they think that previous studies have come up with a different answer. The response:

Most previous studies of vicarious pain point out its similarity with somatic pain, and it is thus widely believed that the two experiences rely on the same systems. Why are our findings and conclusions different? There are three main reasons.

First, many previous studies focused only on the points of similarity—mainly identified in two isolated brain regions, dACC and aINS—ignoring dissimilarities [but cf. (Lamm et al., 2011), who write, ‘our results reveal more differences than similarities…’ p. 2500]. Here, we aimed for an unbiased assessment of the two processes.

Second, most previous work has identified ‘pain-related’ activation by contrasting pain with loosely matched control conditions (e.g., neutral or innocuous stimuli). Overlapping activity in such contrasts may be caused by many processes that are not pain, including general negative affect, attention, and arousal [see (Wager et al., 2016) for discussion]. Engagement of these processes may be responsible for the similar activation in previous work. By contrast, in our study, we attempted to isolate pain-relevant patterns that predicted the magnitude of experienced intensity, and examined the similarity of those patterns.

Third, most previous studies focused on voxel-wise activation ‘blobs,’ rather than on multivariate patterns, which can be sensitive to information at finer spatial scales (Shmuel et al., 2010), possibly even below the intrinsic resolution determined by the voxel size (Kamitani and Tong, 2005). This property, combined with our experimental approach targeting within-person variations in pain intensity, suggests that the multivariate patterns we identified are more likely to reflect specific representations contained in meso-scale neural circuits.

Based on our findings, we infer that the overlapping activation in the dACC, aINS, and other areas is not related to shared pain experience. Interestingly, on close reading, the few previous multivariate pattern-based studies agree broadly with this interpretation. The brain patterns they identified as shared across somatic and vicarious pain were not specific to ‘pain,’ as these patterns were also activated by other, non-painful types of negative affect (Corradi-Dell’Acqua et al., 2016; Zaki et al., 2016). As in our study, in these studies what is shared does not seem to be particular to pain per se.

Anjali Krishnan et al
Somatic and vicarious pain are represented by dissociable multivariate brain patterns
eLife (2016). DOI: 10.7554/eLife.15166

Photo: Alex Guerrero/Flickr