When we first began our research on nonsuicidal self-injury (NSSI; e.g., self-cutting, burning) around 2007, a major question was why anyone would engage in these behaviors. Seminal work by Drs. Matt Nock and Mitch Prinstein (2004/5 in J Consulting & Clinical Psych, J Abnormal Psych) suggested that there were many potential reasons, but "to reduce bad feelings" was the most commonly endorsed. Other researchers later obtained similar findings, but our major question was why people reported this -- pain normally makes people feel worse, not better. There were many possible explanations, including: 

1. This is a self-report bias whereby people are justifying their behavior by saying it makes them feel better when it, in fact, does not

2. People who engage in NSSI are somehow "wired differently" to enjoy pain

3. Pain is a potent distractor, so if people are feeling bad, then maybe NSSI distracts them from these negative thoughts and feelings

As described below, our experimental work did not support any of these answers.

Do people who engage in NSSI feel better after pain?

Answer: Yes... but so does everyone else (and so do most other species).

Our first study set out to answer this question with a complex experimental and psychophysiological paradigm. We measured a physiological indicator of negative emotion/defensiveness (startle eyeblink reactivity) at baseline, during preparation for a stressful speech task, and immediately after a painful task. Our goal with this paradigm was to try to mimic an NSSI episode (baseline, stress, pain), and we wanted to use a physiological measure to get around potential self-report biases. We expected that people with no history of NSSI (Control #1) would feel worse after pain (indicated by higher startle reactivity) and that people with a history of NSSI (NSSI) would feel better after pain (indicated by lower startle reactivity). Because we were worried about a phenomenon called startle habituation (i.e., where startle reactivity drops throughout an experiment) contaminating our results, we gave everyone a second dose of stress and pain. Because we were worried that something other than pain might affect our results, we ran a second control group with no history of NSSI that went through the exact same paradigm except that they sat in a room for a few minutes instead of undergoing the painful task (Control #2).

The basic results are presented below:

Startle Eyeblink reactivity across the experiment

(note: higher startle eyeblink reactivity = more negative feelings)

As expected, we did see some startle habituation throughout the first part of the experiment (see Lane, Franklin, & Curran, 2013 in Int J of Psychophys for more info on this phenomenon); however, results across the latter part of the experiment were clear. These results showed that, yes, people who had a history of NSSI did actually feel better after pain.... BUT, so did the group without any history of NSSI (Control Group #1). As indicated by Control Group #2, which didn't undergo any pain during the experiment,  the drop in negative feelings in the NSSI and Control #1 groups was likely due to the painful stimulation these groups received.

Because people who engage in NSSI find pain itself to be unpleasant (see the SITB Barriers portion of this website) and because we measured startle eyeblink reactivity after the removal of pain, we knew that something about this "pain removal" probably made people feel better. Upon searching through the basic pain literature, we happened upon studies dating back to the 1950s showing that the removal of pain causes intense relief -- this effect is found in humans, rats, fruit flies, and many other animals. We realized that we had accidentally re-discovered this same effect; we call it pain offset relief. One potential explanation for this is evolutionary:

(a) We find pain and things associated with pain to be unpleasant because pain is indicative of injury and potential death.

(b) HOWEVER, the removal of pain should be reinforced... whatever action you take to get away from something painful and potentially deadly is important to remember and to repeat again if you encounter a similar situation in the future.

For example, putting your hand in a fire is painful -- it should feel very unpleasant and that unpleasantness should motivate you not to do it again. But pulling your hand out of the fire should be rewarded; you should definitely be motivated to pull your hand away should it ever wind up in a fire. This study showed that NSSI may work similarly: the pain involved in NSSI is bad, but the removal of pain is a pleasant experience. For most of us, pain offset relief with NSSI is not a great option because it involves pain and other bad things, but if someone is having a hard time feeling better, it may seem like an attractive option.

Recommended citation for this information: Franklin, J.C., Hessel, E.T., Aaron, R.V., Arthur, M.S., Heilbron, N., & Prinstein, M.J. (2010). The functions of nonsuicidal self-injury: Support for cognitive-affective regulation and opponent processes from a novel psychophysiological paradigm. Journal of Abnormal Psychology, 119, 850-862.

Does pain offset relief reduce negative feelings or increase positive feelings?

Answer: Both, at the same time.

A big question in both the basic emotion literature and in the NSSI literature concerned whether relief = an increase in positive feelings, a decrease in negative feelings, or somehow both. Because it is very hard to measure positive and negative feelings at the same time (and people often confuse changes in these on self-report measures), this question had been difficult to address. So we developed an experimental paradigm that involved physiological measures specific to changes in negative feelings (startle eyeblink reactivity) and positive feelings (postauricular reactivity). Importantly, these measures have a near-zero correlation, suggesting that changes in negative feelings may not necessarily accompany changes in positive feelings (and vice versa). Compared to our 2010 study noted above, we used a much more fine-grained pain offset relief paradigm that was better able to examine the time course of pain offset relief and the relationship between pain intensity and pain offset relief.

The basic results were as follows:

Similar to our earlier study, startle eyeblink reactivity dropped substantially after pain offset. Adding to that study, here we found that postauricular reactivity simultaneously increased by nearly 50%. We also found that eyeblink effects subsided after several seconds, but postauricular effects remained strong. Overall, these findings suggested that pain offset relief is associated with both increased positive and decreased negative feelings.

Recommended citation for this information: Franklin, J.C., Lee, K.M., Hanna, E.K., & Prinstein, M.J. (2013). Feeling worse to feel better: pain-offset relief simultaneously stimulates positive affect and reduces negative affect. Psychological Science, 24, 521-529.

Is NSSI experience associated with heightened pain offset relief?

Answer: No... pain offset relief seems to be the same for people who have never engaged in NSSI and people have have engaged in NSSI thousands of times.

Pain offset relief seems to follow an "opponent process" pattern (see Solomon, 1980 and Joiner, 2005 for more on opponent processes). One of the main features of this kind of pattern is that the opponent process (here, relief) increases when the primary process (here, pain) is stronger. Although it isn't detailed above, the Franklin et al. (2013, Psych Science) study did not detect this pattern -- if anything, less intense pain was associated with greater relief. This led us to wonder about a second prediction of the opponent process theory: across repeated experience with a primary process (here, pain), the opponent process (here, relief) should become stronger. This has been proposed as one of the major avenues through which prior self-injurious behaviors (and painful and provocative experiences more generally) increase risk for future self-injurious behaviors. In other words, across repeated experiences, self-injury should become less painful/scary and more relieving. We tested this by examining pain offset relief in people with and without a history of NSSI, and by examining the correlation between lifetime NSSI episodes and pain offset relief magnitude.

Here's what we found:

Results again showed that pain offset relief was associated with simultaneous diminished negative feelings (eyeblink) and increased positive feelings (postauricular). However, pain offset relief was similar regardless of NSSI history. This was further highlighted by correlation analyses, which showed that a greater number of NSSI episodes was not associated with increased pain offset relief (and sometimes associated with slightly less). These findings suggested than pain offset relief may not work as an opponent process, and that increased pain offset relief is probably not an explanation for the link between prior and future self-injury. These findings also suggested that pain (and pain offset relief) does not work purely as a distractor because pain offset relief was associated with increased positive feelings whereas distractors would only be hypothesized to reduce negative feelings.

Recommended citation for this information: Franklin, J.C., Puzia, M.E., Lee, K.M., Lee, G.E., Hanna, E.K., Spring, V.L., Prinstein, M.J. (2013). The nature of pain offset relief in nonsuicidal self-injury: A laboratory study, Clinical Psychological Science, 1, 110-119.

Future Directions

This work has helped us to understand why some people engage in NSSI, but it's not the whole story. In our opinion, the most interesting work in this area is currently being conducted by our collaborator, Kathryn R. Fox, in the lab of Dr. Jill M. Hooley at Harvard University. She is directly comparing pain offset relief, distraction, and self-criticism validation as potential benefits of NSSI.

Our own ongoing work seeks to understand the potential (perceived) benefits of suicidal self-injury and the link between prior and future self-injury.