Emotion Fluctuations in Anorexia and Bulimia Nervosa: A Rollercoaster or Not So Much?

Restricting, bingeing, and purging are powerful ways to regulate emotional states. However, these behaviours probably play different roles in emotional regulation. Whereas restriction is hypothesized to pre-empt the onset of highly emotional states, bingeing and purging is thought to act as a coping mechanism to deal with overwhelming emotional states once they’ve already been activated. 

In BN there is abundant evidence that the binge–purge cycle functions as a means of emotion regulation. Binging [and I would arguing purging too! ] facilitates a temporary suppression of painful self-awareness and helps the self to dissociate from painful emotions or to block negative affect as demonstrated in both laboratory studies and diary studies in daily life.

In AN, restrictive eating patterns have been linked with a narrowing of emotional functioning, flattening of affect and lack of outward display of emotion. As such, Waller, Kennerley, and Ohanian (2007) argue that both binge– vomit-cycle and restriction are emotion suppression strategies which are just utilized at different times.

If this is true, we would expect to see more variability (or fluctuation) in the intensity and types of emotional states experienced by bulimia nervosa patients as opposed to restricting-type anorexia nervosa patients. The problem is that we know asking people to tell us how they’ve felt in the last week is not a very accurate way to get data. Our memories are not very accurate, we are prone to forgetting, and we are prone to bias, too.  Besides, what we often report is often (whether we know it or not) strongly influenced by how we want to appear and the image we want to portray — both to others and to ourselves.

So how can researchers get around that? By using something called ecological momentary assessment (EMA). I’ve written about studies that have utilized EMA before (here and here) but it is worth repeating what it is and why researchers might want to use it:

Ecological momentary assessment (EMA) involves repeated sampling of subjects’ current behaviors and experiences in real time, in subjects’ natural environments. EMA aims to minimize recall bias, maximize ecological validity, and allow study of microprocesses that influence behavior in real-world contexts. EMA studies assess particular events in subjects’ lives or assess subjects at periodic intervals, often by random time sampling, using technologies ranging from written diaries and telephones to electronic diaries and physiological sensors. […] EMA holds unique promise to advance the science and practice of clinical psychology by shedding light on the dynamics of behavior in real-world settings. (Source)

In this study, Vansteelandt and colleagues utilized EMA to find out if there are differences in variability in  the intensity and/or types of emotions experienced by anorexia nervosa and bulimia nervosa sufferers.

(What I really liked about the study, is that the authors utilized two newly developed concepts called pulse and spin in order to quantify and measure emotional (affective) variability. Pulse refers to the variability in the intensity of affect (happy or really happy) whereas spin reflects the variability of types of emotional states (sad versus happy). I would strongly urge those interested to check out the paper for more details. I think it is a nice way to conceptualize, analyze, and even visualize data.)

The participants were 21 AN-R, 16 AN-BP, and 20 BN sufferers; the mean age of the participants was 21.3. The participants had to answer a few questions 9 times a day over the course of a week when they were signalled to do so on a handheld computers they were given.


  • No differences between groups in the variability of emotion intensity (Figure 1)
  • Significant differences in the variability of types of emotions experienced (Figure 2)
Vansteelandt - 2013 - Figure 1

The authors then wanted to see whether distinguishing the groups not by their eating disorder diagnoses but by the presence or absence of purging behaviour would yield the same results. As you can see in the Figures below, it did.

MAIN RESULTS: Restricting (AN-R) vs. Purging (AN-BP and BN)

  • No differences in the variability of emotion intensity (Figure 3)
  • Significant differences in the variability of types of emotions experienced (Figure 4)
Vansteelandt - 2013 - Figure 2


The results from the study suggest that individuals with AN-R, AN-BP, and BN experience the same variability (change) in the intensity of their emotions but that individuals with AN-R have reduced variability in the types of emotions experienced. In other words, it seems that all groups have the same variability in the extent to which they feel a certain emotion (happy to really happy, sort of sad to devastated) but that individuals who restrict (AN-R) have a shallower range of the types of emotions they experience — they tend to stick to the same emotion and not fluctuate as much as the AN-BP and BN groups between positive and negative emotions.

The results also suggest that distinguishing between restricting versus purging type might be more beneficial when it comes to examining emotional regulation (as well as other things) than distinguishing by DSM criteria (no surprises here!)

This suggests that restricting might not necessarily serve as a way to pre-empt the activation of emotion per se but rather restrict the types of emotions experienced. So they tend to stick to the same type of feeling (happy or sad, as opposed to fluctuating between them, but the change in the intensity of feeling happy, for example, is no different between the diagnostic groups).


As always, I like to mention the limitations of the study and caution against reading too much into the data. The study is exploratory and the sample sizes are small. It is possible that differences in the variability of the intensity of emotions would become apparent once you study a large enough sample.

It would also be interesting to see what the the data for a healthy control group would look like, as well as for individuals with other disorders that are highly comorbid with different eating disorders, such as obsessive-compulsive disorder and borderline personality disorder.

There is a lot that I’d love to see explored in future studies. For example, would the results change if you follow individuals for a month? Would that add enough richness and information to the data to be able to untangle finer details of emotional regulation between different eating disorder groups? Moreover, what would happen if you follow AN-R patients as they go through treatment and restore their weight? What about tracking the changes that occur when individuals experience diagnostic crossover?

Moreover, a larger sample size would allow researchers to stratify and group individuals based on the presence or absence of other behaviours that commonly co-occur with eating disorders, such as self-injury or based on experienced in early childhood (like abuse).

Ah, there’s so much to explore! I’m excited for future studies on these topics, and I admit, a large part of it is because I just really like the kind of data you can gather from EMA studies. Perhaps it is because having filled out questionnaires before that asked me to assess or summarize my mood or mental state over some certain time period, well, I have little faith that the data gathered from the questionnaires can do little more than provide a rough picture of what is going on.


Vansteelandt, K., Probst, M., & Pieters, G. (2013). Assessing affective variability in eating disorders: Affect spins less in anorexia nervosa of the restrictive type. Eating Behaviors, 14 (3), 263-8 PMID: 23910763

Liked it? Take a second to support Tetyana on Patreon!


Tetyana is the creator and manager of the blog. She has an Honours BSc in Neuroscience and an MSc in Medical Science. She can be reached at tetyana[at]scienceofeds[dot]org.


  1. Although I do agree that just by asking people how they feel is not going to get the researchers accurate results, I do feel like the researchers could have used a better research method than EMA to get more accurate results. By using EMA I feel that there is not enough random sampling going on. We need to choose small groups of people from different areas in the country, with different weather, different races, age,gender, and activities around because that could also play a role in the emotions or behavior. You could go about the study with a naturalistic observation that way these people are being observed in their natural everyday surroundings. I think to confirm what your observing you can conduct a survey with absolutely no bias in the question to confirm what you have observed. Then compare each participants from different the different areas in the country and the different environments to confirm that emotions do fluctuate in anorexia and bulimia.

    • Hi Megan,

      Thanks for your comment. While I agree that your proposals certainly would make for a richer data set, I do think they are a tall order.

      There are several points I want to make here:

      1. This was an exploratory study to see if there is even anything there to study further.

      2. I did not mention this in the post, but a MAJOR part of the study was looking at analyzing the data using the concepts of pulse and spin as opposed to what has been done before with EMA studies and testing whether a different way of analyzing data would provide more information (it did). I omitted all of that from the post even though it was a large part of the study. And, actually, a part I personally found quite interesting but thought was probably too technical to go into depth in this post.

      3. You need to establish that there ARE differences between two groups before you can start analyzing what can moderate or modulate those differences. Doing a complicated study with many variables (weather, gender, etc.) would create a lot of confounds, and I think you’d need quite a large sample size to be able to tease all of that information apart. Besides, how can we look at how those things play into emotions if we don’t know if there are differences between AN-R, AN-BP and BN? The purpose of the study was to test a specific hypothesis, which wasn’t how weather, gender, or environmental activities affect emotions.

      4. When you are doing qualitative studies (and I’d argue even quantitative ones)there is always bias. Bias isn’t a dirty word, but a lot of the time we want to either minimize it or be aware of it.

      5. It costs a lot of money and takes a lot of time to conduct large studies. I don’t think it is reasonable to expect that given that we have no idea of what the funding levels for this study were. Of course, if we had unlimited money we could do all sorts of things, but conducting a large study is *a lot* of money.

      6. I find studies that test specific hypotheses like this to be much more satisfying than ones that test a whole lot of things and then find false positives all over the place because they don’t properly correct for multiple comparisons.

      7. I don’t see the purpose of doing a naturalistic observation given that we are interested in how individuals feel about their own emotional state, not how they appear to be. I think reporting 9x a day when prompted does a fine job.

Leave a Reply

Your email address will not be published.

2 × 4 =