How Emotions Get So Out of Control in Borderline Personality
A new comprehensive model of borderline tracks the dynamics of its emotions.
Posted Jul 21, 2020
For people with borderline personality disorder (BPD), loss of control of negative emotions, otherwise known as emotion dysregulation, is a regular fact of life. An event triggers anger, fear, or sadness which, in turn, further stimulates a process of rumination in which they continue to dwell upon how upset they are. In the process, the emotion’s intensity increase. At some point, individuals find that the only way to turn down the intensity is to shift their attention to some sort of physical outlet, such as harming themselves. When the next event begins the cycle all over again, the individual once again is at risk of physical harm.
If you know someone with borderline personality disorder, you are undoubtedly aware of how disabling this cycle of events can be. When the self-harm can take extreme forms, the individual may go so far as to make a suicide gesture or attempt. Talking the person down from this peak of emotional intensity can take far more energy and ability than you can muster, try as you might to re-establish some form of equanimity and calm.
According to Rutgers University Edward Selby and colleagues (2020), research in psychopathology in general, and personality disorders in particular, is limited in its ability to provide an empirical understanding of this cycle from emotion dysregulation to rumination to self-harm. As Selby et al. note, this limitation is due to the inappropriate use of statistical methods to test the chain of linkages from emotions to behavior. In the words of the authors, these often make the real world conform to our modeling methods, rather than the other way around.”
To explain this problem, consider what happens in a typical research study of a patient population. The investigators might obtain one measure of, say, emotion dysregulation (how out of control the individual feels) and relate scores on this measure to incidents of self-harm. The research may also take into account the extent of the rumination. Plugging these scores into a statistical program leads to a correlation, or estimation of the extent to which the three factors all relate to each other.
However, in the “real world” of people with borderline personality, these factors don’t occur simultaneously. That chain leading from emotion through rumination through outcome behaviors occurs in a progressive fashion with a considerable degree of feedback going “backward” from self-harm to emotional escalation. An estimate at one point in time won’t be able to test how much this negative loop makes the individual’s symptoms that much worse.
To overcome this limitation of single-point statistical approaches, the Rutgers University team suggests using a completely different method called “Bayesian” estimation. Consider that you have 3 factors (e.g. emotions, rumination, and self-harm). Call these A, B, and C. In this Bayesian method, the researcher calculates the likelihood of “B” happening, given that “A” has occurred. This updating of information based on prior events allows researchers to examine the conditional probabilities of events as they happen over time.
Even without knowing the ins and outs of the statistical method used to establish the pattern of "A->B->C" relationships, it is probably clear to you by now that such a process-oriented approach has the potential to capture the kind of emotional intensification that occurs when people react to events that provoke the reactions that build on each other over time. The Rutgers investigators use the term “emotional cascade model (ECM)” to refer to “the positive feedback process between negative emotion and rumination [that] results in progressively intensifying levels of both negative emotion and rumination.”
Using the Bayesian model to understand the ECM can be valuable because, in the words of the authors, “the ECM posits that there are simultaneous, bidirectional effects of rumination and negative emotion on each other over time, which remains difficult to provide evidence for.” Indeed, the feedback model proposes that it isn’t just A which can lead to B, etc., but that B can feed back on A in a reverberating fashion. The new statistical method not only can estimate these effects but it can also provide a test of the entire process, or the “gestalt” level that takes into account all factors at a time.
To establish the validity of the ECM, the authors enhanced their approach by using a data collection method known as “experience sampling.” In this method, participants complete assessments several times a day to report on their current state (“experience”). Other studies have used experience sampling as well, but not with the type of statistical tests that could allow for those feedback loops to be estimated. Additionally, the research team wished to determine how specifically the ECM would apply to people with borderline personality disorder as compared to other disorders involving self-injurious behavior and rumination; namely major depressive disorder and post-traumatic stress disorder.
The study’s participants were 47 self-injuring adolescents and young adults ages ranging in age from 15 to 21 years old, (Mean age of 19.1 with most between 17 and 21). After completing diagnostic assessments, the participants received a smartphone mobile application which would deliver random prompts during the day to complete brief (3-5 minutes) reports on their experiences during the previous 2 to 3 hours. These assessments were conducted over a 2-week period.
In each momentary assessment, participants rated their affect along 19 dimensions (using a 0-10 scale) that included the following 11 negative affect items: sad, angry, hurt/emotionally rejected, frustrated, anxious/ afraid, lonely, empty/numb, guilty, physically numb, ashamed, and overwhelmed. They also rated how much they had ruminated, or dwelled upon, each emotion, defined as thinking that is “repetitive, passive, difficult to control, and negatively focused.”
Finally, participants indicated whether they had engaged in the following “dysregulated” behaviors, including alcohol use, illicit drug use, impulsive shopping, or binge eating, items that were summed to create a total scale score.
Before moving on to the findings, you can try a thought experiment yourself to see how you would respond, at different times during your own past day, to each of these 3 types of scales. You can also ask yourself, further, whether your answers now would be the same or different than your answers a few hours ago. If not, what events might have transformed your mood? Then ask yourself whether having been put in a bad mood led you to ruminate more, which further accentuated your bad mood. In terms of outcomes, did you then find yourself turning to some online shopping or perhaps having an extra snack?
Turning now to the results, the authors found that, using their Bayesian model, they could accurately predict a BPD diagnosis in 90 percent of the cases and allowed for a distinction to be made between the other two diagnostic categories. Then, taking advantage of the backward-step method in the statistical method, Selby and his collaborators were able to predict, once the diagnosis was factored in, the extent to which earlier levels of rumination and emotion would lead to an increase in later levels of those same ECM components. Indeed, perhaps even more importantly, the authors could predict with 95 percent accuracy whether people with BPD would engage in later self-harm based on earlier emotion and rumination ratings.
Imagine now how useful this technology could be for prevention. As the authors state, the experience sampling method could potentially allow clinicians “to use real-time data provided via [experience sampling] methodology to predict dysregulated behaviors among those with BPD.” From a theoretical point of view, the findings also support the proposal that the cascading effects of negative experiences through rumination through self-harm could help explain in part the difficulties that people with this disorder have in navigating through their everyday lives.
To sum up, technology can’t solve all the problems that people with personality disorders have in managing their symptoms and preventing harmful outcomes. However, this novel approach to studying people with these disorders may help provide insights into how their lives, and those of the people who try to help them, can short-circuit those negative emotional cascades.
Facebook image: Behind-the-Lens/Shutterstock
LinkedIn image: pathdoc/Shutterstock
Selby, E. A., Kondratyuk, S., Lindqvist, J., Fehling, K., & Kranzler, A. (2020). Temporal Bayesian Network modeling approach to evaluating the emotional cascade model of borderline personality disorder. Personality Disorders: Theory, Research, and Treatment. doi-org.silk.library.umass.edu/10.1037/per0000398.supp (Supplemental)