Can negative emotions affect learning and decision-making, leading to suicidal behaviors in at-risk individuals?

painting of face of woman with muted colors

People diagnosed with borderline personality disorder (BPD) can exhibit problem behaviors that result from impulsivity. PhD Student Elinor Waite and Associate Professor Katherine Dixon-Gordon of the Clinical Affective Science Lab (CASL) have theorized that during these impulsive moments, an individual may not have learned from the negative consequences of their past actions. The researchers wanted to find out if emotions affect learning in a way that makes processing potential consequences more difficult.

Their study, published in the Journal of Psychiatric Research, focused on people at-risk of engaging in suicidal behaviors, testing whether their emotional state could ultimately impede decision-making. Increased suicide risk is commonly associated with dysfunctional emotional responses to life events and problems with decision-making, conditions also common in individuals with BPD. Previous research has also shown that suicidal individuals have a reduced capacity to learn from reward and punishment.

The research team first recruited individuals 18–55 years of age that possessed high BPD features in their personality. Participants completed initial interviews as well as questionnaires to collect demographics and a history of their psychopathology.

In the lab, participants first underwent a neutral mood induction, by performing a simple counting task with onscreen visuals. Next, they began a training phase to prep for the upcoming learning task. This involved viewing two Japanese characters onscreen, choosing one of them, and receiving probabilistic feedback (‘correct’ or ‘incorrect’ text, or a more socially based happy or angry face image). Six character pairings were used in total, with certain individual characters deemed correct more often. The team wanted to see how a participant’s learning updated after receiving these responses.

The learning task was similar to the training, but they could be presented with any combination of the six stimuli and received no feedback. The team now wanted to uncover what the individuals learned from the training phase and if they were able to generalize the implied probabilities with new pairings.

Participants also repeated the learning task after undergoing a negative mood induction. At an earlier visit, the individual was interviewed about a recent interpersonal stressor. Details of the stressor like emotions, body sensations, and thoughts were documented, and a script was created—retelling the events that took place. When the participant returned to complete the next learning task, a recording of the script was played, prompting them to recall the experience.  

With the help of Professor Andrew Cohen, the team used computational modeling to identify how the subjects were making decisions during the task, and if they had learned from the training phase. Waite describes this evaluation, “Are their choices based on feedback? Or are they just exploring, because a big part of learning is exploring our environment and seeing what gives us punishment, what gives us reinforcement. How much are they just exploring and trying things versus making their decisions based on what they perceived as feedback?”

The researchers concluded that there were differences in learning rates. Dixon-Gordon relates, “After experiencing this negative stress, people with higher suicide risk were more likely to disregard this pattern of past learning in favor of just flip-flopping around, depending on what worked just recently.

“We can think about this clinically that in those moments of suicidal crisis, when people are in their feelings and feeling hopeless, we really don't want them to disregard past learning. We don't want them to disregard times when they overcome challenges in the past. We want them to be able to hold on to that.”

A larger line of research for CASL and their collaborators, funded by NIH, is looking at individuals who have recently been in the emergency room for suicidal crisis, and investigating whether temporary disruptions in their ability to learn interact with their emotional state to predict near-term suicidal risk behaviors. These acute shifts in a person’s learning ability could represent a period of heightened risk for suicide—a time when the individual should be more closely monitored and receive additional help.

This research will shed light on the emotional and cognitive processes involved with suicidal thoughts and actions, leading to earlier detection of these behaviors and more informed interventions for people at risk.