Doubting what you already know: Obsessive-Compulsive Disorder

Folk wisdom has attached Obsessive-Compulsive Disorder (OCD) to many stereotypes: the clean freak, the person who is a stickler for symmetry and geometric perfection, or the person who is perpetually locked into a series of repetitive routines. 

Clinicians identify compulsive behaviors to make diagnoses for OCD, but it’s still unclear to researchers how these compulsive behaviors actually work. A common preconception about OCD is that it is a display of mental inflexibility, causing perseverations (pathological persistent behavior repetitions). A new study by researchers from the Hebrew University of Jerusalem used computational modeling techniques to uncover a different perspective about obsessive-compulsive (OC) behaviors. They found that OC symptoms can be better understood as a distrust of past experience, which challenges previous understandings of OCD as some deficit of flexible behavior change.

This image has an empty alt attribute; its file name is Handwashing.jpg

The Intuition

OCD affects a person’s ability to learn about the environment. “Environment” can define any unobservable aspect of the world that can have any number of possible states. For example, my environment can be the cleanliness of my hands (clean or dirty) or my friend’s true opinion of me (she likes me or she doesn’t like me anymore). Evidence is anything that helps me reduce my uncertainty and help me make inferences about my environment’s true state. For example, evidence can be ‘I just touched that doorknob’ or ‘she responded to my text message after two hours with just, ‘k’’.

Suppose a person experiences compulsions to constantly wash his hands thinking they are dirty, or constantly need his friend’s reassurance that everything is fine between them. Is this because of rigid belief patterns that are insensitive to feedback? Or rather, is it actually because of excessive uncertainty about the environment’s state that motivates the constant checking behavior? 

The Experiment

The researchers constructed an experiment to observe two main things. They wanted to see how uncertain different participants were about their previous beliefs. They also wanted to see how well the participants could learn from evidence in one environment, then switch gears to locate new evidence once they sensed the change in the environment. In other words, how much time after a good hand-wash in the past does it take for it to lose relevance in the present, and how well does a hand-wash undo the belief of dirtiness from previous surface-touchings?

To measure these behaviors, the researchers gave a spatial cueing task to a varied group of people who had different levels of OC symptoms (see figure). The objective of the task is to correctly predict the spot of a hidden dot on the screen, either left or right of center, in light of seeing some arrow signals on the screen. First, for every trial, participants briefly see a column of 3 arrows that either point left or right. Participants are told that one of the three arrows is reliably predictive of the hidden dot’s current location, while the other two are sham signals. After some trial and error, most people can quickly learn which arrow is the informative one, helping them subsequently locate the dot with perfect accuracy.

Now, the key feature: The experimenters informed participants that at some point in the experiment, an invisible cue shift will occur, where one of the sham arrow cues becomes the new informative one for all the remaining trials.

Knowing to be alert for this cue shift, participants start the experiment, trying to perform the dot-finding task with the highest possible accuracy.

The Spatial Cueing task from the original paper


The researchers compared the behavioral results to several different theory-based simulations to see which theory explained the true behavioral results the best. Winning out for best model was the Bayesian Change-point (BCP) model. OC symptoms correlated with a specific type of uncertainty called ‘transition uncertainty’ in the BCP model. High transition uncertainty means feeling extremely unsure about whether the hand-wash you did two minutes ago really made your hands clean because you technically didn’t go between all the fingers as you missed the thumb and you might have counted 25 seconds instead of the usual 30. It means being unable to trust that the keys are still in your purse after you put them there thirty seconds ago, so you decide that you need to reach inside to check once more.

An external file that holds a picture, illustration, etc.
Object name is pcbi.1007634.g005.jpg
Positive correlation between OC symptoms and transition uncertainty in the probabilistic and deterministic version of the spatial cueing experiment

This result is important because it tells a different story from what researchers previously believed about OC behavior. It’s not that compulsive hand washers have this inflexible belief that their hands are perpetually dirty; handwashing makes hands feel cleaner, regardless of one’s OCD inclinations. Rather, those with high OC tendencies have an intense mistrust of their previous experiences which explains the constant checking and looking for reassurance. 

Bigger Picture

In the bigger picture, insights like this about the inner mechanics of mental illnesses are essential for developing new treatments. If psychiatric researchers are able to understand the specific moving parts responsible for a person’s condition, they’ll have a better idea about which drugs and therapies would fix the problem at its source. Any step that can move mental illness treatment away from today’s prevalent approach of “here’s a cocktail of drugs that kind of works for some people, let’s see what happens” is a step in the right direction.

Source Cited:

Fradkin, I., Ludwig, C., Eldar, E., & Huppert, J. D. (2020). Doubting what you already know: Uncertainty regarding state transitions is associated with obsessive compulsive symptoms. PLoS Computational Biology, 16(2).

1 Comment

Leave a Reply