Because of its circular nature, it is challenging to pinpoint what drives obsessive-compulsive disorder (OCD). Traditionally, obsessive thoughts and compulsions have been framed as downstream, natural consequences of anxiety. After all, patients engage in compulsive actions or thought patterns to avoid specific outcomes that they deeply fear, and an overactive amygdala – the key brain locus of anxiety – is seen across OCD in all symptom dimensions.

Research has uncovered a significant lifetime co-occurrence of OCD with other anxiety disorders: 22% with specific phobia, 18% with social anxiety disorder (also known as social phobia), 12% with panic disorder (PD), and 30% with general anxiety disorder (GAD). Therefore, anxiety is clearly a core part of OCD, as evident from both the neurobiological correlates as well as its psychological symptoms.

However, a shift in research in the past twenty years now places OCD in the family of obsessive-compulsive disorders (OCRDs) in the DSM-5, also including hair pulling, binge eating disorder, and hoarding. While being very heterogeneous, they are all characterised by high trait obsessionality and the urge to perform distressing and time-consuming compulsions.

Stepping away from viewing OCD as an anxiety disorder runs the risk of minimising the established role that fear plays in its onset and maintenance. Regardless of the theme (e.g., contamination OCD, harm OCD, relationship OCD), OCD involves core fears, which are often acquired through trauma; one study involving 265 patients found that 54% had experienced at least one traumatic life event, and 10% dealt with clinically diagnosed PTSD.

But, this new frame offers one key advantage: it encourages research into the cognitive features, or ‘building blocks’, that make up the obsessionality inherent to OCD, which may be what allows patients’ anxiety (trauma-induced or not) to escalate towards compulsions in the first place. Elucidating these cognitive features may help us develop more convincing models of OCD that are explanatory rather than just descriptive, leading to more precise and effective treatments.

Cognitive feature 1: cognitive inflexibility

At the core of OCD lies a deficit in cognitive flexibility, referring to the brain's ability to transition between multiple concepts or perspectives. This is intuitive; regardless of how someone’s OCD manifests, whether it pertains to contamination, identity, incrimination, responsibility over others, or personal safety, they find themselves unable to shift their mind away from their triggers and ‘worst-case scenario’ outcomes. Note: Such inflexible thinking may not necessarily be diametrically opposed to what we consider ‘creativity’. While there is limited research on the topic, the content-rich fears observed in OCD could indicate a high level of trait creativity rather than its absence.

Some could argue that the cognitive inflexibility seen in OCD is a natural side effect of anxiety. After all, the brain’s fear circuitry, containing the amygdala, is over-sensitised in OCD, measurable by fMRI and PET imaging—and we have evolved to fixate on things that the amygdala registers as perceived threats. The amygdala sits within the salience network, assigning emotional significance to external and interoceptive (internal) stimuli and communicating this to the frontal regions involved in high-level cognitive control and attentional processes.

However, a growing body of evidence frames cognitive inflexibility as a potential OCD endophenotype: a measurable characteristic that is inherent to predisposed people, potentially preceding symptom onset (and eventually coalescing with other factors to trigger symptoms).

Indeed, this cognitive deficit is evident in OCD patients’ performance on cognitive tasks. Individuals find it difficult to switch their attention between sets of cards, as measured by the classical Wisconsin card test, and score poorly in a newer task called the intradimensional-extradimensional (IDED) task.

These tasks measure 'attentional set-shifting', the ability to shift focus from one set of stimuli to another—e.g., from pink circles to black lines—in response to changing demands. OCD patients show deficits with a medium-large effect size independent of age and IQ, and this is unlikely to be attributable to anxiety. Decreased scores in the IDED task have been shown in OCD patients with no comorbid anxiety (or other) disorders, and one study found that young people with anxiety disorders did not show a decrease in performance on the IED task, instead outperforming their peers who presented with regular levels of anxiety.

In light of these findings, cognitive inflexibility seems vital to the OCD phenotype rather than merely reflexive, fear-related hesitation. Neuroimaging studies on OCD support this notion, pointing to altered activity within brain regions central to cognitive flexibility, such as the orbitofrontal cortex, caudate, and putamen—all of which are part of the cortico-thalamo-cortico-striatal (CSTC) loop, the main OCD-relevant brain circuitry.

Cognitive feature 2: a tendency to form habits

According to the 'habit theory' of OCD, the condition may stem from an issue within the brain's system responsible for habitual and intentional behaviours. We typically consider habits automatic, repetitive actions performed under various conditions. OCD compulsions could be likened to habits, since they come at the cost of more productive, goal-oriented thinking. Recent research suggests that OCD involves deficits in the goal-oriented system, allowing the habitual system to be overactive, based on studies assessing experiential learning in patients.

One key brain area involved in habit formation is the caudate nucleus, due to its strong connections to the dorsolateral prefrontal cortex. One study reported more caudate nucleus activation in OCD patients during habit learning compared to healthy controls. And within the OCD group, those who had relatively higher caudate engagement reported a heightened urge to perform these behaviours, a contrast not seen in the control group. These findings support the idea that OCD may involve a failure of the cortex to initiate and control goal-directed actions, allowing compulsive, hard-to-break habits to dominate in the form of compulsions.

Cognitive feature 3: low confidence

Another cognitive feature inherent to OCD is decreased confidence. Often called the 'doubting disorder', OCD is characterised by a pervasive sense of uncertainty regarding feared outcomes, which drives compulsive behaviours aimed at seeking reassurance or achieving certainty.

Confidence is the estimated probability that a decision is correct based on the available evidence and can readily be self-reported by patients. It is assessed through behavioural tasks that involve decision-making scenarios (free or forced choices), challenges, or competitive games involving rewards. Patients are tasked with making decisions and subsequently rating their confidence in those decisions. The resulting disparity, or lack thereof, between the accuracy of their decisions and their confidence level, provides insights into whether they are a) overconfident, b) accurate, or c) underconfident.

OCD patients exhibit significantly low confidence across a broad spectrum of cognitive domains far beyond their fears, including memory, perception, and action.

In OCD, low confidence manifests in negative beliefs

In patients, this diminished confidence is observable in overvalued pessimistic beliefs, relating not only to their specific OCD themes and fears but also to their perception of self and others. For example, someone with contamination OCD may have poor confidence in:

  • Their surroundings or the world itself: They believe they are likely to become seriously ill if they do not perform their compulsive cleaning rituals.

  • Themselves: They have difficulty trusting their own judgments and perceptions, leading to a lack of confidence in their ability to accurately assess situations or make decisions. This can be seen as a bias in metacognition, which describes our ability to think about our own capabilities and thought patterns.

  • Others: Whether knowingly or not, they will also possess low confidence in other people’s ability/willingness to assist them in the very slight chance they did become contaminated by something deleterious. They lack the subconscious ‘safety net’ social beliefs that most people hold, doubting that people will intervene effectively if required.

Collectively, low-confidence beliefs fuel compulsive behaviours in an attempt to alleviate doubt and uncertainty. They also harm decision-making, with patients often exhibiting a heightened threshold for making decisions, leading to deliberation over seemingly trivial choices.

Can therapy target and treat the cognitive features of OCD?

So, we can consider a propensity for cognitive inflexibility, a tendency to form habits, and low confidence inherent to the OCD phenotype. Under this model, the following are symptoms of OCD, rather than causal drivers:

  • Feelings of uncertainty.
  • Obsessions/intrusive thoughts.
  • Attempts to self-soothe with compulsions.
  • Poor decision-making.

OCD is a persistent and relentless disorder that can severely disrupt an individual's life. About 25% of OCD cases start by age 14, and OCD is often chronic, with roughly 50% of individuals experiencing symptoms for over 40 years. Gearing treatments towards treating the causes of the OCD cycle, rather than the symptoms that describe it, might prove ground-breaking. But can the cognitive features that drive OCD be targeted?

To date, no existing studies have investigated the possibility of OCD-targeted therapy addressing the first two cognitive aspects: cognitive inflexibility and a tendency for habits. However, low confidence can be treated, potentially altering cognitive rigidity and habit formation through beneficial downstream brain adaptations.

How do we know this?

  • Interestingly, deep brain stimulation applied to the ventral anterior limb of the internal capsule (vALIC) increases self-reported self-confidence in treatment-resistant individuals.
  • Evidence shows that decreases in dysfunctional beliefs correlate with clinical improvements after therapy.
  • A study by Radomsky et al. (2020) found that therapy both a) decreased patient's excessive feelings of responsibility and b) increased their confidence in their own memory, and that both of these factors predicted treatment outcome.
  • Solem et al. (2009) reported improvements in metacognition (including confidence) during ERP therapy to predict the degree of symptom reduction.

How can confidence be improved in therapy? By helping patients prove to themselves that their negative beliefs aren’t helpful, reasonable, or relevant predictors of events.

Exposure-response prevention (ERP) improves confidence by updating beliefs

Beliefs are cognitive structures, stored in complex, interconnected neural networks that sprawl across the brain, but the prefrontal cortex plays a significant role in forming, maintaining, and updating them. The most effective way to alter negative beliefs is to provide the brain with evidence that they are not helpful representations to hold onto. And, this is done by synthesising moments where corrective learning can occur—moments during which the brain’s primitive emotional circuits fire to induce anxiety, but the cortex informs them that there is no real risk—and starts to inhibit them.

Consequently, the brain simultaneously changes its mental representations within the cortex, giving rise to new, more positive top-down beliefs. These beliefs serve as predictions about the world and integrate new understandings of the existing triggers, associating them with more positive possibilities (affordances) instead of just negative ones.

Importantly, this must be done in the language the emotional brain circuitry speaks; patients may feel tempted to use coercion or logical self-talk, but it won’t achieve the desired goal. Patients need to run experiments to prove their brain’s specific feared hypotheses through personalised and precise ERP that a) exposes them to their specific fears while b) providing their brain with sufficient, benign contextual and sensory information.

Before they can do this, they need to precisely identify their core underlying fears – since these reveal their feared hypotheses about what ‘might’ go wrong if they don’t perform compulsions. Core fears take the form of: “If I don’t do X compulsion, Y will happen and I’ll be subjected to Z emotional state”. These are obvious in themes that mainly involve mental compulsions like thought suppression/distraction, such as Harm OCD, Psychosis OCD, and Sensorimotor OCD, but are also present in themes that appear less ‘cognitively complex’, like Just Right OCD (which compels sufferers to perform specific actions until they get a comforting ‘just right’ feeling).

ERP: not just indiscriminate ‘exposure therapy’

The goal of ERP is to 'demonstrate' to the brain that feared stimuli, thoughts, and outcomes are not real threats. Joseph Wolpe, a South African psychiatrist, laid the groundwork for ERP in the 1950s through his work on systematic desensitisation, which involves gradual exposure to feared stimuli The fundamental idea revolves around 'habituation', the natural process by which we acclimate to stimuli or distractions in our environment when they are presented to us repeatedly.

During ERP, patients are prompted to envision their fears or engage in actions that elicit dread, all the while consciously avoiding mental and physical compulsions. Building on Wolpe's work, British psychiatrist Vic Meyer further developed the technique in the 1960s. Edna Foa, an Israeli psychologist, further refined ERP and gave it its name in the 70s and 80s, leading to its establishment as an effective first-line treatment.

However, terms like 'habituation' can lead to the false notion that ERP is just about exposing patients to generic stimuli related to their fears while they sit and tolerate discomfort. While adapting to harmless background noise happens naturally, fully desensitising to an OCD trigger—to the extent it no longer induces fear via the amygdala’s output pathways—requires more than passive acceptance of fear and uncertainty. The brain clings instinctively to associations borne out of anxiety, seeking to keep us secure from perceived threats.

It is only when the brain is presented with sufficient evidence dispelling the belief in the existence of an actual threat that it then becomes open to reconfiguring its protective neural associations, a process called inhibitory learning or 'fear extinction'. It begins to develop new, top-down GABAergic inhibitory connections from the medial prefrontal cortex to suppress the fear response in the amygdala, where fear-related associations are stored. This is the closest the human brain gets to the concept of ‘forgetting’ or 'unlearning' fears.

Disproving scary hypotheses: the goal of ERP is to provide the brain with ‘safety’ evidence

What does the brain consider ‘evidence’ that something isn’t a real risk? When a patient believes something terrible will happen and puts this hypothesis to the test. When the negative, worst-case scenario outcome inevitably doesn’t happen, this generates prediction errors and prompts the brain to rewire itself. A helpful supplementary tool to facilitate this is a technique called paradoxical intention, coined by Victor Frankl; this involves trying to make your worst fears come true by running targeted, time-limited experiments.

Viktor Frankl described paradoxical intention as follows: “The patient is encouraged to do, or to wish to happen, the very things he fears (the former applying to the phobic patient, the latter to the obsessive-compulsive).”

In ERP, mental and real-life experiments should be run, since OCD involves many layers, with individuals fearing real-life triggers but also their own thoughts and what they mean. For instance, an individual who has contamination OCD may start practising ERP by:

  • Initially thinking about their worst-case scenario outcome for five minutes, imagining it in great detail.
  • Then going to the supermarket without washing their hands afterwards, while accepting potential contamination.

Paradoxical intention aligns neatly with ERP and could be seen as a more targeted form of it. Patients try to make their fears come true in order to demonstrate to their brain that they cannot occur, expediting ERP’s goal: to encourage the re-evaluation of associated triggers and thoughts as inconsequential and non-threatening.

Why ERP doesn’t always yield satisfactory results

Overall, it's estimated that about 60% of OCD patients show a significant reduction in symptoms after undergoing ERP therapy. Excluding discontinuation, there are three main reasons why some patients fail to see improvements with ERP:

  • Practitioners often fail to help patients identify their true underlying core fears, (which are always related to the emotional state, not just the outcome, that they fear; OCD involves the fear of something specific happening and feeling a certain way as a consequence).
  • It takes the form of overly generic exposure rather than pinpointing which specific hypotheses to test (i.e., the core fears they believe will come true if they don't perform their compulsion) and designing appropriate experiments. When deep brain stimulation works for OCD, it most likely does so by relaxing the weight of top-down priors, or expectations, concerning negative ideas and outcomes. This can be achieved under normal circumstances by testing irrational fears, as this generates prediction errors when they don’t happen and prompts the brain to modify its priors naturally.
  • Patients do not engage in it fully, or do, but only for short bursts of time. ERP demands a certain boldness, as patients have to be willing to experience their least favourite anxiety symptoms. Fear memories in the amygdala can only be overwritten by the medial prefrontal cortex when both regions are activated31, and feared hypotheses can only be disproved if an experiment accurately representing the fears is set up. ERP is always temporarily uncomfortable, underscoring the importance of aligning its practice with neuroscience principles to enhance its efficacy and efficiency.

Could psychedelic therapy streamline the ERP journey?

While ERP is the most robust treatment we have for OCD, some patients find tools helpful to ease them through the process. Selective serotonin-reuptake inhibitors (SSRIs) are the primary medications recommended for OCD. However, they typically provide only modest symptom improvement, and 40-60% of individuals don’t respond. Emerging research suggests a novel approach to treating OCD through psilocybin therapy, with a small 2006 proof-of-concept study demonstrating low-moderate doses to be both a safe and effective treatment for OCD patients.

Psychedelics chiefly bind to serotonin 5HT2A receptors in the cortex, where they facilitate the relaxation of rigid belief systems and prompt the brain to reassess its existing new mental constructs. This occurs through the temporary weakening of established top-down priors, allowing the formation of new priors (and beliefs) concerning alternative possibilities. As a result, a patient’s thinking around their OCD theme (and beyond) may become more lateral and positive, as long as they are adequately supported.

Psilocybin-assisted therapy could be a valuable tool to accelerate the ERP healing journey. By making beliefs temporarily more malleable, it might enhance patients’ confidence, while also improving cognitive flexibility, and potentially promoting the predominance of goal-oriented thinking over habitual compulsions by its known modulation of the prefrontal cortex and other areas within the CSTC circuitry. These effects could considerably improve patients’ tolerance for ERP and its efficacy, allowing them to rise above their previously entrenched thought patterns with greater ease.