Exposure Works When Learning Happens

“Face your fear” usually refers to deliberately entering situations that trigger difficult thoughts, emotions, and bodily sensations.

Why would anyone do that—when it feels so bad?

Exposure is an effective treatment with large effect sizes, evidence shows (Olatunji et al., 2010). When exposure fails, it is rarely because exposure itself does not work, but because it is misunderstood—or used to manage anxiety rather than to learn from it. Its effectiveness depends on whether it leads to new learning.

Exposure is structured as an experiment, where we test different kinds of predictions: whether feared events will occur, or whether one’s anxiety and thinking processes are dangerous or uncontrollable.

Across anxiety disorders, exposure on its own is often as effective as exposure combined with cognitive reappraisal, meta-analyses show (Deacon & Abramowitz, 2004; Olatunji et al., 2010).

A main reason exposure works so well on its own is that experience is persuasive.

How do we benefit the most from exposure therapy? Exposure is emotionally demanding, and dropout rates are therefore higher than in many other therapeutic interventions. 

For this reason, effective exposure requires a strong and shared treatment rationale, a carefully designed experiment tailored to the client’s beliefs, and a clear agreement before the work begins.

My hope is that this post can help both therapists and clients gain the benefits of effective exposure.

Cognitive-Behavioral Exposure: Changing Beliefs Through Experience

In cognitive-behavioural therapy (CBT), exposure is used to explore and gradually change catastrophic assumptions about feared situations. One way this is done is through cognitive reappraisal—learning to look at the situation in a more balanced and flexible way. The underlying idea in CBT is that anxiety is maintained by beliefs that overestimate danger.

When a person stays in a feared situation without escaping, avoiding, reassuring themselves, or trying to make the anxiety go away, important learning can take place. Over time, people often realize that:

• “The feared outcome did not happen.”
• “My thoughts were exaggerated or inaccurate.”

This pattern of learning has been demonstrated across various anxiety disorders, such as panic disorder, social anxiety, OCD, and specific phobias (Salkovskis et al., 1996; Whittal et al., 2005; Wells 2009).

Interestingly, changes in beliefs often happen even without directly challenging thoughts. Many people find that their thinking shifts naturally as a result of lived experience rather than effortful mental debate.

Studies show that cognitive change frequently follows behavioural exposure, rather than preceding it (Longmore & Worrell, 2007).

How exposure is done matters. When a lot of effort is put into controlling or reducing anxiety during exposure—through distraction, reassurance, or constant mental adjustment—learning can become weaker. 

Research suggests that these strategies may unintentionally teach that anxiety is only manageable when it is actively controlled (Salkovskis et al., 1999; Telch et al., 2014).

Because the urge to manage anxiety is very human and common, this is something therapists and clients need to keep noticing together. When exposure is allowed to be an experience rather than a coping exercise, it is more likely to support lasting change.

Metacognitive Exposure: Changing Your Relationship to Anxiety

Metacognitive exposure takes a different starting point than the traditional CBT approach. 

In metacognitive therapy (MCT), exposure is used to test and modify beliefs at the metacognitive level, i.e. beliefs about the mind itself—such as the belief that thoughts must be controlled in order to cope in a situation.

From a MCT perspective, anxiety is maintained by self-focused attention and prolonged thinking, known as the Cognitive Attentional Syndrome (CAS):

• The urge to monitor, analyse, reassure, or control
• The impulse to respond mentally to perceived threat with worry

The aim of exposure is to disengage the CAS and, for example, realize that you are not losing control over your mind. More specifically, you will discover that: anxiety regulates itself, flexible focus and action remain possible while anxiety is present.

This aligns with the metacognitive model, which emphasise reducing engagement with anxiety rather than changing its content (Wells, 2009; Nordahl & Wells, 2017). A long-term follow-up study shows that treatment gains from metacognitive therapy are maintained up to 9 years after treatment.

From a metacognitive perspective, the central learning is:

• “Nothing needed to be done.”
• “Anxiety can be present without needing a response.” 
• “I’m safe feeling this way.”

Metaphorically speaking, metacognitive exposure isn’t about controlling the weather. You don’t try to stop the rain or argue with the forecast. You learn that rain can happen without requiring a response. You can keep walking, working, and living while it rains. The weather may change. It may not. What changes is not the weather —it is that weather no longer decides what you do.

Does The Brain Get Used to Fear Stimuli? (The Theory of Habituation)

A widespread misunderstanding is that exposure works because anxiety decreases during the session.

However, multiple studies show that within-session fear reduction does not reliably predict treatment outcome (Craske et al., 2008). 

Instead, outcome is better predicted by between-session change and reductions in avoidance and safety behaviours (Craske et al., 2014).

From a learning-theory perspective, this is expected: exposure does not work by habituating the brain or erasing emotions, but by producing new learning.

Indeed, research findings support this idea:

• Reduced amygdala activation (habituation) is not a reliable mechanism explaining anxiety reduction (Peterman, et al., 2016)
• Greater or sustained amygdala engagement during exposure can be associated with better clinical outcomes (Lor, et al., 2023)
• Changes in amygdala–prefrontal functional connectivity are more consistently linked to improvement than raw fear reduction (Zhu, et al. 2017; Klumpp, et al., 2014)
• Clinical improvement often reflects reduced behavioural control by anxiety rather than reduced emotional intensity: Fear memories remain intact and are inhibited by new learning (Bouton, 2004)

The inhibitory learning model formalises this: exposure works by creating new, competing associations, not by eliminating old ones (Craske et al., 2014; Jacoby & Abramowitz, 2016). 

Metaphorically speaking, exposure does not remove “the fire alarm”: It teaches you that the alarm can go off without there being a fire. The alarm system (anxiety) remains intact and sensitive. What changes is that you no longer: Evacuate the building, call the fire brigade, start checking for smoke. You learn that nothing needs to be done when the alarm sounds. This is learning — not raw desensitisation.

Also, calm is not required for learning. As mentioned earlier, attempting to control anxiety during exposure can actually interfere with new learning (Craske et al., 2008; Telch et al., 2004). 

Summing Up

Taken together, evidence points to the conclusion that although reduced amygdala activation may occur during repeated exposure, it does not explain why people get better.  

Anxiety does not need to fade for change to happen. What matters is a shift in how anxiety is responded to during exposure.

Lasting improvement is not driven by feeling calmer during exposure, but by doing less to avoid, control, or neutralise anxiety. As safety behaviours fall away and behavioural freedom increases, anxiety loses its power. Not because it disappears, but because it no longer governs attention, choices, or action.

For this reason, exposure works best when it is approached as a focused experiment with clear, client-relevant predictions. The question is what the person needs to discover to get better:

“What would it mean for you to discover that X does—or does not—happen? How would that experience change how you respond to anxiety the next time?”

The aim is not to reduce fear in the moment, but to test the assumptions about danger, control, or the need to respond to anxiety at all. 

If this way of understanding exposure raises questions, you are welcome to continue the conversation below in the comments.

Thank you for your time.