Cognitive behavioral therapy (CBT)

What CBT is

CBT is a short-to-medium-term, goal-focused talk therapy. It grew out of two older traditions, behavior therapy and cognitive therapy, which merged over the second half of the twentieth century into the approach used today. The behavioral side came first, built on the observation that what we do, especially what we avoid, keeps problems alive. The cognitive side, developed by the psychiatrist Aaron Beck and others, added the insight that the way we interpret events, not just the events themselves, drives much of how we feel. Put the two together and you get CBT: a therapy that works on thoughts and behavior at the same time, because in real life they're never separate.

A few features set CBT apart from open-ended talk therapy. It's structured. Sessions have an agenda, and the overall course has a direction. It's collaborative. The therapist isn't a silent listener or an authority handing down interpretations. They work alongside you, more like a coach than an oracle, and a good CBT therapist will say so out loud. It's present-focused. CBT pays attention to the past, but mainly to understand how older experiences show up in today's patterns, not to excavate the past for its own sake. And it's skills-based. The point of CBT is to leave you with tools you can keep using once therapy is over. People sometimes describe a good course of CBT as becoming your own therapist.

It's also worth saying what CBT is not, briefly, because the name gets stretched. CBT is a specific, manualized family of approaches with a defined structure, not a loose label for any practical-sounding therapy. When it's done well, it follows a recognizable shape. When a clinician says they "use CBT techniques," that can mean a lot of things, and it's fair to ask what they actually mean.

The core idea

CBT rests on one model, and it's simple enough to draw on a napkin. A situation triggers a thought. The thought triggers an emotion and a set of physical sensations. The emotion and sensations push toward a behavior. And the behavior loops back, feeding the next situation and the next thought. Thoughts, feelings, and behavior aren't a chain so much as a circle, each one influencing the others.

The part that surprises people is the role of the thought. Most of us assume situations cause feelings directly. Something bad happens, so we feel bad. But two people can sit through the exact same event and feel completely different things, because they interpreted it differently. A short reply from a boss reads as "she's busy" to one person and "she's angry at me" to another. The email is identical. The feeling isn't. What sits between the event and the emotion is an interpretation, and that interpretation is often fast, automatic, and unexamined. CBT calls these automatic thoughts. They flash through so quickly that we usually notice only the feeling they leave behind.

Automatic thoughts aren't always wrong. Often they're accurate and useful. The trouble starts when they're distorted, and distorted thinking tends to run in predictable grooves. Catastrophizing jumps to the worst case. Mind-reading assumes you know what someone thinks. All-or-nothing thinking sorts everything into total success or total failure. Overgeneralization turns one bad event into a never-ending pattern. These are the cognitive distortions, and once you can name them, they get easier to spot.

Behavior matters just as much, and here the key concept is the maintenance cycle. Anxiety is the clearest example. A worried thought makes you anxious, so you avoid the thing you're worried about, or you do something to feel safer. Avoidance brings instant relief, which feels like the right call. But it also robs you of the chance to learn that the feared thing might not have happened, or might have been survivable. So the fear stays intact, ready to fire next time. The behavior that's meant to protect you is quietly keeping the problem going. Depression has its own version. Low mood pulls you to do less, doing less removes the small sources of reward and accomplishment that lift mood, and the lower mood pulls you to do even less. CBT works by getting inside these loops and changing them at a point where change is actually possible, which usually means the thought or the behavior, because those are the parts you can get hold of.

For what changes in the brain over a course of therapy, see how therapy changes the brain on shrinkMD.

What a course of CBT looks like

A course of CBT has a shape, and knowing it ahead of time makes the experience less mysterious.

It starts with assessment. In the first session or two, the therapist gets a clear picture of the problem: what brings you in, when it started, what makes it better or worse, and how it affects daily life. Out of that comes a formulation, which is a working map of how your particular problem is being maintained. The formulation isn't a diagnosis. It's a hypothesis about your loops, and it can be revised as you learn more.

Then you set goals together. CBT works toward specific, concrete targets rather than a vague sense of feeling better. "Be able to give the team update without rehearsing it twenty times" is a CBT goal. "Have more confidence" is harder to work with until it's broken into something observable. Good goals give both of you a way to tell whether therapy is working.

The middle of the course is the working phase, and it follows a steady rhythm. Most sessions have a similar structure: a brief check-in and mood rating, a look at the homework from last time, an agenda for the day, the main work, and a plan for the week. The agenda is set together at the start of each session, which keeps the time focused. This structure can feel businesslike at first. It's deliberate. It makes sure the hour goes toward the problem rather than drifting.

Homework, often called between-session practice, isn't an optional extra in CBT. It's where most of the change happens. The therapy hour is for learning a skill. The other hundred-odd waking hours in the week are for using it. Research consistently finds that people who do the between-session work tend to do better. The tasks are meant to be specific and doable: track a few thoughts, try one experiment, take one avoided step.

CBT is time-limited by design. A typical course runs roughly 8 to 20 sessions, usually weekly to start, though more complex problems such as PTSD or longstanding difficulties can take longer. As things improve, sessions often spread out, every other week, then monthly, which gives you a chance to run things more independently while still having support. Near the end, good CBT spends real time on relapse prevention: naming early warning signs, planning for setbacks, and writing down what worked so you can use it again. The aim is for the ending to be planned, not abrupt, and for you to leave with a usable toolkit rather than a dependence on the room.

What a session actually looks like

The abstract description only goes so far. It's easier to see CBT working through a concrete example, so here's one you can follow start to finish.

Imagine someone, call her Maya. At work, she sends a report with a real mistake in it, a wrong figure in a table. Her manager replies: "Can you take another look at this?" Nothing more. From there, a familiar sequence kicks off. The automatic thought arrives almost instantly: "I'm going to get fired." The emotion is sharp anxiety, and her body joins in, a tight chest and a buzzing, restless feeling. The behavior follows. She fixes the figure, then keeps refreshing her email for a follow-up. She rereads the manager's message a dozen times, hunting for hidden anger in the wording. She drafts a long, over-apologetic reply, deletes it, drafts another. She gets very little else done that afternoon. By evening she's exhausted and convinced her job is hanging by a thread.

Here's how a CBT session would actually work with that.

The first move is noticing the thought. Maya and her therapist slow the moment down and pull it apart, often on paper, sometimes using a worksheet called a thought record. They separate the strands she's been experiencing as one blur. The situation: a mistake in a report and a brief reply from the manager. The automatic thought: "I'm going to get fired." The emotion: anxiety, which Maya rates around 85 out of 100. The behavior: checking email, rereading the message, drafting and deleting. Just naming the thought as a thought, rather than a fact, already changes something. She's used to living inside "I'm going to get fired." Now she's looking at it.

The second move is examining the evidence. The therapist asks Maya to step into the role of someone weighing a claim rather than someone braced for a verdict. What's the evidence that supports the thought? She has made a mistake, and that's real. What's the evidence against it? She's been in the role two years with solid reviews. The manager's message was neutral, not angry, and "take another look" is a normal request, not a threat. People at her workplace make mistakes regularly and aren't fired for them. She's never seen anyone fired for a single error in a table. Laid out side by side, the evidence is lopsided. The catastrophic thought has one fact behind it and a stack of facts against it.

They also name the pattern. "I'm going to get fired" is catastrophizing, jumping to the worst possible outcome, and it's fueled by mind-reading, assuming she knows the manager is angry from a message that doesn't say so. Naming the distortion isn't about scolding the thought. It's about recognizing the groove, because it's a groove Maya's mind has run many times before.

The third move is testing the thought, and this is where CBT goes past discussion. A balanced thought is useful, but a thought tested against reality is stronger. Maya and her therapist design a small behavioral experiment. Her prediction: if she sends a brief, normal reply and doesn't monitor her email, the manager will be cold or angry, and the anxiety will stay at 85 all day. The experiment: send one short message, "Thanks for catching that, I've corrected the figure and reattached it," then deliberately not check email for two hours. They agree in advance on what would count as the feared outcome and what would count as evidence against it.

The fourth move happens the following week, when they review what actually occurred. Maya sent the reply. The manager wrote back "Great, thanks." Her anxiety, which had been at 85, dropped into the 30s within the afternoon, faster than she expected once she stopped feeding it with checking. The prediction failed, and not in a vague way. She has a specific, concrete instance where the catastrophic thought was wrong and the calmer reading was right. That's the changed response. Over a course of CBT, Maya does this enough times, with enough different situations, that the automatic thought loses its grip. She still has the thought sometimes. She just no longer treats it as the truth, and she no longer organizes an afternoon around it.

That sequence, notice the thought, examine the evidence, test it, change the response, is the engine of CBT. The content changes from person to person. The mechanics stay the same.

The main techniques

CBT isn't one technique. It's a toolkit, and a therapist draws on different tools depending on the problem. A few are central to most courses.

Cognitive restructuring. This is the work done in Maya's example: catching automatic thoughts, examining them against the evidence, and arriving at a more accurate and balanced view. The everyday tool is the thought record, a simple grid for writing down the situation, the thought, the feeling, the evidence on each side, and a revised thought. It feels artificial at first, the way any new skill does. With practice, the steps become quicker and eventually internal, something you can run in your head in a few seconds. The goal isn't a cheerful thought. It's a true one.

Behavioral experiments. Thinking through evidence is useful, but the most convincing evidence is the kind you collect yourself. A behavioral experiment turns a belief into a testable prediction and then runs it in the real world. "If I admit I don't know the answer in a meeting, people will think less of me" becomes an experiment: admit it once, on purpose, and watch what actually happens. Experiments often teach faster than discussion, because lived experience tends to outweigh a logical argument when the two disagree.

Exposure. For anxiety problems, fears, panic, social anxiety, OCD, PTSD, exposure is often the most powerful component. The principle is to approach the feared thing, gradually and deliberately, instead of avoiding it, and to stay with it long enough to learn something new. You usually build a ladder, from mildly uncomfortable steps to harder ones, and work up it at a manageable pace. Two things change. The fear tends to settle the longer you stay, and, more importantly, you learn that the feared outcome either doesn't come or is more bearable than predicted. In OCD, the specific form is exposure and response prevention, where you face the trigger and resist the compulsion that usually follows. Exposure is uncomfortable on purpose, and a good therapist makes it collaborative and paced, never a matter of being thrown in the deep end.

Behavioral activation. This is the core behavioral tool for depression. Depression pulls people to withdraw and do less, which removes the small experiences of pleasure and accomplishment that feed mood, which deepens the depression. Behavioral activation works the cycle in reverse. Rather than waiting to feel motivated before acting, you schedule activity first, in small, planned steps, and let mood follow action rather than lead it. It sounds almost too plain to be a treatment. It has strong evidence behind it, and on its own it can be as effective as fuller CBT for depression.

Between-session practice. Homework isn't a separate technique so much as the thread running through all of them. Thought records, experiments, exposure steps, scheduled activities, these are practiced between sessions, in the situations where the problem actually lives. CBT works because you use it, not because you discuss it. The people who get the most from CBT are usually the ones who do the practice, even imperfectly.

A few other elements show up often: psychoeducation, simply understanding how anxiety or low mood works, which itself reduces fear; relaxation and breathing skills, useful for the physical side; problem-solving for real-world difficulties that genuinely need solving; and worry-postponement techniques for people whose minds run all day. Which tools get used, and in what mix, depends on the person and the problem.

What CBT is used for and what the evidence shows

CBT has the largest and strongest evidence base of any psychotherapy, built from hundreds of randomized controlled trials and many meta-analyses, including Cochrane reviews. Major clinical guidelines, including those from the UK's National Institute for Health and Care Excellence and the American Psychological Association, recommend it as a first-line treatment for a range of conditions.

Anxiety disorders. This is CBT's strongest ground. For generalized anxiety disorder, panic disorder, social anxiety disorder, and specific phobias, CBT is a first-line treatment with consistent benefit across trials. For phobias and panic in particular, the exposure-based components can produce substantial change, sometimes quickly.

Depression. CBT is one of the best-supported psychotherapies for major depressive disorder, recommended as a first-line option for mild to moderate depression and used alongside medication for more severe depression. A useful and well-replicated finding is that CBT appears to lower the risk of relapse after treatment ends, more so than medication stopped at the same point, which is generally attributed to the skills staying with the person.

OCD. The specific form of CBT for obsessive-compulsive disorder, exposure and response prevention, is the first-line psychotherapy for OCD and has strong evidence. General talk therapy tends not to help OCD much. ERP specifically does.

PTSD. Trauma-focused CBT, including approaches such as cognitive processing therapy and prolonged exposure, is a first-line treatment for post-traumatic stress disorder, recommended in major guidelines.

Beyond those. CBT also has good evidence in insomnia, where CBT for insomnia, known as CBT-I, is the recommended first-line treatment ahead of sleeping medication. It's used, often in adapted forms, for eating disorders, chronic pain, substance use problems, and as an adjunct in the management of conditions such as bipolar disorder and psychosis, where it doesn't replace medication but can help. It's effective across formats too: individual, group, and increasingly through well-designed online and guided self-help programs, which matters because they can extend access where in-person therapy is hard to reach.

One honest qualifier runs through all of this. "Effective" means it helps a large share of people meaningfully, on average, in trials. It doesn't mean it works for everyone, or that it's the right first choice for every person. The evidence is strong, and it still leaves room for individual judgment.

What CBT is not

A few misconceptions follow CBT around, and clearing them up helps people know what they're actually getting.

CBT isn't positive thinking. This is the most common confusion, and it matters. CBT doesn't ask you to swap a dark thought for a cheerful one, or to repeat affirmations until you believe them. The work is closer to fact-checking than to cheerleading. You take a thought and ask whether it's accurate and whether it's useful, and you let the evidence decide. Sometimes the honest answer is reassuring. Sometimes it isn't. The thought "I'm going to get fired" is usually distorted. The thought "I made a real mistake and should fix it carefully" might be entirely true. CBT lands you on the accurate thought, which is often more realistic rather than more upbeat, and a realistic thought is something you can actually act on.

CBT isn't telling yourself everything is fine. If something genuinely isn't fine, CBT doesn't paper over it. A person facing a real loss, a real diagnosis, a real injustice, isn't helped by being told to reframe it away. In those situations CBT shifts toward problem-solving the parts that can be changed, and toward a different and harder kind of work: accepting the parts that can't, without piling distorted thoughts on top of an already difficult reality. Pretending isn't the goal. Accuracy is.

CBT isn't a way of dismissing feelings. Because CBT works on thoughts and behavior, people sometimes assume it treats emotions as a nuisance to be argued away. It doesn't. Emotions are taken seriously in CBT, as important information about what matters to you and as the very thing the therapy is trying to ease. The reason CBT works through thoughts and behavior is practical, not dismissive. Those are the parts of the loop you can get hold of directly. You can't simply decide to feel less anxious. You can change how you respond to an anxious thought, and the feeling tends to follow.

CBT isn't a lecture, and it's not the therapist telling you what to think. Done well, it's collaborative. The therapist brings a method and asks good questions. You bring the expertise on your own life. The conclusions are meant to be ones you reach and test yourself, which is exactly why they tend to stick.

CBT isn't a quick fix. It's often time-limited, but time-limited isn't the same as effortless. The skills work because you practice them, in the weeks between sessions and after therapy ends. Hearing about a thought record changes nothing. Using one, repeatedly, changes a lot.

How it compares and where it fits

A reasonable question is how CBT stacks up against the alternatives, and the honest answer is that it depends on the problem and the person.

CBT versus medication. For many common conditions, including depression and the anxiety disorders, CBT and medication produce broadly comparable benefit on average. They aren't competitors so much as different routes, and they have different profiles. Medication can work faster for some people and asks less active effort. CBT asks for real work and time, and in return it tends to leave skills behind, which is the likely reason it lowers relapse risk after treatment stops. Neither is universally better. The right choice depends on severity, preference, what's available, and what has or hasn't helped before. They also combine well. For more severe depression, for example, medication plus CBT is a common and well-supported plan, with medication bringing symptoms down enough that the therapy work becomes possible. CBT doesn't require stopping medication, and starting CBT isn't a verdict that medication failed.

CBT versus other therapies. CBT is the most studied psychotherapy, but it isn't the only one that works. Interpersonal therapy has good evidence for depression. Psychodynamic therapy helps many people and suits those who want to work in a more open-ended, exploratory way. Newer approaches that grew out of the CBT tradition, sometimes called third-wave, including acceptance and commitment therapy and dialectical behavior therapy, have their own evidence and their own strengths. DBT in particular is the leading treatment for chronic suicidality and borderline personality disorder, areas where standard CBT isn't the first choice. The practical takeaway is that CBT is an excellent default for many problems, with the strongest evidence base, and it's still one good option among several. If a course of CBT doesn't help, that's a reason to try something else, possibly a different therapy, possibly a different therapist, not a reason to conclude that therapy can't help.

Two factors deserve as much attention as the brand of therapy. The first is the fit between you and the therapist, the working relationship, which predicts outcome across every type of therapy. The second is whether the approach matches the problem: ERP for OCD, trauma-focused work for PTSD, CBT-I for insomnia. A good clinician will help you match the tool to the task.

How to find CBT and what to look for in a therapist

Finding good CBT takes a little knowing-what-to-ask, and it's worth the effort.

Start with the practical routes. A primary care doctor can often refer you, and is a fine first stop. Your health insurance, if you have it, will usually have a directory of covered therapists. Professional bodies maintain searchable directories, and reputable ones let you filter for CBT specifically. In some countries, including the UK, you can refer yourself to publicly funded talking-therapy services without going through a doctor. University clinics and community mental health centers can be lower-cost options, and many therapists offer a sliding scale.

When you're choosing, a few things matter. Look for someone properly licensed in your region, a psychologist, clinical social worker, licensed counselor, or psychiatrist, who has specific training in CBT rather than a passing familiarity with it. It's entirely reasonable to ask directly: "Are you trained in CBT, and do you use it as a structured approach?" Ask whether they have experience with your particular problem, because CBT for OCD, for PTSD, and for insomnia each involve specific methods. A therapist who treats panic regularly will work differently from a generalist.

You can also tell something from how the work actually runs. Real CBT tends to feel structured. Early sessions involve assessment and goal-setting. Sessions have an agenda. There's between-session practice, and progress gets reviewed. If, several sessions in, the therapy feels entirely open-ended with no plan, no goals, and no practice, it's fair to ask whether what you're getting is CBT, and whether it's the right fit.

And pay attention to the relationship. You should feel reasonably comfortable, taken seriously, and able to be honest. The bond between you and the therapist predicts outcome across every kind of therapy. It's normal for the first session or two to feel a bit awkward. If after a few sessions it still feels wrong, it's completely acceptable to look for someone else. That's not failure. It's matching, and a good therapist will understand.

A note on cost and access. Trained CBT therapists aren't equally available everywhere, and waitlists can be long. If in-person therapy isn't reachable, evidence-based online programs and guided self-help are a legitimate route, not a consolation prize. Group CBT, often more available and lower-cost, is also effective for many conditions.