Functional Communication Training, Explained

Functional communication training, or FCT, teaches a person a safe way to ask for what they want. It replaces a challenging behavior with a clear request that gets the same result. If a child screams to escape a hard task, we teach them to ask for a break instead. The request works, so the screaming has less reason to happen.

FCT matters because it does two jobs at once. It lowers behavior that hurts or disrupts, and it gives the person a real skill they can use for life. That is why so many clinicians reach for it first. But doing it well takes more than picking a phrase, and the experts below show why.

Why FCT is such a strong intervention#

Many behavior analysts, the pros who study why behavior happens, call FCT the workhorse of the field. Matt Harrington makes that case as plainly as anyone.

FCT is the biggest intervention in behavior analysis. Like, I don't think it's arguably. I think it is. From the talk — Matt Harrington

Harrington also explains the math under the hood. FCT works because the new request pays off almost every time, while the old behavior stops paying off. He calls that gap in "contingency strength," a measure of how reliably a behavior pays off, the real engine. The request wins because it is the sure path to the reinforcer, the reward that keeps a behavior going. The severe behavior loses because it now leads nowhere. Prompts and materials help, but the payoff gap does the heavy lifting.

The old payoff usually has to stop#

Done the classic way, FCT has two parts. You reinforce the new request. You also put the old behavior on extinction, which means it no longer earns its reward. Harrington frames it as reinforcing the mand, or request, at a high rate while the severe behavior gets nothing.

He also points out that extinction is not the only road. In one talk he shows a team that reached the same result without ever placing the child on extinction. They shrank a precursor behavior, a smaller behavior that comes right before the big one, and built up the request instead. The lesson is that the effect matters more than the recipe. A copy-paste protocol from a paper is not the goal.

The response has to make sense to strangers#

A request only helps if people outside the trained team can understand it. Harrington built his whole day-five lesson around this idea.

Functional communication training was really built with social validity and generalization in mind. From the talk — Matt Harrington

His test is simple and worth stealing. Could a substitute teacher honor this request with no notes left behind? If the answer is no, the request is too private to travel. The point of FCT is to let a person enter a brand new room and still ask for what they need. A response only the home team can read does not pass that bar.

A good request starts with a good function assessment#

FCT needs a request that matches the reason behind the behavior. To find that reason, you need a functional behavior assessment, a careful look at why a behavior keeps happening. Harrington says the tie is not optional. An FCT plan needs a function-based response, and that means an accurate assessment comes first.

Nicole Stewart adds a sharp warning through an ADHD lens. Imagine an assessment says a behavior is escape-based, meaning it works to get out of demands. But the real driver is a need for stimulation. Teaching the child to ask for a break will not fix it, because a break does not deliver the missing input. The request has to replace the true need, not the surface story. Dr. Jessica Osos makes a related point. She calls FCT a strong, well-supported choice for escape-based behavior. She also shapes the modality, sign or voice, around the family's wishes.

One rigid phrase is rarely enough#

Insisting on a single exact sentence can backfire. B. Kuerine Gray sees this often with demand-avoidant profiles, learners who push back hard on being told what to do. To them, a stiff script is just another demand. Her advice is to count any honest attempt as an FCR, short for functional communication response, the trained request.

if you're understanding that the intent is an FCR that you count it as an FCR, because if you're saying like, Oh, well, you didn't say this... you're setting another demand instead of engaging in the intent of skills-based treatment From the talk. B. Kuerine Gray

Dr. Shane Spiker points to research that backs a broader request. Teaching an "omnibus mand" like "my way" can hold behavior down better than only teaching narrow, moment-specific requests. In the study he cites, problem behavior stayed low even after teams switched to teaching more specific requests.

Offering a few equal responses, rather than one, also tends to make the treatment last longer. Several modalities give the person room to communicate their way.

When the prompt will not fade#

A common FCT snag is a request that only shows up when a helper prompts it. Harrington uses a stuck case again and again to teach this. A child uses the request after a motor imitation prompt, where a helper models the action to copy. The moment the team delays that prompt, property destruction returns. The plan looks fine on paper, yet the request never becomes independent.

His fix is to turn the vague struggle into a searchable question. "FCT isn't working" becomes "prompt fading within FCT is not working," which points to real literature and a real next step. Thinning the schedule of reinforcement, so the request is not honored every single time, is its own careful stage. It comes after the teaching.

Writing FCT into school goals#

In schools, FCT often lives inside an IEP, a student's written learning plan. Nikki warns against writing a goal too far above the child's current skill. Do not ask for a full polite sentence when the student can only say "no" today. She also teams up with the speech therapist, or SLP, since the goal may sit under communication or behavior.

Lauren Weaver shows the sequence in action. Her team taught a student to say "please give me some more time," then taught him to stay calm when the answer was no. Kelly Brzak keeps early goals realistic too, splitting FCT trials across two family members so the skill travels beyond one adult.

What the research says#

The evidence base for FCT is deep and mostly encouraging. A meta-analysis pools results from many studies. One such analysis covered 34 single-case studies of young children with autism (Blair et al., 2025). It found large effects for reducing challenging behavior. It found moderate-to-large effects for building the replacement skill. It also flagged weak reporting on maintenance and generalization.

A mega-review, a review that combines other reviews, examined six research syntheses in schools (Corr et al., 2025). It reported consistent gains against challenging behavior. Results shifted by setting, communication mode, and who ran the plan.

Newer work looks at the tricky parts. One study chained a compliance procedure with FCT for escape-based behavior (Ferris et al., 2026). Children preferred FCT, and low behavior and high compliance held through fading. Another study tested a fixed-lean schedule to thin reinforcement, meaning rewards come less often, after FCT (Ramirez et al., 2025). Most participants kept behavior low and still knew when to ask.

If you want the full clinical walkthrough, 5 Days of Manding Mastery builds the whole setup step by step. For turning a stuck FCT case into a research question, see Solving Clinical Challenges with Research. For flexible responses with complex profiles, see PDA Caregivers, Complex Profiles, Replacement Behaviors, and Being Trauma Informed.

Frequently asked questions#

What is the difference between FCT and just teaching a child to talk? FCT is not general speech practice. It teaches one specific request that matches why a behavior happens. The request has to earn the same result the behavior used to earn. That tight match to function is what makes it work.

Do I always have to use extinction with FCT? Not always. The classic version reinforces the request and stops the old payoff. But clinicians can reach the same result other ways, like reinforcing the request while shrinking a precursor behavior. The goal is the payoff gap, and there is more than one path to it.

Why does my client only use the request when prompted? This is a fading problem, not a failed plan. The request often shows up under a helper's prompt first. You then fade that prompt and thin the reward slowly, so the request stands on its own. Rushing the fade is a common reason behavior comes back.