- All right.
Good afternoon, everybody.
Good afternoon, Dr. Vollmer.
It's a pleasure to have you with us.
I just wanted to mention a couple of logistics as we're getting started here.
For participants online, I did send out a copy of Dr. Vollmer's PowerPoint presentation.
You'll notice that he's going through his presentation.
There are a few slides missing from the handout that I sent to you.
That's simply because a number of the figures in Dr. Vollmer's presentation are copyrighted.
So we didn't have permission to share those outside of his presentation.
So you'll see those as they come up, in his presentation.
Also to let that Dr. Vollmer will stop at a few key points during the presentation to take questions.
If you have a question at any time when he's presenting, just go ahead and type it in the chat box, I'm going to be monitoring the chat box and writing down questions.
And then, because we don't have everybody accessing the phone line to be able to talk to us, I will ask your questions on your behalf to Dr. Vollmer when he takes those breaks.
So if you type them in the chat box, I'll keep track of them.
And we will, I'll ask those for you as we go along.
So welcome to our February autism and education webinar this year.
We're thrilled to have Dr. Timothy Vollmer with us.
Dr. Vollmer received his PhD from the university of Florida in 1992.
And from 1992, until 1996, he was on the psychology faculty at Louisiana state university.
From there until 1998 he was on faculty at the university of Pennsylvania medical school and returned to the university of Florida in 1998.
And is now a professor of psychology.
His primary area of research is applied behavior analysis with emphasis in developmental disabilities, reinforcement schedules and parenting.
Dr. Vollmer has published more than 100 articles and book chapters related to behavior analysis, and he was the recipient of the 1996, the F Skinner new researcher award from the American psychological association and received another APA award in 2004 for significant contributions to apply behavior analysis.
He's currently the editor in chief of the journal of applied behavior analysis.
And we are thrilled Dr. Vollmer to have you with us this afternoon.
- Thanks, it's great to be here.
I think it was about a year ago or so that I was up in Canada and we were talking about severe behavior disorders and we focused on, socially reinforced problem behavior.
For example, problem behavior that is reinforced by attention from adults or by escape from instructional activity, things of that sort.
And a few of us got to talking and, I acknowledged that the more difficult problem is when behavior is automatically reinforced.
And that is that it's not anything that anyone else is doing per se, that keeps the behavior going, but the behavior itself produces the reinforcement.
And so we kind of agreed that it'd be nice to reconvene someday and talk about that.
So I think that was the original impetus of us getting back together today.
And so as was mentioned, I'm happy to take questions.
This is meant to be informal.
I've never done a webinar before, so this is new for me because strange not to see the audience, but I'm glad to know you're there.
And I, the thing I like most about, giving seminars and chatting with other professionals is, sharing of questions and comments and things like that.
So feel free to jump in.
And if you have a few points in the talk where I specifically we'll pause for question.
So we had a really long title, something about, it's not always our fault or something to that nature, but I just kind of chopped it down to automatically reinforced behavior.
That the point of that title was that a lot of times, when you tell parents, teachers, care providers, others responsible for the care of individuals with intellectual disabilities, when you tell them that it's, for example, your attention that is maintaining the behavior, they sometimes interpret that as, "oh my gosh, I'm doing something wrong.
"It's my fault that the behavior is occurring." Of course, in those cases, it's not their fault per se.
Usually they're doing what any care provider would do to try to attend to the situation.
But in this case automatically reinforced behavior, it's even more explicitly a problem that, is difficult to address because the behavior is producing its own source of reinforcements.
So in terms of what we will be discussing today, I wanted to give you a little bit of background about, what I do in my research.
And I want my current research interests are and how they relate to this problem with automatically reinforced behavior.
I wasn't quite sure of everyone's backgrounds or put that into my prior presentation or anything like that.
So I wanted to start with a brief overview of the functions, the known functions, of severe problem behavior based on the recent research literature.
And I'll talk about how, treatment logic is usually derived from, what we know about the functions of behavior and why we're automatically reinforced behavior, that treatment logic doesn't always apply.
We'll talk about some, new research suggesting there are subtypes of automatically problem behavior, and then we'll go through some treatment approaches, for automatically reinforced behavior.
One thing that I will be stressing is that the prognosis for intervention with automatically reinforced behavior is not as good as it was when we know that it will, behavior is maintained by social consequences.
However, despite the limitations of the known treatment for automatically reinforced behavior, there is some silver lining.
And, I'll try to stress that, but just to give you a little hint about what I'm talking about, let's suppose that someone engages in 100 instances of severe self-injury per day that's a lot of self injury, but I think many of us who see the behavior occurring at that rate if not higher, and, although we may not be able to develop treatments that bring the behavior to zero, as we often do with socially reinforced behavior, suppose that we reduced the frequency to 80 times per day.
Well, that's a reduction of 20 responses per day, 140 responses per week, 140 responses, times 52 for a year and so on.
So it could be just the slightest intervention, effect might have pronounced long-term effects by reducing instances of problem behavior by dozens, hundreds, or even thousands in a month or a year span.
We'll talk about some assessment approaches for automatically reinforced behavior.
And I'm kind of doing this in a reverse order.
Normally we talked about assessment and then treatment, but I think in this case understanding how the interventions work will, make the assessment logic more extensible.
And then I'm not sure how much time we'll have, we'll probably have some time at the end.
I wanted to touch base on some new considerations that we're looking at related to skills training and how the development of skills seems to help replace the occurrence of automatically reinforced behavior.
And also, we'll talk about some strategies for parents and care provider training.
So, my background was introduced a little bit at the beginning.
I studied here at the university of Florida in 19.
I got my PhD received my PhD in 1992.
Some of you may be familiar with the work of Brian Iwata.
He was my major professor my PhD work where Iwata is the individual who initiated the functional analysis methodology for self-injurious behavior when he was a professor at Johns Hopkins university.
And he had recently come to the university of Florida and I was among his first cohort of graduate students.
And it turned out to be a very fortunate time to be a graduate student because all of the treatment studies, following functional analysis were yet to be completed.
So, it was a good time to enter graduate school and be able to be involved in the lab that was developing all of the treatments.
So after my, graduate program, I moved to Louisiana state university.
And the reason that I bring that up here is because, it was my first opportunity to really work with younger children in the schools.
And that kind of expanded my experience because with Iwata I had been working in a large state operated residential facility for adults.
And so we were able to bring the functional analysis methodology into the schools in the early mid 1990s, and then a full switch was done again when I went to the medical school where we were seeing inpatient individual whose problem behavior was so severe that they required hospitalization and those were children and young adults.
And then I came back here to the university of Florida in 1998.
And, I will describe some of the current research sites and interests that I have.
I continued to be interested in assessment and treatment of severe behavior disorders.
Primarily my graduate students, colleagues and I have focused on developing differential reinforcement and non contingent reinforcement as treatment for these behavior disorders.
I have become interested in this notion of automatically reinforced problem behavior, because as I mentioned at the outset, I think this is an underrepresented area in the literature and it's underrepresented because it was much more difficult to develop, safe and effective interventions.
And then finally, I am interested in parent and care provider training.
I view our job as, researchers and practitioners as being incomplete, unless we have the intervention in place in all, the of the areas where the individual resides studies, works and so on.
My current research takes place in, we have a clinic right here on campus to behavior analysis research credit.
We collaborate with a private agency called the Florida autism centers.
And we work with them throughout the state of Florida.
We do work with the state of Iowa department of education.
We collaborate with the Gainesville area schools.
I live in Gainesville, Florida, and then we also work with the rural school districts surrounding Gainesville.
Also Gainesville is sort of a university town, but surrounding Gainesville is the old South in the United States with a lot of rural areas and not much in the way of services for individuals with intellectual disabilities.
So we contract out and our board certified behavior analysts go to those surrounding school districts.
So that's kind of what I do and where a lot of his information comes from.
I wanted to make sure we're all on the same page in terms of what I'm talking about when I say severe problem behavior, I may say severe behavior disorders.
I kind of use those interchangeably as I also want to be clear that I'm not intending to say that the individual is a problem.
If clearly, sometimes the behavior that the individual engages in is a problem.
It's a problem because the behavior may produce self-injury.
So the individual individuals getting hurt by their own behavior.
Sometimes they may need to be hospitalized because of that injury.
Aggression is problematic, often because others in the environment are getting hurt and might be children, other children, it might be other adults.
It might be adults in the environment.
Property destruction can be very damaging and especially when their expensive equipment is being broken, like computers, windows, car windows, things like that.
Severe tantrums would fall under a severe problem behavior often because they keep the family or staff or classroom teachers from taking the individual into community environments, such as shopping malls or restaurants.
And it's also disruptive to the classroom environment.
I included a couple of miscellaneous examples, but this list could go on.
We sometimes see children who engage in dangerous climbing.
So they'll climb up on furniture or bookshelves or jump down from them and things like that, or elopement, which basically means running away from an environment without permission to do so.
So when I'm talking about severe problem behavior, really this list is what I'm referring to.
And sometimes it might even include things like PICA, such as eating inedible objects.
And I'll talk about that problem just a little while.
So before we get into the discussion strictly about automatically behavior, I wanted to summarize what is known from the research literature about the operant functions of severe problem behavior.
And when I say operant functions, I don't believe that all of you in the audience are behavior analysts.
So you might not know what I mean when I say operating functions, what I'm referring to is a type of behavior that is, strengthened by its consequences.
So for example, when you turn on the light switch, the behavior turning on the light switch is strengthened by the fact that light comes on after you looked at that switch.
If you didn't flip, if you flipped the switch and the light didn't come on, you might test it out a couple more times, but then you would stop.
You wouldn't keep engaging in flipping on the lights with, because the light doesn't come on.
Well, similarly, severe problem behavior occurs by and large because of the consequence it produces.
If it produces a favorable consequence for the individual, who's engaging in the behavior, that behavior will persist.
Some of the known favorable consequences, if you will include positive reinforcement in the form of attention.
So a lot of people tend to think of positive reinforcement as being only a good thing like, it sounds good, doesn't it?
It's positive, it's reinforcement.
Its some form but positive reinforcement can maintain a lot of bad behavior.
So for example, when, a child displays severe self-injury and natural parenting response is for the parents to come up and comfort the child, maybe hold the child in their lap and say, it's okay, you're gonna be okay.
Or it might even be in the form of a teacher, reprimanding a child while the reprimand was meant to scold or punished the child's behavior.
But for some children, that type of attention can actually positively reinforce the behavior.
So all we really mean by positive here is that something is presented.
The attention is presented as a consequence to the behavior.
And reinforcement means that the behavior is strengthened.
So that attention that we provide that is meant to comfort the child or reprimand the child has been shown repeatedly to serve as positive reinforcement.
In fact, for self-injurious behavior about 25% of the cases of self-injurious behavior have been shown to be reinforced by attention.
Other types of positive reinforcement include, access to tangible items.
So for example, a child might have a major tantrum and that's reinforced when the parent gives the child a drink for example, they're trying to calm the child down.
And the parent's behavior is reinforced by the offset of the tantrum.
So once they figure out what the child wants to drink or wants to eat, or what toy they want to play with, their iPad or whatever it might be when the child is calm.
So the parent's behavior of giving the item to the child is reinforced.
And the child's probably behavior is reinforced by virtue of getting access to that item.
Another category is negative reinforcement in the escape from the form of escape from instructional activity, self care activity, aversive social context, maybe loud noises like in a gymnasium or something like that, when the problem behavior starts and someone in the social environment turns off the diversity of stimulation.
So they might stop the instructional demand like, okay, well, we'll do that later, or let's get him out of this from gymnasium.
He obviously doesn't like this.
And so the behavior is negatively reinforced in the sense that some stimulation is at least temporarily attenuated or turned off.
And that's all that's meant by negative is it's turned off.
Positive if it's turned on, negative it's turned off in this case the loud noises, the instructional activity is turned off and that the individual learns to escape from that activity.
The key thing to notice about these first three operating functions is that, although they're very common, I mentioned that positive reinforcement is about 25% negative reinforcement in epidemiological studies seems to run between 33 and 37% of the cases, especially for self-injury, but also seemingly for aggression.
But in all three of these types of reinforcement, the reinforcer is provided by another person in the environment.
So attention is presented by another person.
The tangible item is delivered by another person.
The instructional activity is turned off or the self care activity is turned off by another individual.
So as complicated as those arrangements might sound, once people in my field see those relationships, we start going like this because we know we can treat that behavior.
And here's how, for example, with attention, we know let's minimize attention when the problem behavior occurs and teach the person a new way of getting attention, maybe by saying talk to me or using sign language or picture exchange cards.
So we can differentially reinforce some alternative behavior.
And the literature is now filled with hundreds of examples of successful treatment of behavior maintained by social mediated positive reinforcement in the form of both attention and tangible and treat as treatment for negatively reinforced behavior.
Now where the literature is not so cool is when we get down to the next categories and that is automatic positive or negative reinforcement.
Sometimes it is the case that the behavior simply provides its own source of reinforcement.
So an individual could be in a room all by themselves, perhaps, especially when they were all by themselves.
And the behavior continues to occur, even though no one else is there, no one is providing reinforcement.
Why is that?
Well, the behavior is producing its own source of reinforcement.
There are a number of theories about what that source of reinforcement might be, and we'll touch on those as we come through the discussion today.
But the key point here is that these forms of reinforcement are not socially mediated.
In fact, this term, automatic reinforcement dates all the way back to B.
F.
Skinner the famous behavioral psychologist who passed away many years ago, he used the term automatic reinforcement simply to me that the reinforcement is not delivered socially.
It's not socially mediated.
So that is the special problem that we are addressing today.
I could go on and on about social mediated reinforcement.
In fact, I usually do, I can talk about that for an entire semester or multiple day workshop.
And that's what we did when I came up to Canada last time.
But this special problem, I think deserves some discussion in its own, right?
How do we develop interventions?
How do we figure out what's going on with the behavior is producing its own source of reinforcement.
It's much easier to tell someone, stop attending to that behavior, or don't allow escape when that behavior happens, allow escape when they finish their work, that's easy to do like cutting off sources of reinforcement that are automatically produced or competing with reinforcement that is automatically reinforced is very difficult.
You can think of, analogies with, populations who are not intellectually disabled.
And think about things like cigarette smoking.
Cigarettes smoking isn't typically maintained maybe at the beginning, it is by socially enforcement, but after a time most cigarette smokers will smoke when they're alone or maybe even especially when they're alone, because the behavior is producing its own source of reinforcement.
Same with, substance abuse, alcohol consumption, things like that.
And automatic reinforcement doesn't always have to be a bad thing just as positive reinforcement isn't always a bad thing.
We do many things that are automatically reinforcing, like when we read a novel, do a crossword puzzle, things of that nature, persist in the absence of any kind of social reinforcement.
But in this case, we're talking about automatic reinforcement, maintaining problem behavior.
So I mentioned that that list that I just described is from the research literature there's a very strong empirical basis for what I just asserted hundreds of research studies that supported those findings.
So how does one usually go about identifying the function?
Well in practice most commonly, there are three approaches used to make a reasonable hypothesis about what the function of behavior is.
One is through the use of indirect assessments.
These are typically in the form of questionnaires, checklists, open-ended questionnaires to get the conversation going about what a teacher or a parent or an administrator or another professional might hypothesize is reinforcing the behavior.
We also use descriptive methods which involves collecting data on the behavior to see when in time that is most likely to occur is it during instructional activities is it during self care activity is when the teacher's behavior is, or excuse me, teacher's attention is diverted temporarily from the students.
And it allows us to kind of watch in the natural environment and figure out what might be going on.
But the most sound methodology, the one that is most supported in the empirical literature is the functional analysis method.
And this is the approach where the consequence is intentionally withheld during the assessment until the problem behavior occurs.
When the problem behavior occurs, the therapist or the individual doing the assessment intentionally provides the consequence they're testing.
So for example, in an attention condition, they would intentionally provide attention following an instance of the problem behavior.
I and others have often used the analogy of an allergy test to describe the logic of a functional analysis.
When you go in for an allergy test, your, allergists will do something like poke you in the back and look to see what swells up.
If she sees that you swell up in the presence of a certain type of grass or in the presence of a certain kind of food, she will say you are allergic to this type of grass or this type of food.
Well, similarly, in a functional analysis, we're trying to see where the behavior swells up to tell us what is maintaining the problem behavior.
Well, we test things like attention as a reinforcer, tangible items as a reinforcer, escape from instructional activity as reinforcer but we also have a condition most pertinent to today's discussion in which the individual is watched with no consequence.
In fact, ideally they're watched when you're in a room alone.
And if the behavior persists under those circumstances, it is most likely, automatically reinforced.
Here's an example of an individual in one of our studies published a few years back, Elizabeth Athens who is now working in Canada, actually.
What we found in this particular, functional analysis was this individual George was engaging in aggressive behavior that was reinforced by attention.
So if you look at these sessions with the open square, these were all the sessions in which we were intentionally giving George attention, following his aggressive behavior.
And these other conditions we were giving him escape, tangible items or no consequence it's, no worthy that only when he was getting the attention did the aggression persist.
Okay.
So, the key thing I saw that some of you can't see the cursor, well, what I'm referring to here is this line, this data path floating above all of the other data paths.
This represent a high rate of aggressive behavior and the condition where he's producing attention.
So we would say that George problem behavior is aggression is socially reinforced.
It is reinforced by the attention that it produces.
Here's another example from that same study, Clark who only engaged in the aggressive behavior when it produced access to his favorite, tangible items, particular toys that he liked to play with in all of the other conditions it hardly ever occurred.
So this is another example of socially mediated positive reinforcement.
Here's a case Henry who only engages in the aggressive behavior when it produces escape from instructional activity.
So in his case, the problem behavior is still socially mediated, but rather than through positive reinforcement, by gaining access to attention and or tangible items, it's by having instructional activity turned off.
Now, I only show you these to contrast with two examples of automatically reinforced behavior.
I selected an early study that we did in 1997, where we found, two essentially subtypes of automatically reinforced self-injurious behavior.
In this upper example for Barry, he engaged in one form of self-injury that we called hand biting, where he would clench his teeth down on his hand.
And it occurred in every condition that we tested.
So he was biting his hand, no matter what test condition he was in.
And then we even conducted a series of no interaction sessions, where he was not getting any consequence for the hand biting whatsoever, except for the automatic consequence and the behavior persisted .
So she is not engaging in this behavior because of some social consequences.
He is engaging in the behavior because of some automatic consequence to the behavior.
It couldn't be the social consequences because there is no social consequence in this condition.
Now his other types of self-injury hand and body heading only occurred in the no interaction condition.
So that's a little bit different than what we see it here, but it represents what might be another subtype of automatically reinforced behavior.
And that is the type that only occurs under conditions of low stimulation.
And I'm gonna come back to this point, because these are the two types of outcomes that we generally see in a functional analysis that would lead us to say the behavior is not socially reinforced, which is to say in another way, the behavior is automatically reinforced behavior.
Okay.
I see one question here that says, when you say automatic, are you meaning sensory, at least in this one case?
A lot of people use the term sensory.
But we typically try to avoid that term because really anything that is, contacted through the senses could be considered sensory.
So even attention is in a sense sensory, right?
Because it's something that person either sees or gears or touches.
We were trying to distinguish here when we say automatic, between socially delivered and not socially delivered.
So that's the distinction that we're making.
So, I gave the example doing a crossword puzzle.
Let's suppose you do a crossword puzzle.
A lot of times when one finishes a crossword puzzle, they look at it kind of smile because they're completed with it.
And then they throw it in the trash can, they're not doing it because someone comes up and says, "oh wow, you're really smart.
"You did the whole crossword puzzle." The behavior produces its own source of reinforcement.
So Chris says.
Now we could call that sensory or not, but that's not the distinction that I'm making here.
The distinction is socially mediated delivery of reinforcement versus not socially mediated delivery of reinforcement.
Does that make sense to the person who wrote that?
I guess it's Robert, he said, I get your distinction, but I would still view it as sensory.
That's fine there may be different terms that people are familiar with or are comfortable with.
But again, the reason I'm using the term automatic is only to delineate whether it's socially mediated or not.
We'll start to get into what the specific sources of reinforcement might be in a few minutes.
Then another question from Christine is, would you say that there is a stark difference between a tantrum and a meltdown?
When I used the term tantrum, I, there's a continuum.
There could be a very mild tantrum, some fussing and crying to what typically people call a meltdown.
Neither of those are technical terms.
So I'm not necessarily distinguishing between the two, but I think that people would refer to a meltdown as a severe version of the tantrum.
So it would be just along the same continuum.
Okay, this is a really good question by Simon.
If a student is disturbed by noise, should he or she be removed from the noise or left to become conditioned to it?
I think that, actually there's not a lot of research on that.
Escaped from noise, which is surprising because all of us who have worked, especially with individuals with autism spectrum disorder, know that it's very common that you'll see kids plugging their ears and starting to get riled up when there's sound.
From a clinical standpoint, what we have tried to do is teach the individual another way of saying, I need to get out of here basically.
So the, what we're reinforcing is the appropriate way of saying I need to get out of here and then allowing them to leave.
The second part of your question is also interesting because it could be the case that there simply needs to be some habituation that would need to occur.
So maybe the noise of the gymnasium is a versus, but with a little bit of time in the gymnasium, the person might start to get used to the noise if you will, and it might become less aversive to them.
So what we do clinically again, is gradually start to increase the amount of exposure that the person has to that sound before we honor the get me out of here communication response.
So you can sort of combine a functional communication approach with a habituation slash desensitization approach.
But there really needs to be a lot more research on that noise phenomenon because it's quite pervasive.
Thanks for that good question.
All right.
So I've sort of already implied the treatment logic for socially reinforced behavior.
And that is the reason that the prognosis is good or socially reinforced behaviors is because by identifying the reinforcer via functional analysis.
Reinforcer reinforcement for probable behavior can be minimized.
In other words, let's suppose we find out that the behavior is reinforced by attention.
We can tell people or suggest to people minimize your attention.
Don't give a lot of attention to that behavior.
And that reinforcer attention can be used to strengthen some alternative behavior like functional communication or appropriate play.
So we can provide the reinforcer either contingent on some alternative behavior or non contingently, give them free attention to reduce their motivation, to engage in problem behavior.
Now, this approach simply said, it just works.
It's, there's no question about it.
That the research is sound on differential reinforcement, non contingent reinforcement.
You can suppress socially reinforced problem here.
That's why the prognosis is good when we find that the behavior is in fact socially reinforced, there are a couple of glitches with that notice in the first bullet I put ideally extinction, while sometimes you can't completely ignore problem behavior or you can't avoid giving the person a great period of escape, such as if they're a very large or dangerous individual.
That's why I'm always careful to use the term minimized.
So you can minimize the reinforcement while maximizing reinforcement for the appropriate behavior.
We've done some research on what is called the matching law.
The matching law essentially tells us that people will allocate their behavior according to the larger payoff.
So if you get twice as many reinforcers for pressing lever A as you do lever B, you're going to allocate twice as much of your behavior to lever A because it produces twice as much reinforcement.
Well, similarly, our research with children has shown us that if we provide twice as much reinforcement for problem behavior, as we do appropriate behavior, well, guess what, they're going to engage in twice as much problem behavior.
And so their behavior matches the schedule of reinforcement and it turns out to be amazingly uniform.
We talked about that last year when I came, but it's quite phenomenal to see.
And what it suggests is if we're able to shift the reinforcement rate for appropriate behavior, we start to see more and more appropriate behavior.
And in fact, that's what we see.
So normally a functional analysis is very helpful in developing these kind of treatments, but not so helpful with automatically reinforced behavior.
Because if you think about it, how do we cut off the source of reinforcement if it's produced automatically?
And also, how do we provide that source of reinforcement for some alternative behavior when we don't necessarily even know what the exact reinforcer is, or how do we, how we could bottle it and deliver it for something else.
So when I conduct a functional analysis, if it shows me that the treatment of socially reinforced behavior, that we're, it's leading to a treatment for socially reinforced behavior, then I feel good.
I feel very confident that we can affectively treat the behavior with good parent and staff training.
When it shows me that the behavior is automatically reinforced, I don't feel so good.
So it would be roughly analogous to if you went to.
You're a doctor and there was a very treatable form of cancer.
The doctors used hundreds of cases that we can treat this and you're not out of the woods, but it's, I feel confident versus one where the physician would say, this is we're down.
We're going down a tougher road here.
And that's what I think we're dealing with with automatically reinforced behavior.
So I want to stress that in this discussion, because we're not talking about magic bullets here, we can kind of have the magic bullet for socially reinforced behavior.
We don't for automatically reinforced behavior.
So everything that I'm going to describe here forward, you're going to immediately see some holes in it and say, yeah, but what about this?
Yeah what about that?
And you're right.
What we have to do is fall back on that silver lining that if we can influence even some percentage of automatically reinforced behavior in the long run, that magnified to dozens or hundreds or thousands of instances of the problem behavior.
So that's what we're shooting for here.
And in the best case, we wipe it out, but that's not gonna happen that often.
One question that typically arises when I'm giving seminars on this topic or in the literature, when we're discussing this topic is, well, isn't it possible that automatically reinforced behavior isn't operant behavior.
Maybe it's not maintained by its consequences.
Maybe it's something else while there is a lot of evidence to suggest that it actually is operant behavior, in most cases, I'm going to show some possible exceptions that need more research.
But the evidence that behavior that persists in the absence of social reinforcement is operant that is controlled by as consequences is as follows.
One access to the behavior itself increases behavior that produces that access.
So that sounds kind of strange, but let's say that someone engages in hand flapping like this, the reason that we know that that's operant behavior is because if you restrict access to the flat flapping and have them do something like press a button to get access to the flapping, the rate of button pressing increases.
So that tells us that the stimulation produced by the hand flapping is reinforcing stimulation, and it's not produced by someone else it's produced by the behavior itself.
Similarly, if you can somehow block the stimulus products to the behavior, it does go away.
Now that's very difficult to do, but let's say in the ideal world, if somebody was doing hand mouthing and you were able to block every attempt to hand mouth, then eventually the behavior does go away.
So that suggest it must be off the operant because it's not getting that consequence that was maintaining it.
Therefore it stops just like you would stop turning the light switch on if it didn't produce light.
And then third, the evidence that automatically or behavior that persists without social reinforcement is in fact awkward behaviors that we can cite numerous examples of behavior.
That's known to be maintained by automatic reinforcement.
Like the ones I talked about a short time ago, substance abuse, cigarette smoking, doing crossword puzzles, reading a novel, those forms of behavior persists in the absence of any kind of social reinforcement.
So it stands to reason that probably behavior could also produce access to automatic reinforcement in individuals who have intellectual disabilities.
Now I mentioned that there is some possibility that the behavior that persists in the absence of social reinforcement could be maintained by other mechanisms.
One such response is biting.
So someone was asking the question about noise a while back.
Have you ever seen, I'm sure many of you have, an individual who as soon as they hear some aversive noise, like a truck starting up or a vacuum cleaner start they immediately bite their hand or they bite somebody who's next to them or they bite their desk or something like that.
Well, there is a phenomenon that's been shown across a range of organisms, including humans.
And this has been known for many, many years that, a burst of stimulation, elicits a biting response.
So in the early experiments with monkeys, monkeys will bite down on a bar.
If it's in front of them when they receive a shock, if the bar is not there the monkey will bite their hand because there's something about aversive stimulation that causes a biting response.
So if in the days before anesthesia people, would bite down on a towel when they were having a bullet removed, or if a woman was in childbirth, she would bite down on a wet towel or something like that.
That biting response seems to go along with aversive stimulation.
So it's possible that something like that is going on when we see behavior like self biting, biting of others, I think we should classify that as open or more research is needed, but it certainly makes some sense.
Another interesting old line of research, I didn't put the date here, but Nathan Azrin did a series of studies in the 1960s showing that organisms of all kinds will attack under certain types of schedules of reinforcement.
So for example, if a pigeon is getting reinforced food reinforcement for packing the key, and then you stop providing the food reinforcement, if you'd have a dummy pigeon like a fake pigeon in the back of the chamber, that pigeon will go over and just start attacking that fake pigeon and hazard showed this across a variety of species.
Now there would be ethical issues about demonstrating this with humans, but if we extend, it might make a certain amount of sense that damage to another organism might reinforce the behavior in some way.
So think about road rage or something like that, where, effectively the person in power never even comes in contact with the other person, but they see their face getting mad.
They see their face getting upset.
So they're flipping them off and shouting at them and things like that.
What would maintain that behavior?
Well, some of them theorized that damage to that other organism is what is maintaining the behavior.
So it's not their attention or reaction per se, but just that the facial grimacing, the pain that it produces in them.
But all of those mechanisms need more research, but I put this slide up to say, I, at least, I don't think we can say we've ruled out these other possible mechanisms with biting and aggression as possible sources of maintaining variables.
So let's talk about how, treatment typically progresses when the finding is that probably behavior is automatically reinforced.
Well, typically the first thing that is done is environmental enrichment.
And what I mean by environmental enrichment is simply that the person has given access to their most highly preferred stimuli.
If you remember that functional analysis where I said, it's a second type of automatically reinforced behavior for that child Barry the highest rate of problem behavior, occurred when he was alone with no stimulation.
So what that suggests is all you really have to do is give the child highly preferred sources of stimulation and problem behavior is attenuated.
And so in a small number of cases, it's as simple as that in times when the person would be bored or left without something to do you find out what their most highly preferred stimuli are, their favorite toys, if they an iPads, favorite leisure materials, whatever it is, and still the environment with those.
An example of that, whoops forgot.
I'm not running my own computer here is depicted here in a study that we did, long ago, 1994, this particular child, Ron engaged in self-injurious behavior in the form of ear picking.
And what would happen is blood would start flying out from the back of his ear and blood would get on all over him and all over the room and things like that.
And in the baseline condition, we measured the percentage of our observation time when he was engaging in the self injury.
And it hovered around 50% of the time.
Now, in his case, all we did was conducted a stimulus preference assessment.
So we put things out in the room and looked to see what he would interact with what he would interact.
And when we use the items that he preferred in the EE plus condition, that's environmental enrichment with preferred items, he played appropriately with those items and he did not engage in the self-injurious behavior, at least not at nearly as high level.
Now, importantly, we switched to a condition called EE minus where we filled the room with an equal number of items, but they were ones that were not preferred and the self-injury came right back up until we put the preferred items back in the room.
So in his case, there was a pretty easy fix to the self-injury.
We just, in times when he would have been in downtime or bored or not having much to do give him something else to do.
Now, unfortunately it's not always that easy.
And that's why we have to progress down this list.
And secondly, we would go to differential reinforcement, and this is where instead of just giving free access to highly preferred stimulate and highly preferred items, we would give access to those, highly preferred items contingent on some alternative behavior.
So let me show you an example of that.
Korey, in the same study, when we gave him the enriched environment with preferred stimuli at first, it increased his appropriate behavior, which is indicated by the open circles and it somewhat decreased his self-injurious behavior, which was indicated by the closed circles, but it wasn't until we explicitly reinforced toy play with attention and access to bits of food that we saw, the further increase in toy play and a further reduction in problem behavior.
Then we returned to the baseline and the self injury returned.
Then we went back to the differential reinforcement condition and over time his self-injury started to resurface.
So we had to do a new reinforcer assessment to find out things that he was now interested in because his preferences had changed from the final condition when we did the new assessments, that means we found new items for him to play with such as switch toys and his self-injurious behavior reduced further, and he maintained a relatively high level of appropriate play.
So that outcome is different than this previous outcome in the sense that the enriched environment helped, but it didn't completely reduce the behavior to a level that we felt was important.
Now also know that his self-injury never completely goes away.
If you look at the literature on socially reinforced behavior, usually the final condition is pretty much a flat line.
The behaviors gone.
We did not completely get rid of the behavior, even in the best of all circumstances for Korey, but go back to that silver lining that I talked about, we're seeing probably about a 90% reduction here and maybe even a little bit more, but suppose he has 100 instances of self-injury per day.
Now we're talking 10 per day, with that type of reduction.
So 90 per day times 30 days and so on, like I was giving you the example that multiplies by thousands of multiples of 90, per day and thousands of responses per month.
Now going down this progression often environmental enrichment and differential reinforcement are not effective.
So usually some kind of blocking or brief time out or mild punishment has to occur in conjunction with these, particular intervention.
And so in that same study, Rhonda was a girl who had happened to have Rett syndrome.
I'm not sure how many of you are familiar with Rett syndrome.
It's a relatively rare, developmental disability it's seen only in girls.
And what they typically do is, they engage in hand stereotypies in at the midline.
So down the center of their body, and they're also prone to water related stereotypies such as running like winning their hands under water.
In Ronda's case, what she did was she would rinse her hands and keep them in her mouth, such that she had worn the skin down on her hands.
It had become a self-injurious mouthing if you will.
And when we tried the enriched environment approach with her, when it was non-preferred toys and items in the EE minus, we saw a very minor reduction in her mouthing and some level of toy play.
Then when we introduced the preferred items, we were able to reduce it by about 50%.
And her toy play was up to about 50%.
Then we introduced the differential reinforcement like I did with the prior intervention.
And it worked temporarily, but started to wear off.
And it wasn't until we introduced a timeout, which in her case was a brief, I forget this at the top of my head, if it was five or 10 seconds, I believe 10 seconds of hands down at her side, we would hold them and count to 10 and I'd let go.
And it wasn't until we added that time out components that we were able to have a more effective intervention for her.
And even in some cases, the blocking and relatively mild forms of punishment, like time out are ineffective.
And in those cases, if the behavior is severe, that's when, people are forced to make difficult decisions about whether they'll use restraining or more extreme forms of punishment, depending on what is considered acceptable within the therapeutic community, in which they live.
Where I live we typically can't go much more extreme than that and not that I would want to, but in the case of very severe self-injurious behavior, or very severe aggression, it has to be recognized that sometimes even the more, the milder forms of punishment are not as effective, It's also important when dealing with automatically reinforced behavior to consider the possible functions of response blockage, I'm not sure how it is in Canada.
If somebody could respond to me quickly about that, I'd appreciate it.
But in the United States, almost every written behavior plan has response blocking in the plan or maybe it says redirect, which also implies a physical blocking and moving.
Is blocking and redirection fairly common components to behavior plans where you live?
Okay.
Robert answered, yes.
So we see that too.
And I think it's very important to note that, with response blocking conserve really four very distinct functions, I have three listed here, but one is, it can actually positively reinforced the behavior.
So when somebody starts engaging in problem behavior, the blocking might provide physical contact that they actually like.
And so the blocking increases the problem behavior that can function as extinction, such as, as the example I gave about the hand mouthing, where if the person isn't allowed to complete the hand mouthing response, then the behavior goes away.
And in some cases, it actually functions as punishment.
There's there are studies to show that that hand contact punishes the behavior, and you can distinguish between the extinction and punishment, if you block every other response, if you block every other response and the behavior still goes away, it was blocking was punishment.
If you block every other response and the behavior increases or stays the same, then blocking probably was extinction because for extinction to work, you have to block every single response.
So these are three very different effects.
And another one is that it might have no effect blocking might do nothing.
So what I would highly recommend is in any behavior plan where you see redirect the response or block the response, you probably should do a very brief functional analysis to figure out what blocking actually does to this behavior for this particular individual is thrown in there in almost every plan that I see.
But with little information about what it is likely to do.
So a quick little task might look like this.
You have one condition that is play, where they just can have preferred items, another condition where you provide no consequences to the target behavior and a third condition where you block it.
If the blocking increases the behavior is probably positive reinforcement.
If the blocking decreases the behavior, it's either an extinction or punishment.
If it does decrease the behavior, you move to a second comparison, which is block every other time versus block every time if you see the effect, when you block every other time then blocking is probably a punishment procedure, I think it's just important to understand what all the little components that might seem like add ons to the behavior plan.
They might influence behavior in important ways.
So let's pause again for any questions that you might have.
I see one from Amy that says, can SIB be maintained by escape?
Yes.
Now we're here where we're focusing on automatically reinforced behavior, but actually the most common function of SIB is escape.
Usually in the thirties percent of the cases of self-injurious behavior are maintained by escape from something.
So it could be escape from instructional activity, escape from self care activity.
So yes, most definitely it can be maintained by escape.
- Okay, one of the other questions, Robert wondered one issue with extinguishing automatic or sensory behaviors is that they're often replaced with another similar or sometimes more problematic behavior.
Any thoughts on that?
- Yeah, and we're gonna be coming to that.
My group and Cathleen Piazza's group have done some research on that.
And there are assessment methods that can be used to identify appropriate replacements.
So it's interesting that sometimes even the highly preferred stimulus will not necessarily replace the problem behavior.
So you have to do an assessment to see not only is the stimulus preferred or the toy or the item preferred, but what is its influence on the problem behavior?
So we try to make it so that by environmental enrichment and differential reinforcement, instead of them moving to some other form of stereotypic behavior or other form of self-injurious behavior, that they would shift allocation to at least sometimes interacting with the item.
Now that question is, really important because it's one of the reasons that the prognosis is not as good for automatically reinforced behavior, because then they can just switch to another automatically reinforced form of behavior.
So we kind of take that silver lining approach and say, can't we cut into this by 10%, 25%, 50% and take what we can get.
Ideally we want to kind of do a much more, but it requires very extensive, assessment and even then it's not always effective.
And I will show some data from those studies in a few minutes.
- Okay another question was around the fact that occupational therapists sometimes recommend an item called jewelry or like a necklace or a bracelet that a child can chew on and in lieu of engaging in a problem behavior, is there any research that you're aware of on that?
Or what are your thoughts on that?
- Yeah, well, sometimes behavior analysts and occupational therapists come at things from conceptually different angles.
But from a standpoint of why something might work or be effective, we might look at it conceptually differently, but procedurally, it ends up being the same thing.
So, whereas an occupational therapist might say, this is a component of a sensory integration type of therapy.
We might look at it as this is an alternative response that produces the same general type of stimulation, or maybe even a more preferred type of stimulation.
And if the alternative is that they would suck on their hand and wear down the skin on their fingers, I'd rather have them chewing on the item.
So until I can find something more age appropriate or something more socially appropriate, I would be in favor of having them chew the chewy or whatever it's called, because it reduces the likelihood of, spreading germs.
It reduces likelihood of infection of the, from the open skin wounds and things like that.
So the conceptual differences, I can look at it as a component of sensory integration, but that type of therapy might be effective for a given individual, but I would put it into the assessment that we're going to talk about shortly to see if it actually does compete with the target behavior and to see if it actually is a preferred item for the individual.
I often get questions about PICA, which is eating inedible objects and vocal stereotypy.
These are common, relatively common forms of problem behavior.
So, and they're usually automatically reinforced.
In fact, in my clinic in the last five years, we have not had a single case of vocal stereotypy or PICA out that was not automatically reinforced.
In other words, it usually persists in the absence of any social consequence.
And in fact, it's most likely to occur when there is no social consequence.
Oh, that didn't turn out good on your, the one that you have.
Well, let me just kind of summarize the approach we've been taking to PICA.
One is that we try to get or teach the individual to trade in the PICA item for a highly preferred food reinforcer.
So for example, one guy that upper figure, which you cannot see very well, would eat crayons and chew them and actually swallow the crayons.
So we had crayons sitting out and we taught him to a simple, responsive hand.
The therapist is Fran, and then he would get grapes and other preferred food reinforcers.
And, then the therapist gradually increased her distance away from him until he actually had to go out of the room to give her the crayons and get the grape.
So that of course depends on the grape to being more powerful reinforcer as the crayon but it shows that it is possible to compete with the PICA in that way.
And the lower figure, which you also cannot see very well.
We taught a girl who engaged in sand and dirt PICA here in Florida the ground is very Sandy.
And so when people walk into the clinic and things like that, where there's carpet, they will leave little grains of sand.
And it's almost impossible to keep up with that, even if you vacuum it constantly.
So this girl would flop down and start eating the sand grains.
So instead of having her trade in the sand grains like we did for this boy, we thought it'd be more appropriate for her to throw them away.
So we taught her to throw them away, and then she would get access to bites of her preferred food, which in her case was Oreos.
So we were fortunate in these cases that we were able to find food items that were more preferred than the PICA items, but it also highlights the fact that it is possible to do so.
So if you can, you can find some, appropriate alternative, like handing the PICA item to the therapist or drawing the PICA item away, depending on what it is.
There is a procedure called response, interruption and redirection that is commonly, that is currently very commonly used, for vocal stereotypy.
We in fact adopted it and started using it in our clinic.
But one of my former doctoral students Carol Wunderlich who is now at the university of Georgia, was running into negative side effects.
So every time she would do the procedure, what this procedure involves is as soon as the child engages in vocal stereotypy, you would start placing demands on them and would continue to place the demands until the vocal stereotypy stopped for a period of time.
And there, there is a lot of research supporting it.
But we kept getting kids for whom it wouldn't work.
And we were wondering what are we doing wrong?
Well, what Carol Wunderlich found out was that if you plot the data, such that you only count the vocal stereotypy that occurs before the intervention that is before you start repeatedly asking the child demands over and over again, it looks wonderful.
The upper panel shows what appears to be a nice treatment effect, for vocal stereotypy, but you'll see on the right.
I called it a false positive because when we included the data from the time when we were placing the demands on the child, after the vocal stereotypy, there's almost no difference between the treatment and periods and the non treatment periods.
So we're sort of cautioning people that response intervention or response interruption and redirection for vocal stereotypy might at times be a false positive.
And if you're interested in that study, it was published in 2014 in the Journal of Applied Behavior Analysis.
So what do you do with the stereotype of vocal stereotypy?
We have had some successful cases.
Here's one example where we found that when an adult was present giving attention, the child was less likely to engage in the problem behavior.
So we took advantage of this by, at first, having an adult present at all times.
And then we would have a contingency in place where we would present demands when the vocal stereotypy happened just like in response interruption and redirection, but also we had a response cost contingency which is the RC.
So when the vocal stereotypy occurred, we would take a preferred item away from the individual.
You can see that, right when we started that treatment, the vocal stereotypy went down to near zero.
Then we went back to the baseline, the vocal stereotypy came back, went back to the full treatment.
It went away and then we pulled out bits and pieces of the treatment to see what was required to maintain the effects.
And it works best when all three components were in place, but you can't have an adult there all the time.
So the next phase of the intervention evolved gradually decreasing the amount of time that the adult was there, such that we started out by having the adult present for 60 seconds and absent for 20 seconds, then absent for 30 seconds and 40 then 50.
And then so on up to the point where eventually the adult was gone for 300 seconds or five minutes, and we were still able to maintain zero levels of the behavior.
So vocal stereotypy is treatable.
It just might require so sort of idiosyncratic testing of what is going to be most effective on an individual basis.
So to summarize the treatment findings, there are some considerations.
One is, you should always identify stiff preferred stimuli.
You should evaluate what the effects of the preferred stimuli are on problem behavior.
So, if a person is chewing on the chew toy and they're still engaging in problem behavior, well, that doesn't do much good.
You have to see that there's a negative correlation between the two.
You should see what the effects of response blocking are on an individual basis.
And then one we haven't talked about yet is you should consider the role of skill development.
And there's an interesting literature that shows that the more skills that an individual has, the less likely they are to engage in stereotypic or automatically reinforced behavior.
So we're sort of taking a long range approach in some of our work where very early on we're teaching kids routine skills, like play skills, social skills, self-care skills in an effort to build a repertoire that will again compete with at least some percentage of the occurrence of the automatically reinforced behavior.
And then finally, as I mentioned at the beginning, none of this matters unless the people taking care of the individual are trained to implement the procedures that you are using.
I mentioned the Cathleen Piazza has done some work.
What she does in her assessment is she compares interaction with the item.
So if you look way over to the left, this person interacts a high percentage of the time with shaving cream and the black bar represent how much time the individual engages in the problem behavior, which in this case was saliva play.
Well while interacting with the shame cream, this person hardly engaged in any saliva play.
So the important part of this, is you would wanna look for things like, well with, lotion, the person interacts with lotion quite a bit, but they're still engaging in the problem behavior, the saliva play during the lotion play.
So by measuring the amount of the problem behavior and the amount of play with the alternative item, you can get a handle on what would be the best replacement.
So going back to the question about chew toy you could do this by giving a chewy toy and measure does it compete with hand mouthing or hand bite or whatever that particular child, response pop out happened to be?
This is a case where we worked with a child with severe emesis.
What she would do is engage in self induced emesis, where she put her hands into her mouth and down causing herself to vomit.
When, what we did in this intervention was simply at first blocked every single response.
And what we found was that her attempts in the upper panel attempts to engage in this health induced emesis gradually went down to near zero.
Her actual instances of emesis were at zero with the exception of one session around session 25.
She had one instance of emesis but when we reduced the blocking to 50% toward the end, the behavior returned, she started putting her hands back at her mouth, and she started engaging in the emesis again.
So going back to that earlier point I made about is blocking extinction or is locking punishment.
In her case, blocking was extinction.
Because as soon as you reduced it to 50%, the behavior came back.
If it had been punishment blocking every other time would have reduced the behavior to zero.
So we have this study coming out shortly in the journal of applied behavior analysis.
The last thing I'll mention because of our time limitations is with care provider training.
What we typically use is behavioral skills training.
It's, we have found that it's not enough to get intervention going to simply tell someone what to do, or to give them a written instruction of what to do even the most highly educated, and well-trained individuals have difficulty doing that.
So we follow this method where we first discuss the intervention or procedure.
We model it and ask if there are any questions that the individual has.
We role play, we demonstrate the procedure to them, with them serving as the client and then reversing roles.
Then we go in sit you and provide them immediate feedback.
And then finally, when they're performing with 100% accuracy or close to that, we do delayed feedback where we watch them for an interval of time and only give them feedback after say a 10 minute observation or a 15 minute observation.
So the point here is that until you see the person doing it effectively, you can't assume that the intervention, will translate into the home or classroom or residential environment.
Here was a case where, this was the little girl actually with Rett syndrome that I talked about earlier.
We use behavioral skills training, to teach the intervention to her mother and to her sister.
And so, where we did the behavioral skills training was right after the baseline.
We went through that whole procedure that I just talked about, and only under those circumstances, did the mouthing decrease in the presence of her mother and her appropriate play skills increased.
And then it was maintained at low levels at one in one month, in five months followup.
And then also we trained her 12 year old sister to engage in the procedures and with the exception of one high session of marauding, where we did retraining of her sister after that, she maintained low levels of the mouthing.
So the point of this is you don't have to be a behavior analyst to do these procedures, as long as the training has occurred, to teach those, intervention skills.
Sorry to rush here at the end, but I know you all have places to go.
I just wanted to summarize that.
I definitely don't wanna come across as saying, like we've got the treatment for automatically reinforced behavior.
In fact, the whole purpose of this presentation is quite the opposite and that is to emphasize that when the behavior socially reinforced, the prognosis is good.
It's not so good.
You've got a rough road when you found that the behavior is automatically reinforced.
And I want to emphasize that all I need by automatic reinforcement is that it's the absence of social mediation with that reinforce the contingency.
So the typical intervention typically involves a progression from environmental enrichment, just providing alternative stimuli and the environment to differential reinforcement, which would be explicitly reinforcing contact with those alternative items and possibly mild punishment such as response blocking.
But if you use response blocking, make sure the effect on that particular individual behavior.
And then finally care providers should be trained to competency using the behavioral skills training method, such as what I just quickly described.
I think we have just a couple minutes left for any last minute questions.
- There are a couple of questions there.
I'll go back to one of the earlier questions that I captured.
One of the questions was if you could provide an example of schedule induced aggression for a school aged student, what would that look like for them possibly?
- Okay, well, so it's possible that what we typically view as behavior that's reinforced by getting access to a tangible.
So let's say the teacher says it's time to stop using your computers and then the child breaks the computer or lashes out and hits the teacher.
Currently, our assumption is, "oh, that's been reinforced by getting access "to the computer again." While it might just be that the source of reinforcement was cut off similar to Azrin's studies, where when the reinforcement is cut off the organism then aggresses toward another organism.
So in that case, what I was suggesting is we don't know enough yet to rule out the possibility that that's not just some kind of elicited response where those of us who are typically developing that gets shaped out of our repertoire because it's socially inappropriate, but individuals with intellectual disabilities or autism spectrum disorders may not be sensitive to the same types of social contingencies and that sort of elicited aggression might occur.
And I caution that that's speculative, but I wanted to throw it in there to say we have not necessarily ruled that out and there is some animal research to support it.
So I think we need to do more research on that for now.
- Thanks so much we really appreciate your time this afternoon, and maybe we'll have an opportunity to follow up with you again in our next webinar series.
- That would be great.
Take care.
- All right, thanks Dr. Vollmer, take care.