Teaching Advanced Verbal Behavior Using Speech- Generating Devices

Alice Shillingsburg, BCBA-D Author

04/20/2022 Added

138 Plays

Description

Tri State Webinar 2022

Searchable Transcript

Search:

[00:00:00.734]Hello everyone.
[00:00:02.110]My name is Alice Shillingsburg
[00:00:03.580]and today I'll be presenting a talk
[00:00:05.509]on teaching advanced verbal behavior
[00:00:07.880]to children with autism using speech generating devices.
[00:00:13.200]So, as most of us are well aware,
[00:00:14.800]autism is a developmental disorder characterized by deficits
[00:00:18.260]in both social communication and the presence
[00:00:21.390]of restricted and repetitive behaviors.
[00:00:24.820]It's also widely recognized now that the social deficits
[00:00:29.030]are really the hallmark feature of the disorder
[00:00:32.860]and that language deficits are no longer required
[00:00:36.160]for the diagnosis to be made.
[00:00:38.090]However, we do certainly still see deficits in language
[00:00:42.090]and language use and these deficits range
[00:00:45.870]from mild to severe.
[00:00:49.050]Studies have found approximately 65% to 75% of children
[00:00:53.490]with autism exhibit moderate to severe language delays.
[00:00:58.051]In a study by Anderson et al. in 2007,
[00:01:01.240]a sample of children diagnosed with autism at age two
[00:01:04.870]were then followed and assessed
[00:01:06.470]for their expressive language level at age nine.
[00:01:08.827]And the majority of the participants in this study
[00:01:11.360]did not have fluent language skills
[00:01:13.430]and almost 30% were not using spoken words consistently
[00:01:18.000]to communicate.
[00:01:22.090]So for children with more significant difficulties
[00:01:24.480]with spoken communication,
[00:01:26.110]augmentative and alternative communication systems
[00:01:28.740]are often successfully used.
[00:01:32.767]Augmentative and alternative communication systems
[00:01:37.610]provide a means for communicating when speech is delayed.
[00:01:41.190]And despite a common concern expressed by parents,
[00:01:44.271]there's no evidence to suggest
[00:01:46.330]that augmentative and alternative communication systems
[00:01:49.830]or AAC, will impede the development of spoken language.
[00:01:55.940]AAC includes both aided and unaided systems.
[00:02:01.600]Sign language is an example of an unaided system.
[00:02:05.230]Whereas picture exchange communication
[00:02:07.470]is an example of an aided system.
[00:02:10.400]There are also low tech and high tech AAC.
[00:02:14.313]A picture based system consisting of laminated icons,
[00:02:18.670]and a Velcro book for example,
[00:02:20.300]is an example of a low tech system,
[00:02:23.650]whereas a tablet or an iPad with an application on it,
[00:02:30.030]perhaps with an automated digitized speech output
[00:02:33.470]would be an example of a high tech option.
[00:02:39.170]High tech speech generating devices,
[00:02:41.820]or SGDs may be dedicated devices,
[00:02:45.560]such as a tablet or computer
[00:02:48.080]that are produced really exclusively
[00:02:49.990]for the use as a communication device,
[00:02:53.310]or they may be a commercially available tablet
[00:02:56.440]that's just equipped
[00:02:57.560]with some aftermarket speech generating apps,
[00:03:01.160]such as Proloquo2Go, TouchChat,
[00:03:04.740]any number of applications.
[00:03:07.610]Most SGDs offer touch screens
[00:03:10.810]and have what are called dynamic screen displays,
[00:03:14.050]meaning that the user's selection of one button
[00:03:17.200]causes the display to change
[00:03:19.241]and then they navigate to another screen
[00:03:22.580]with additional folder options.
[00:03:24.500]And this can facilitate faster communication
[00:03:27.130]and allows for a really a very large number of icons
[00:03:31.060]to be provided in one system.
[00:03:38.020]High tech speech generating devices on handheld tablets
[00:03:41.410]have been described as having some advantages
[00:03:44.300]over more low tech systems.
[00:03:47.530]Mobile devices are ubiquitous in society now
[00:03:51.610]and are relatively low cost.
[00:03:54.240]The systems are easy to modify and upload new icons
[00:03:58.370]that can take place very quickly.
[00:04:00.590]And they're pretty easy to carry and transport.
[00:04:03.330]I would guess pretty much everyone here today
[00:04:06.161]has a high tech device on them somewhere.
[00:04:10.840]And because of some of these advantages,
[00:04:15.370]some have started to really wonder
[00:04:18.370]if the widespread use of these devices
[00:04:20.980]as speech generating devices with autistic individuals
[00:04:26.560]has really gotten ahead of the research.
[00:04:33.570]The majority of research on the use of AAC
[00:04:37.170]with the autism population
[00:04:39.130]focuses on teaching requesting skills
[00:04:42.360]or what behavior analysts call mands.
[00:04:45.930]This holds true when examining
[00:04:47.820]high tech speech generating devices as well.
[00:04:50.960]In a recent meta-analysis of tablet use
[00:04:53.650]to teach communication skills to children with autism,
[00:04:56.910]the authors examined only 15 studies
[00:05:00.530]and the majority of these studies
[00:05:02.220]focused on simple manding or requesting,
[00:05:05.446]using a static icon display,
[00:05:08.320]as opposed to a dynamic display.
[00:05:11.340]A few studies also included single word tax,
[00:05:15.260]or what we call labeling, greetings,
[00:05:17.890]some social niceties, such as please and thank you.
[00:05:20.410]But for the majority of these studies,
[00:05:23.900]the skill that's being taught is requesting.
[00:05:27.580]And all but one study targeted single step communication.
[00:05:31.860]And this means that there's no navigation through screens
[00:05:34.850]on the devices at all.
[00:05:36.480]And instead, the participant in the studies
[00:05:38.520]simply had to select an icon
[00:05:40.290]from a display on a single screen.
[00:05:44.280]So, although most apps have dynamic screen displays,
[00:05:48.040]allowing for navigation to thousands of words,
[00:05:51.440]much of the research doesn't tell us how to teach this
[00:05:54.430]or how to move on to more complex communication skills
[00:05:57.063]over and above straightforward requests,
[00:06:01.390]particularly with the autism population.
[00:06:06.380]So in today's presentation, I'll present a set of studies,
[00:06:11.160]teaching advanced communication skills
[00:06:13.920]to children with autism
[00:06:15.610]who use high tech speech generating devices.
[00:06:19.710]And all three of these studies are actually replications
[00:06:23.050]of previous studies that were conducted
[00:06:25.030]with vocal participants, with autism.
[00:06:27.810]People who communicate predominantly
[00:06:29.980]with their vocal language.
[00:06:31.780]And we're gonna cover a study to teach mands for information
[00:06:36.710]procedures to teach individuals
[00:06:39.270]to report past experiences or past behaviors
[00:06:42.700]and labeling, expressive labeling
[00:06:45.329]or tacting using noun-verb combinations.
[00:06:48.080]So two word utterances using their speech generating device.
[00:06:54.420]Okay, so let's get started with the first set of studies.
[00:06:58.670]During typical child development,
[00:07:00.320]early mands include requests for preferred toys,
[00:07:03.940]snacks, or people.
[00:07:05.900]This repertoire becomes increasingly complex
[00:07:08.940]by beginning to include mands for items
[00:07:11.550]that maybe aren't present,
[00:07:12.780]mands for things in the future,
[00:07:15.180]such as when are we going to the playground,
[00:07:17.870]and mands for information.
[00:07:20.080]Although neurotypically developing children
[00:07:22.070]appear to learn to ask questions fairly effortlessly,
[00:07:25.810]some children with autism who also experience
[00:07:28.470]significant language and communication delays
[00:07:31.280]may require some specialized intervention
[00:07:33.650]to learn to mand or request for information from others.
[00:07:40.380]So a widely recognized key element
[00:07:43.000]of effective mand training in general
[00:07:45.880]is manipulating what are called motivating operations.
[00:07:49.430]Now, for those of you attending this presentation
[00:07:52.970]who maybe aren't behavior analysts,
[00:07:54.520]you can easily substitute the word motivation
[00:07:57.300]for the term motivating operations.
[00:08:01.330]But before we begin for the behavior analysts
[00:08:03.580]in the audience, as well as everyone else,
[00:08:07.810]we'll talk a little bit about a quick overview
[00:08:09.980]of how motivating operations or motivation
[00:08:13.630]is involved in mand training or request training.
[00:08:19.420]Okay, so quickly a definition of motivating operations,
[00:08:24.370]motivating operations really have two components to them.
[00:08:28.160]The first is that when a motivating operation is present
[00:08:31.880]or a motivation,
[00:08:33.150]there is a change in the reinforcing effectiveness
[00:08:36.100]of other stimuli.
[00:08:37.360]In other words,
[00:08:38.280]other stimuli the environment can become a reinforcer
[00:08:41.870]or lose that effect.
[00:08:43.930]And when that happens,
[00:08:45.390]the second part of a motivating operation
[00:08:47.560]is that we see a change in the frequency
[00:08:49.630]of the occurrence of behaviors
[00:08:51.110]associated with those reinforcers.
[00:08:53.500]In other words, if a reinforcer becomes more effective,
[00:08:56.440]it might evoke behaviors.
[00:08:57.860]And if a reinforcer loses its effectiveness
[00:09:00.392]then those behaviors might abolish or abate.
[00:09:06.730]Okay, let's look at an example.
[00:09:09.280]So in the behavior analytic literature,
[00:09:11.550]we talk about motivating operations in two ways,
[00:09:15.200]it can be an establishing operation,
[00:09:17.020]which is really the establishment of motivation.
[00:09:21.220]And when this happens,
[00:09:22.550]we see a change in the value
[00:09:24.120]of something in the environment.
[00:09:25.620]And when something becomes more valuable,
[00:09:27.450]we see a change in behavior.
[00:09:29.190]So for example,
[00:09:31.110]let's say there's been a period of time without food,
[00:09:34.530]food deprivation sounds a little bit harsh,
[00:09:36.600]but essentially it just means it's been a little while
[00:09:39.940]since you've had a snack.
[00:09:42.020]The value of that food,
[00:09:44.880]any kind of snack then would increase
[00:09:46.800]and it would increase any behaviors that lead to food.
[00:09:51.110]Alternatively, what might happen in the environment
[00:09:53.720]is that you have what's called an abolishing operation,
[00:09:56.720]where the value of something actually decreases
[00:10:01.300]so there's no motivation.
[00:10:02.580]And this might happen for example, when you're satiated.
[00:10:05.960]So instead of it being a long time
[00:10:07.570]since you've had something to eat,
[00:10:08.890]you're actually full or satiated.
[00:10:11.070]And when this happens,
[00:10:12.380]then the value of food actually decreases.
[00:10:15.270]And it makes it a little less likely
[00:10:17.360]that you're gonna engage in any behavior
[00:10:19.280]that might lead to food.
[00:10:20.990]So, a little bit more succinct here.
[00:10:23.980]Let's say it's been five hours since breakfast.
[00:10:26.680]Food now becomes valuable
[00:10:28.130]and you might engage in any of these types of behaviors
[00:10:31.050]in order to contact food.
[00:10:32.930]You might get up and walk over to the fridge.
[00:10:35.180]You might look up a menu on your phone,
[00:10:37.440]or you might ask somebody for a snack.
[00:10:40.310]Let's say you just finished a giant buffet lunch.
[00:10:44.190]In this case,
[00:10:45.023]there would be no motivation for additional snacks or food.
[00:10:49.430]So from a motivating operation perspective,
[00:10:52.431]there's a decrease in the value of that food.
[00:10:55.720]So the food loses value.
[00:10:57.250]And so you don't engage in behavior that leads to food.
[00:11:00.630]Instead, you might take a nap, watch a football game,
[00:11:03.170]and you certainly aren't gonna ask somebody for a snack.
[00:11:06.510]Okay, so in this example,
[00:11:08.630]the change in behaviors that we see
[00:11:10.440]only a couple of them are actually relevant
[00:11:12.520]to manding or requesting.
[00:11:15.720]When you're motivated
[00:11:16.880]only one of those three behaviors is a mand.
[00:11:19.520]The other two, you just get up and get it yourself.
[00:11:23.200]But in this case you might ask for a snack.
[00:11:25.940]When there's no motivation,
[00:11:28.390]what we would hope to see
[00:11:29.960]is that you wouldn't ask for a snack.
[00:11:34.890]So, when we set up intervention
[00:11:38.200]to teach requesting skills or manding skills
[00:11:43.170]to individuals who haven't acquired this skill,
[00:11:47.030]it might look something like this.
[00:11:48.860]So, we try to establish motivation
[00:11:53.630]for something that we're gonna teach a child to ask for.
[00:11:56.740]So for example, it's been a little while since we've eaten,
[00:11:59.910]it's now snack time.
[00:12:01.520]And at this point there's an increased value of snack.
[00:12:05.000]What we might see in individuals who have developed language
[00:12:07.920]is they might say chip or chip please,
[00:12:09.740]or I'd like some chips.
[00:12:11.030]And the reinforcer for that behavior is of course,
[00:12:13.410]to give somebody their chips for their snack.
[00:12:16.240]Now, with kiddos who haven't developed
[00:12:18.050]the ability to ask for things,
[00:12:19.810]to mand for things when they want it, the skill is absent,
[00:12:24.470]our intervention is gonna require
[00:12:26.270]that we do something right here.
[00:12:28.410]And in the case of behavior analytic, ABA services,
[00:12:32.300]what this might look like
[00:12:33.310]is we're actually gonna prompt them to mand.
[00:12:36.470]Now, with kiddos who use spoken language,
[00:12:38.470]we might prompt, say chip,
[00:12:41.180]with kids who use AAC or speech generating devices,
[00:12:45.130]we would prompt them to press the button on their device
[00:12:49.900]for that to give the output of chip.
[00:12:53.170]When that happens then we would reinforce with chips.
[00:12:56.250]Okay, so that's simple manding.
[00:12:57.930]Today we're gonna talk a little bit more
[00:12:59.590]about more advanced manding,
[00:13:01.630]or what's called manding for information,
[00:13:03.530]or requesting for information.
[00:13:05.340]And the setup is actually pretty similar.
[00:13:08.440]So imagine a child asks for something that he can't find,
[00:13:12.370]maybe he's doing an art project
[00:13:13.690]and he can't find his crayons
[00:13:15.130]or whatever it may be.
[00:13:17.201]He's told by the adult in the room,
[00:13:20.840]oh, they're in the cabinet,
[00:13:22.280]but potentially not told which specific cabinet.
[00:13:25.450]And so in this case,
[00:13:26.950]the child needs more information
[00:13:29.030]in order to get the thing that he wants.
[00:13:31.870]So, just like in the previous example,
[00:13:34.470]we've set up a motivation or an establishing operation.
[00:13:38.430]And this is that we've told him some information,
[00:13:41.890]but not enough information
[00:13:43.980]to access the thing that he's looking for.
[00:13:46.410]And so the motivation really is related to the fact
[00:13:49.810]that information is withheld or not known.
[00:13:52.790]And when that happens,
[00:13:54.040]that increases the value of knowing that information
[00:13:57.760]with an individual who has an intact
[00:14:00.077]requesting for information
[00:14:02.040]or manding for information repertoire,
[00:14:03.820]they may say, which cabinet or which one?
[00:14:06.950]And when that happens, the information,
[00:14:09.110]oh, the cabinet on the left or cabinet number two,
[00:14:12.840]might be provided,
[00:14:13.930]and the person can then use that in information
[00:14:16.240]to access the reinforcer that they were looking for
[00:14:19.090]in the first place.
[00:14:20.310]Now, again, with kiddos
[00:14:21.510]who don't have a manding for information repertoire,
[00:14:24.550]we've got to intervene.
[00:14:26.130]And similar to simple mand training,
[00:14:29.600]we're gonna do something right here.
[00:14:31.510]And that is, we're gonna prompt the mand to ask which one,
[00:14:35.850]and when that happens,
[00:14:37.168]we're essentially providing the information.
[00:14:40.970]They say which one, we provide the information,
[00:14:43.220]that's the reinforcer.
[00:14:44.690]And then they use that information
[00:14:46.590]to access that ultimate reinforcer.
[00:14:51.400]Now, just like with the simple mand training,
[00:14:56.080]there are situations
[00:14:56.913]where we don't want kids to ask for information
[00:15:00.090]because they've already got it.
[00:15:01.450]So in this case,
[00:15:02.340]a child might ask for something he can't find.
[00:15:04.860]And instead of being given sort of vague information,
[00:15:08.640]he's told the very specific cabinet where or the item is.
[00:15:12.160]In this situation,
[00:15:13.170]we've set up what's called an abolishing operation
[00:15:15.550]in the sense that we've already said
[00:15:17.010]it's in cabinet number two.
[00:15:18.600]And if that's the case, the information is already provided.
[00:15:22.400]There's no increase in the value of more information
[00:15:25.490]because we don't need any more information.
[00:15:27.940]And in this case, no mand is admitted, no request.
[00:15:31.300]There's no information additionally provided
[00:15:33.620]because we don't need it.
[00:15:34.900]And instead,
[00:15:36.910]the individual can just use the information
[00:15:38.940]that was already provided to access the item.
[00:15:43.150]Now functional mand training
[00:15:45.290]requires that a person be able to discriminate
[00:15:48.490]between those EO and AO conditions.
[00:15:51.410]When inform is needed,
[00:15:52.700]versus when information is not needed.
[00:15:55.470]And that's why when we teach mands,
[00:15:57.730]we typically conduct practice opportunities or trials
[00:16:03.890]where sometimes the information is needed
[00:16:06.200]and sometimes the information is not needed.
[00:16:09.090]Because if you see manding or requesting occur
[00:16:13.100]in that AO condition,
[00:16:14.880]when information is not needed,
[00:16:16.960]that's a signal to you
[00:16:18.500]that you haven't taught a discriminated mand.
[00:16:20.920]And the mand might actually be more of a rote response.
[00:16:25.120]So this is a really good opportunity
[00:16:26.940]to sort of check for that and program for that
[00:16:29.330]by alternating between EO and AO conditions,
[00:16:32.680]to make sure your mands or your requests are functional.
[00:16:37.230]Okay, so this is a study that I published in 2014
[00:16:42.310]with a couple of co-authors of mine,
[00:16:44.420]teaching mands, organ request for information,
[00:16:48.445]specifically teaching kiddos with autism
[00:16:51.540]to ask who and which questions.
[00:16:54.280]And we made sure that we alternated
[00:16:56.380]between establishing and abolishing operations.
[00:16:59.520]So when you need the information
[00:17:01.060]versus when you don't need the information.
[00:17:04.670]So just walk you through this.
[00:17:06.343]This study was actually published
[00:17:07.970]with kiddos who use spoken communication,
[00:17:10.000]and then I'll walk you through how we did it
[00:17:11.850]with kids using speech generating devices.
[00:17:14.600]So again,
[00:17:15.433]we created or contrived this sort of motivation
[00:17:18.800]and establishing operation
[00:17:20.740]and abolishing operation conditions.
[00:17:23.340]In the EO present condition,
[00:17:25.670]the information regarding the location of a preferred item
[00:17:29.250]was not given to the child.
[00:17:31.410]So we kind of gave some vague information.
[00:17:33.677]And this was how we were contriving motivation
[00:17:36.570]for that information,
[00:17:37.650]so that they would want to ask us questions.
[00:17:42.180]In the EO absent condition, this is abolishing operation,
[00:17:46.220]the information regarding the location of the thing
[00:17:49.080]that they were looking for was just given straight away.
[00:17:51.950]And so now there's no motivation for additional information.
[00:17:56.290]And what we were looking to measure are dependent variables,
[00:17:59.290]where we wanted to see did they ask who has it,
[00:18:02.790]or which one, when that EO was present,
[00:18:06.650]when the information was needed.
[00:18:08.170]But we also measured to make sure that they never asked
[00:18:11.290]when the motivation was absent,
[00:18:12.890]so that we could ensure we were not teaching rote responses.
[00:18:17.920]So it looked a little bit like this,
[00:18:19.976]in the EO present condition,
[00:18:21.840]we hid some preferred items in a container,
[00:18:26.290]and the containers were sort of a set of them.
[00:18:29.220]So there were similar containers and we didn't specify
[00:18:32.580]which container the item was hidden in.
[00:18:34.650]So if a kiddo was looking for a crayon
[00:18:36.480]or wanted a snack, something like that.
[00:18:41.150]In our AO condition,
[00:18:43.164]or when the motivation was gonna be absent,
[00:18:46.200]we hid a preferred item in the container, again,
[00:18:48.920]amongst similar containers.
[00:18:50.700]But in this case,
[00:18:51.533]we did tell them which container it was in.
[00:18:54.860]So the child's asked for a cookie, we say, sure,
[00:18:57.970]it's in one of those boxes,
[00:18:59.650]contriving the motivation to learn which box it's in.
[00:19:04.220]And then other condition, the child asks for a cookie
[00:19:07.270]and we say, sure, it's in the yellow box.
[00:19:09.560]And this indicates that there's or sets the stage
[00:19:13.610]for not needing any additional information.
[00:19:17.170]Okay, so let's look at some data from this study.
[00:19:20.230]Now these data are not percent correct.
[00:19:26.500]I know sometimes when we look at data for teaching skills,
[00:19:29.740]we're looking at, did he get 90% correct
[00:19:33.917]and that type of thing.
[00:19:34.980]But instead these data are presented cumulatively.
[00:19:39.320]What that means is that each one of these dots here
[00:19:46.210]represents an opportunity to request.
[00:19:48.820]It's just each single one is an opportunity.
[00:19:52.240]And if you see the lines start to go up,
[00:19:54.080]that means that the child engaged in the behavior
[00:19:57.020]we were looking for.
[00:19:58.540]And if it stays flat, that means that they didn't.
[00:20:02.970]Okay, so what we're seeing here
[00:20:05.170]are mands for information or requests asking which,
[00:20:10.660]so which cabinet, which box, which drawer,
[00:20:14.080]that type of thing, on this side.
[00:20:16.410]And over here is when I deliver that information.
[00:20:20.300]So I say, oh, it's in the yellow box.
[00:20:23.290]Does the child approach that box
[00:20:26.500]and access the thing that they were looking for?
[00:20:28.950]So basically this is that use of information.
[00:20:31.430]Do they use the information that they're told?
[00:20:34.400]Okay, so the black diamonds represent the EO condition.
[00:20:39.250]Now keep in mind that's when the information is needed.
[00:20:42.010]We want kids to ask us for information in that condition.
[00:20:46.320]The AO condition is when the information
[00:20:48.830]was already provided,
[00:20:50.130]we want this line to stay low
[00:20:53.000]because we've already told them the information.
[00:20:55.190]So we don't want them to ask.
[00:20:57.580]What you can see
[00:20:58.413]for all three of our participants with autism,
[00:21:00.640]that in baseline,
[00:21:01.559]whether we told them the information or not,
[00:21:04.420]they never asked which questions.
[00:21:08.050]And you can see that we're alternating
[00:21:10.570]between the two trial types and they never ask.
[00:21:13.940]Over here though,
[00:21:15.290]what you'll see is that
[00:21:16.820]when we don't provide them the information in these trials,
[00:21:21.460]they can't use the information to go find it
[00:21:24.630]in the yellow box.
[00:21:26.020]However, in these trials, right over here,
[00:21:29.940]we are providing the information right out of the gates.
[00:21:32.940]And they're able to use that information.
[00:21:37.190]So we know now that if this child
[00:21:40.130]could just ask for information, they could use it
[00:21:43.415]to find things better in their environment.
[00:21:47.510]Okay.
[00:21:48.520]So then we ran treatment trials,
[00:21:52.220]teaching them how to ask for information.
[00:21:55.300]And what you see here is,
[00:21:57.800]as we've taught contriving opportunities
[00:22:01.310]for them to ask for information
[00:22:02.870]and prompting them to request that information,
[00:22:05.560]we see all three of our participants ask for information
[00:22:09.240]only when the information is withheld
[00:22:11.730]and develop a functional manding repertoire,
[00:22:14.960]because they aren't asking when the information is withheld.
[00:22:19.420]Okay.
[00:22:20.253]And then you can also see over here
[00:22:22.290]now that these kids can use their language skills
[00:22:27.300]to ask for information,
[00:22:28.790]they're able to use information in both conditions
[00:22:31.570]to then go and find the item that they were looking for.
[00:22:35.230]Okay, so those data were with three children with autism
[00:22:39.960]who were vocal.
[00:22:41.020]We wanted to see if kids
[00:22:42.400]who were using speech generating devices
[00:22:44.350]could also mand for information under similar conditions.
[00:22:49.260]This study was published in 2019
[00:22:52.200]in the Journal of Applied Behavior Analysis.
[00:22:54.510]And we've essentially just replicated these procedures.
[00:22:58.240]So here are just some information about the participants.
[00:23:01.800]I typically get asked, okay, for which kiddos?
[00:23:04.561]So you can see here all of our participants
[00:23:07.010]were between the ages of three and six years old.
[00:23:09.550]If you're familiar with the VB-MAPP,
[00:23:11.540]or the verbal behavior milestones assessment,
[00:23:14.060]you can see that all of these kiddos
[00:23:16.860]were scoring in the upper level two and three
[00:23:19.250]of the VB-MAPP.
[00:23:22.010]They had really high mand scores,
[00:23:25.130]which means that they learned new mands
[00:23:27.220]without a lot of work.
[00:23:28.340]Now these are mands for items.
[00:23:30.950]But none of these kids were manding for information
[00:23:33.490]when we started this study.
[00:23:36.970]Okay, so here's how these mand for information trials
[00:23:40.490]would go for our kids using speech generating devices.
[00:23:44.160]This is just an example of the home screen
[00:23:47.430]that a child could initiate a mand for a preferred item.
[00:23:53.090]So the child would click on here.
[00:23:55.513]These are their preferred items that are in our clinic.
[00:23:59.160]And so they would go to that location.
[00:24:01.157]Then they might click on the folder labeled favorite things.
[00:24:04.820]When they got in there,
[00:24:06.010]then the child would mand a request
[00:24:08.560]by selecting the icon of that preferred item.
[00:24:11.440]Now they've asked for a specific item
[00:24:13.310]and the response here is, sure, one of your teachers has it.
[00:24:16.550]Now this is information, but it's vague.
[00:24:19.520]And so this is setting up the situation
[00:24:23.520]where a child actually needs more information.
[00:24:26.240]And so that sets the stage for manding for information.
[00:24:30.560]The child would then go back out to their home screen,
[00:24:34.100]select the questions folder,
[00:24:35.950]and then select the correct question out of an array.
[00:24:39.550]So in this case, it would be who has it?
[00:24:42.290]Asking about the particular teacher
[00:24:44.990]who has their preferred item.
[00:24:48.600]One participant,
[00:24:50.400]there was no navigation because he simply typed.
[00:24:55.180]So here are the data from our recently,
[00:24:57.040]this published study with the three participants
[00:25:00.530]who use speech generating devices.
[00:25:02.940]So here we taught mands for information using which.
[00:25:06.630]Now the data should look similar to you.
[00:25:08.990]It's presented the same way as the previous data set
[00:25:11.870]that we just went through.
[00:25:13.410]As you can see, none of the participants
[00:25:15.900]were manding for information or asking WH questions,
[00:25:22.350]whether they needed the information or not in baseline.
[00:25:25.720]But similar to the other kiddos,
[00:25:27.980]if we provided the information in that AO condition,
[00:25:31.960]all of them could use the information
[00:25:34.220]to access the thing that they were looking for.
[00:25:36.420]And that's critical to know
[00:25:39.090]since use of the information really is that reinforcer.
[00:25:44.020]All of these participants also learned to discriminate
[00:25:48.207]when they needed information to ask questions.
[00:25:51.200]And when they knew the information
[00:25:53.250]they refrained from emitting like a rote response,
[00:25:56.650]which is really also important to know.
[00:25:58.800]And then additionally,
[00:26:00.250]just like the kiddos who use vocal language,
[00:26:03.720]these kiddos, once they learned to request information,
[00:26:07.340]could then use that information
[00:26:09.170]to access the thing that they couldn't find originally.
[00:26:16.420]This is just the data for manding for information who,
[00:26:20.620]so we taught both which questions and who questions,
[00:26:26.020]similar results here.
[00:26:29.040]Now, we also wanted to make sure
[00:26:31.220]not only would the kiddos
[00:26:33.850]be able to functionally use their new language skill
[00:26:37.010]to ask which questions when they needed to know which,
[00:26:41.380]and who questions when they needed to know who
[00:26:44.070]and to refrain from using those questions
[00:26:46.520]when they didn't need to use them.
[00:26:48.360]We also then interspersed them together
[00:26:52.530]to make sure that they didn't accidentally
[00:26:54.950]ask which questions when they meant who,
[00:26:57.220]and who questions when they meant which.
[00:26:59.850]And this is an important thing to assess as well.
[00:27:04.380]So what you can see here is that every time,
[00:27:06.840]this is Justin had no problem
[00:27:10.280]selecting the correct question to fit the right scenario
[00:27:15.230]and never made mistakes here, neither did Emma.
[00:27:18.900]However, one of our participants,
[00:27:20.790]once we sort of started interspersing
[00:27:23.460]the different types of WH questions, did have some errors.
[00:27:29.110]Now, he didn't ask questions when he didn't need them,
[00:27:31.950]but he was using who and which kind of interchangeably.
[00:27:34.990]So we then taught that to kind of straighten that out
[00:27:37.730]a little bit for him.
[00:27:38.870]And that's important to know for any learner
[00:27:41.430]that you might be working on this skill with.
[00:27:46.230]So in conclusion,
[00:27:47.330]all three participants engaged in discriminated manding.
[00:27:50.190]They maned for information when it was needed,
[00:27:52.200]refrained from manding for information
[00:27:53.900]when it wasn't needed.
[00:27:54.930]And they used the appropriate mand frame, who or which,
[00:27:58.860]under the right conditions with the one little guy
[00:28:02.010]needing a little bit of training to help him with that.
[00:28:06.970]So let's move on to another series of studies
[00:28:11.980]to address another commonly identified parental concern.
[00:28:15.360]And that is teaching kids to be able to answer questions,
[00:28:20.370]or to describe events and things that have happened to them.
[00:28:24.960]And we've examined this through teaching kids
[00:28:27.410]to talk about the events in their day.
[00:28:29.800]Now in day to day interactions,
[00:28:31.720]children are frequently expected to report or recall
[00:28:34.300]their past behavior accurately.
[00:28:35.970]For example, that adult might ask,
[00:28:38.670]did you finish your homework?
[00:28:40.160]Or who did you see at school today?
[00:28:42.310]And having difficulty answering questions like this
[00:28:45.340]is really a common caregiver concern
[00:28:48.000]and can become more concerning when it's things such as,
[00:28:51.900]how did you get this bruise or did you get hurt today?
[00:28:56.030]Now the development of this repertoire
[00:28:58.280]could be conceptualized as what we call a self tact
[00:29:01.280]or a self label.
[00:29:02.880]Basically self tacting or self labeling is verbal behavior
[00:29:06.960]that's controlled by a current stimuli,
[00:29:09.440]including things that are going on
[00:29:11.420]inside the speaker himself
[00:29:14.630]that are sort of generated by the question
[00:29:16.580]that they've just been asked.
[00:29:18.470]And what you've been asked,
[00:29:20.470]what you're thinking about or feeling about on the inside,
[00:29:23.350]in combination with what has been conditioned
[00:29:26.690]through reinforcement really comes together
[00:29:30.215]to develop this skill.
[00:29:32.970]Now, Palmer describes this as,
[00:29:35.660]what we call intraverbal control.
[00:29:37.730]So for the behavior analysts in the audience
[00:29:40.760]that this might sound familiar,
[00:29:42.250]but essentially the difference is,
[00:29:45.920]let's say, I say Mary had a little.
[00:29:52.270]And if you're listening to this, you probably said lamb.
[00:29:55.570]That's an example of kind of answering a question
[00:29:59.110]or filling in something that's pretty rote, right?
[00:30:01.490]It's an intraverbal but it's pretty rote.
[00:30:04.030]Versus something like,
[00:30:05.910]tell me something, tell me what makes a good pet.
[00:30:09.880]And probably many of you might be saying dog, cat,
[00:30:12.750]bird, fish.
[00:30:14.220]And the reason we get different answers to that question
[00:30:17.630]is because your answer is evoked
[00:30:19.540]not only by the question that I asked
[00:30:21.550]or the words that I said,
[00:30:22.750]but also your thoughts and feelings about that,
[00:30:25.390]which are things that are going on inside of you.
[00:30:28.660]Now let's look at an example
[00:30:30.540]of reporting past behavior or events in children.
[00:30:34.430]So a caregiver may ask a question, oh,
[00:30:36.728]what did you eat for snack today?
[00:30:39.020]And essentially,
[00:30:42.454]what happens here is the child might think, hmm,
[00:30:47.570]say to himself, what did I eat for snack?
[00:30:49.420]And think back and maybe visually imagine
[00:30:52.950]the snack that they had that day.
[00:30:56.420]And then says gummy bears.
[00:30:59.070]And then the caregiver responds, oh, yummy gummy bears,
[00:31:03.360]to sort of reinforce that response.
[00:31:06.490]But what happens if a child doesn't ask, I'm sorry,
[00:31:11.100]answer correctly and just gives a rote answer?
[00:31:13.360]Maybe they just hear the word snack and say any snack,
[00:31:16.770]as opposed to thinking about their own personal experiences.
[00:31:22.130]Now, what typically happens in child development
[00:31:26.010]is that our verbal community really helps us learn
[00:31:29.950]how to respond correctly to questions.
[00:31:32.260]Essentially, if you may ask a question,
[00:31:37.700]who did you see at school today?
[00:31:39.240]And let's say that that question
[00:31:40.420]is a little too complicated,
[00:31:42.543]perhaps the child can't answer that question.
[00:31:45.150]We would swoop in with a little bit of help
[00:31:47.100]and say, was Jessica there?
[00:31:49.260]Now that might be a question that a child can answer
[00:31:51.990]and you can see through lots of interactions like this,
[00:31:55.510]that a child might learn that this more complicated question
[00:31:59.500]is actually referring to people,
[00:32:01.640]because you see the who, and then was Jessica.
[00:32:03.980]And over time,
[00:32:06.230]kids learn to be able to answer to more complex questions.
[00:32:09.680]And this is how reporting past behaviors
[00:32:11.660]is shaped in typical development.
[00:32:16.210]Now deficits in the development of reporting past events
[00:32:19.500]or behaviors may happen for a couple of reasons
[00:32:22.640]for kiddos with autism.
[00:32:25.140]One is just errors in stimulus control,
[00:32:28.810]meaning that the question is potentially too complicated.
[00:32:32.650]And the child just hasn't learned
[00:32:35.938]to be able to respond to questions like that.
[00:32:40.497]The second reason might be
[00:32:43.090]that a lot of these types of behaviors
[00:32:45.380]are sort of reinforced or developed
[00:32:48.780]through kind of natural social reinforcers.
[00:32:52.257]And if social stimuli aren't potent enough
[00:32:56.980]to really reinforce particular skills,
[00:33:01.310]then this may be more challenging for children with autism.
[00:33:05.753]So in order to overcome these deficits,
[00:33:08.430]interventions might focus on,
[00:33:10.420]specifically on improving the correspondence
[00:33:12.930]between non-verbal and verbal behavior.
[00:33:15.940]And this is known as do/say correspondence.
[00:33:20.100]So essentially the ability to accurately report
[00:33:25.570]or say what you just did.
[00:33:29.070]So in do/say correspondence,
[00:33:34.550]a non-verbal behavior, that's the due part,
[00:33:37.960]and an antecedent stimulus,
[00:33:40.050]so for example, saying, what did you eat for a snack,
[00:33:43.360]occurs, and then a verbal behavior is emitted.
[00:33:46.530]And that's the say part,
[00:33:48.090]cuz that's where we get sort of the do/say correspondence.
[00:33:50.520]Now, if we have correspondence
[00:33:52.130]in the response, I ate gummy bears,
[00:33:55.100]would then be reinforced.
[00:33:57.280]However, we have found that this strategy alone
[00:34:00.470]doesn't always work, just reinforcement.
[00:34:03.810]So, in two studies,
[00:34:07.320]we've looked at examining potential strategies
[00:34:10.040]to improve teaching kids to report past behavior.
[00:34:14.090]In one study, came out in 2017,
[00:34:17.270]we implemented the procedures with vocal children,
[00:34:20.090]and we've found that using vocal imitation prompts
[00:34:23.750]can be effective at teaching them
[00:34:25.950]to tell us what happened earlier in the day.
[00:34:28.220]And the study I wanna talk about today
[00:34:30.650]is a replication of this study that came out in 2019,
[00:34:34.120]where we used similar procedures
[00:34:36.160]to children who were not vocal,
[00:34:37.940]but who used high tech speech generating devices
[00:34:40.410]to communicate.
[00:34:43.380]So these kiddos were actually the same three kiddos
[00:34:45.820]that were in the last study that I talked about
[00:34:49.560]with manding for information.
[00:34:51.240]So you're familiar with their sort of skillset.
[00:34:54.360]They all use their device to mand, tact,
[00:34:56.640]and intraverbally respond.
[00:35:00.070]Now in this study, we scored a response as correct
[00:35:02.960]if the participant admitted the name of the activity
[00:35:06.110]that they completed in their specific location,
[00:35:09.610]using either a picture selection on their device
[00:35:12.540]or text selection,
[00:35:13.700]or for the one little guy who used typing on his device.
[00:35:20.070]So again, for the two kiddos who used icons,
[00:35:23.520]navigation was required,
[00:35:24.990]but for the kiddo who typed, he just typed on his screen,
[00:35:29.760]as you can see there on the bottom, something like that.
[00:35:33.290]Okay, so prior to starting this type of procedure,
[00:35:38.110]you're gonna potentially need to do some pre-assessment
[00:35:40.700]or pre-teaching to ensure that some prerequisite skills
[00:35:44.320]are in place.
[00:35:45.230]So, before we did baseline here,
[00:35:48.210]we taught the names of the activities
[00:35:50.750]and the locations where those activities
[00:35:53.270]were going to be done so that they could tell us about that.
[00:35:58.850]So basically we taught three different activities
[00:36:02.330]and three different locations
[00:36:04.620]that were gonna be used in this procedure.
[00:36:08.840]And in all of the conditions throughout this study,
[00:36:11.830]just so that you know,
[00:36:13.199]the order of the locations that we went
[00:36:16.070]and the different activities that we completed
[00:36:18.240]varied every single day.
[00:36:19.710]So there was nothing rote
[00:36:20.910]about any of the things that we did.
[00:36:23.300]So for example, on one day,
[00:36:25.420]a participant might have colored in the classroom,
[00:36:30.140]went bowling in the playroom and then read in the cafeteria.
[00:36:35.450]And then on the next day,
[00:36:36.980]he or she would've read a book in the playroom,
[00:36:39.690]colored in the cafeteria and so on.
[00:36:42.794]So that there was no way to sort of memorize
[00:36:45.190]the responses to the questions we were asking.
[00:36:48.710]Now, once we started the study,
[00:36:51.700]after they completed their activities,
[00:36:54.861]the kiddo would go about their regular daily routine.
[00:36:58.630]And then at the end of their day,
[00:37:00.630]which was about two hours later,
[00:37:02.590]the therapist would ask them,
[00:37:04.420]what did you do in the cafeteria?
[00:37:07.100]What did you do in the classroom?
[00:37:08.330]What did you do in the playroom?
[00:37:11.540]For all of the locations.
[00:37:13.630]Whoops, there we go.
[00:37:15.810]Now a correct response was reinforced.
[00:37:18.470]So we would give them a reinforcer to help teach them
[00:37:22.410]that these were correct responses.
[00:37:26.490]At the end of the day, what did you do?
[00:37:28.390]Sorry, I got a little behind there.
[00:37:30.420]Okay, now for one of our participants, Emma,
[00:37:33.010]she responded a little bit uniquely.
[00:37:34.650]So I'll show you her data a little bit separately.
[00:37:38.870]Now here,
[00:37:39.850]we're most interested in the open data path
[00:37:43.150]and the squares here,
[00:37:45.790]because this represents her responses at the end of the day,
[00:37:48.670]which is really the socially valid response, right?
[00:37:51.040]When a parent comes to pick you up at the end of the day,
[00:37:52.907]and you say, oh, what'd you do in the playroom today?
[00:37:56.320]Usually that's sort of after a pretty lengthy delay
[00:38:00.020]that you did that fun activity a couple hours earlier.
[00:38:06.070]So at the end of the day,
[00:38:07.040]each participant is asked about their activities
[00:38:09.070]completed in the three locations.
[00:38:10.880]So you can see here,
[00:38:12.100]they could have responded up to three times correctly.
[00:38:15.990]In baseline, Emma admitted no correct responses
[00:38:19.160]to past events.
[00:38:23.020]Okay, here we have the data for the other two participants,
[00:38:27.630]Justin and Bruce.
[00:38:28.720]Justin had low levels of correct responding in baseline
[00:38:32.270]while Bruce had zero.
[00:38:33.930]And for these two kiddos,
[00:38:35.880]we actually did have their caregiver
[00:38:37.490]do some probes for us as well
[00:38:39.720]to see how well it generalized to them.
[00:38:41.930]And you can see,
[00:38:42.763]they did not report past behaviors
[00:38:44.010]to their caregivers either using their devices.
[00:38:50.380]Now, the first phase that we implemented
[00:38:55.790]was really just the introduction of immediate probes
[00:38:58.570]instead of delayed probes.
[00:39:00.500]So this phase was identical to baseline,
[00:39:03.870]except that immediately after each activity was completed
[00:39:07.270]and cleaned up, we would say, hey,
[00:39:08.810]what did you just do in the, whatever?
[00:39:12.550]And one of the reasons we wanted to do that
[00:39:14.700]was we just needed to know
[00:39:16.140]could they tell us at all what they had just done?
[00:39:19.620]Or was it specifically the delay that was a challenge.
[00:39:23.880]So what did you do in the playroom.
[00:39:27.640]If it was,
[00:39:30.080]and then we would ask, do the next activity,
[00:39:32.520]ask them what did you do in the cafeteria?
[00:39:35.410]And then did the next activity,
[00:39:36.650]what did you do in, wherever location they were in.
[00:39:40.250]If they were correct, it was reinforced.
[00:39:43.100]If not, it was sort of neutral response.
[00:39:46.350]And then again,
[00:39:47.183]we would ask after an hour and a half,
[00:39:49.840]or two hours at the end of the day.
[00:39:52.670]Now, once we introduced an immediate probe with Emma,
[00:39:56.520]we actually saw her get the questions correct
[00:40:01.280]immediately following the activity.
[00:40:03.440]And then also got it correct at the end of the day.
[00:40:06.350]So really no intervention here whatsoever.
[00:40:09.950]We just sort of, I think,
[00:40:12.130]clarified what the task was was that we were doing with her.
[00:40:16.010]However, we don't typically ask kids immediately,
[00:40:19.290]what did you just do?
[00:40:20.220]And then again after an hour delay or two hours delay,
[00:40:23.290]so we wanted to make sure
[00:40:24.530]that she would retain this skill
[00:40:27.950]if we removed that immediate probe.
[00:40:30.120]And you can see once we went back
[00:40:31.790]to sort of not asking her immediately,
[00:40:34.960]she was now able to report all three activities,
[00:40:38.010]even when they varied, using her device.
[00:40:43.470]Okay, for Justin and Bruce, it was a little bit different.
[00:40:47.750]Now there was some increase in correct responding
[00:40:50.330]as soon as we added in that immediate probe,
[00:40:52.690]but it wasn't consistent.
[00:40:55.220]You can see for Bruce, there was-
[00:40:59.940]So for Justin,
[00:41:00.773]there was a little bit of a change.
[00:41:02.310]And for Bruce, there was really no improvement.
[00:41:04.440]So that makes it make a little bit more sense
[00:41:08.380]that they couldn't respond at the end of the day
[00:41:10.740]because they couldn't even tell us
[00:41:13.850]right after they did the activity.
[00:41:15.630]And so we knew we needed to intervene here
[00:41:17.870]and teach them how to report about things
[00:41:19.810]that they had just done.
[00:41:23.240]So this is where we added prompting.
[00:41:25.250]So if you have a kiddo who can't respond correctly,
[00:41:27.860]even after a teeny tiny, short delay,
[00:41:30.150]you'll probably wanna teach them
[00:41:31.630]how to immediately respond to that question.
[00:41:34.760]So again, we would do the activities,
[00:41:36.780]we would ask right after coloring,
[00:41:39.420]what did you do in the playroom?
[00:41:41.564]And then we would start prompting.
[00:41:43.540]So, if they got it correct, great.
[00:41:45.690]But if not,
[00:41:46.970]we would say you read a book,
[00:41:49.270]and the kid would say, read a book.
[00:41:51.830]Then we'd ask again, what did you do?
[00:41:53.719]And if their response was correct,
[00:41:57.460]we'd intersperse a distractor trial,
[00:41:59.440]then ask 'em one more time, what did you just do in there?
[00:42:02.700]And if they got that correct, we would reinforce.
[00:42:06.520]So really just giving them some practice,
[00:42:08.310]telling us what they just did
[00:42:10.720]in the location and getting that mastered.
[00:42:18.290]So you can see for both of these kids,
[00:42:20.510]as soon as we introduced prompting,
[00:42:22.500]they were able to tell us immediately
[00:42:24.730]after doing the activity,
[00:42:27.190]and also able to do it at the end of the day,
[00:42:30.100]and were able to report the activities that the did
[00:42:33.840]to their caregiver,
[00:42:34.810]which was really wonderful for them to be able to do.
[00:42:39.210]But just like with Emma,
[00:42:41.150]we wanted to make sure if we removed that immediate probe,
[00:42:44.380]that they would still be able to report the fun activities
[00:42:47.400]they did throughout the day after a lengthier delay.
[00:42:51.080]So we removed it and you can see
[00:42:54.625]when we probed again with the caregiver,
[00:42:56.760]both kiddos had a little bit of a dip,
[00:42:58.490]so we just kind of kept going.
[00:43:00.050]And by the end of this procedure,
[00:43:02.640]both participants were reporting accurately
[00:43:05.130]to their caregivers at the end of the day,
[00:43:08.570]using their speech generating device.
[00:43:12.860]So in general,
[00:43:13.750]all of our participants improved the accuracy
[00:43:15.860]of reporting past behavior at the end of the day.
[00:43:18.600]One participant, Emma,
[00:43:19.750]reported accurately following just the introduction
[00:43:22.400]of immediate probes.
[00:43:23.490]So that's a really simple thing
[00:43:25.120]that you could start interspersing into your day
[00:43:27.170]with learners, is just getting them into the habit
[00:43:29.750]of telling you, what did we just do?
[00:43:32.380]Two participants needed prompts.
[00:43:34.450]So if kiddos aren't reporting immediately
[00:43:37.160]just helping them to learn to be able to do that.
[00:43:40.240]And then once kiddos are able to report immediately,
[00:43:43.170]we saw a hundred percent reporting at the end of the day
[00:43:46.420]and generalization to their caregivers.
[00:43:50.050]Okay, in the last oh, I think 15 minutes,
[00:43:54.010]I'm gonna quickly go through one last study.
[00:43:56.870]We've talked about advanced manding
[00:43:59.210]or request for information.
[00:44:00.660]We've talked about advanced intraverbals,
[00:44:02.540]being able to answer questions
[00:44:04.240]about what we've done throughout the day.
[00:44:05.947]And the last study I wanna share
[00:44:07.690]is actually advanced tacting or labeling
[00:44:10.530]using a speech generating device.
[00:44:14.640]Now, research has found that children with autism
[00:44:16.970]do not combine words into multi-word utterances or phrases
[00:44:22.890]at the same time that typically neurotypically children do.
[00:44:26.880]And this is despite having a similar number of single words
[00:44:31.270]in their repertoire.
[00:44:32.600]So there seems to be something a little bit more challenging
[00:44:35.240]about word combinations.
[00:44:37.990]Instead, we often see rote and inflexible responding,
[00:44:42.546]and a lot of times kiddos with autism
[00:44:45.160]with significant language delays
[00:44:47.430]really are learning only language
[00:44:49.800]that's directly taught to them.
[00:44:52.690]Now, the work we're gonna talk about today
[00:44:56.289]is aimed at interventions
[00:44:59.130]that will promote word combinations that occur
[00:45:01.550]in flexible and novel ways,
[00:45:03.960]as opposed to just through direct teaching.
[00:45:06.370]And so I'm gonna go through a study,
[00:45:08.176]I'll show you some data
[00:45:09.480]with kiddos who use spoken vocalization,
[00:45:11.950]but also then how we replicated that
[00:45:14.030]with kiddos who use speech generating devices.
[00:45:17.930]Okay, so we've started to examine
[00:45:19.910]how to teach word combinations in such a way as to promote
[00:45:23.420]untaught, generative, responding, right?
[00:45:26.040]Flexible responding.
[00:45:28.740]So specifically this is with the kiddos with vocal language,
[00:45:35.070]we taught children to combine nouns and verbs together
[00:45:39.848]when they're sort of watching their instructor
[00:45:43.320]use a little figurine to perform an action.
[00:45:47.030]And we found that if you teach a subset of word combinations
[00:45:50.660]in a certain way,
[00:45:52.170]then you get a lot of novel word combinations
[00:45:55.160]that were never directly taught.
[00:45:57.780]And we do this through what's called matrix training,
[00:46:01.410]which sounds a little fancy,
[00:46:02.600]but it's actually fairly straightforward.
[00:46:05.120]So what do we know about teaching tacts to children with ASD
[00:46:08.490]who use speech generating devices?
[00:46:10.960]We know that there are a few studies that have taught tacts
[00:46:14.730]or expressive labels of pictures.
[00:46:18.810]There's at least one that focuses on labeling objects.
[00:46:23.390]But across these studies,
[00:46:26.000]the use of prompts and reinforcement
[00:46:28.010]have been shown to be effective in training these skills.
[00:46:32.080]But we need to further explore strategies
[00:46:35.690]that are specifically aimed at developing generativity
[00:46:38.680]for kids who are using speech generating devices.
[00:46:44.110]Okay, so we had three goals with this particular procedure,
[00:46:49.270]which was to directly teach noun-verb combinations, right?
[00:46:52.970]So I might ask, hey, what do you see here,
[00:46:55.760]or what's happening?
[00:46:56.930]And I'm acting out a particular action
[00:46:59.230]with a particular figurine, maybe an animal.
[00:47:03.170]Then I wanna assess
[00:47:04.300]what's called recombinative generalization.
[00:47:06.520]And this is, can a child then recombine
[00:47:09.910]those nouns and verbs in novel ways?
[00:47:13.230]And then I also also wanna assess,
[00:47:15.240]can they expressively label or tact
[00:47:17.700]using novel noun-verb combinations
[00:47:20.460]that were never included in any training scenarios.
[00:47:23.150]And we call this generalization.
[00:47:25.790]So, as I mentioned,
[00:47:26.623]recombinative generalization is a process in which
[00:47:29.080]individuals produce novel combinations
[00:47:33.450]when they put sort of known components together.
[00:47:37.370]So matrix training is a method to organize components
[00:47:43.940]so that we can try to promote
[00:47:46.130]this recombinative generalization.
[00:47:47.770]So it's a lot, let me me show you what it looks like.
[00:47:49.670]This is a matrix, essentially.
[00:47:51.840]You can see the nouns
[00:47:54.011]this might be dog, cat, pig,
[00:47:56.600]and then the verbs might be sleeping, eating, drinking,
[00:47:59.660]whatever it might be.
[00:48:00.990]Now, matrix training means that we train down the middle
[00:48:05.340]and then these ones that we probe,
[00:48:08.520]these combinations are recombinative skills.
[00:48:15.910]So, oh, I have some examples here.
[00:48:17.610]Sheep, bear and dog, jumping, sleeping, drinking.
[00:48:20.500]You've got kids who can do label those things individually,
[00:48:25.170]but can't combine them together very easily.
[00:48:28.330]So then we would teach, hey, what do you see here?
[00:48:30.630]What's happening here?
[00:48:32.010]And the kiddo would learn, sheep is jumping,
[00:48:34.510]or the bear is sleeping, dog is drinking.
[00:48:37.507]So our diagonal targets are directly taught.
[00:48:41.695]The non-diagonal targets are then probed
[00:48:44.680]for recombinative generalization.
[00:48:46.820]And we just see, like,
[00:48:48.220]if I just teach you these three things,
[00:48:49.960]can you get these other things correct
[00:48:52.180]on your own and by yourself?
[00:48:54.380]And then we're gonna probe a completely novel set of targets
[00:49:00.470]in what's called a generalization matrix.
[00:49:03.160]Now still the individual,
[00:49:05.355]nouns and verbs are known,
[00:49:07.090]but the kiddo has never combined them together
[00:49:11.250]into a word combination.
[00:49:13.980]So here are some data.
[00:49:15.080]Let me orient you to this graph.
[00:49:16.716]These are four vocal kids from our original study.
[00:49:20.870]You see three different data markers.
[00:49:24.460]The black diamonds are, let me see.
[00:49:29.860]I just lost my spot here.
[00:49:30.890]Yep, the black diamonds are our generalization targets
[00:49:33.910]from a completely novel matrix
[00:49:35.740]that never had any direct teaching.
[00:49:37.550]So that's our ultimate goal.
[00:49:39.430]The black circles are the diagonal targets
[00:49:42.400]from the first matrix that we're gonna directly teach.
[00:49:45.390]And the open data path are those non-diagonal targets
[00:49:48.750]that are the recombination targets.
[00:49:51.080]Those ones were we're hoping will recombine.
[00:49:54.820]So you can see here, we saw,
[00:49:56.840]sorry about that, with all three kids,
[00:49:59.380]no word combinations in baseline.
[00:50:02.170]Here you can see, we'll just look at the top one here,
[00:50:04.850]when we directly teach word combinations,
[00:50:07.300]all of our kids can learn that.
[00:50:08.990]And that's great.
[00:50:10.010]That's nothing to not be proud of,
[00:50:16.070]but if all they've learned is what we've directly taught,
[00:50:19.720]then we don't necessarily have generative language.
[00:50:22.370]So here in our post-test,
[00:50:24.770]we test all of those recombined combinations
[00:50:29.160]that I showed you in the matrix
[00:50:30.860]and that final generalization matrix
[00:50:33.700]that never even had the diagonal targets presented.
[00:50:36.820]And what you see with all of these kids
[00:50:39.452]is that just teaching that diagonal,
[00:50:42.850]those three targets down the diagonal and boom,
[00:50:45.810]they get all of these noun word combinations
[00:50:48.509]sort of for free.
[00:50:50.870]So, we replicated this study,
[00:50:53.420]matrix training to teach tacts or expressive labels
[00:50:56.630]using speech generating devices.
[00:50:59.120]And what did this look like?
[00:51:00.410]This study included three participants
[00:51:03.140]spanning the ages of four to 16.
[00:51:05.810]All of the participants had a diagnosis of autism
[00:51:08.640]and they were in sort of a focused ABA language program
[00:51:14.080]and really had limited vocalizations.
[00:51:18.020]This is what they looked like
[00:51:19.480]on their VB-MAPP administration.
[00:51:21.400]I get asked that sometimes, for the sake of time,
[00:51:23.750]I'm gonna go a little bit quickly through this part,
[00:51:26.430]but again,
[00:51:27.263]they had kinda level two or so on the VB-MAPP level,
[00:51:33.020]but no word combinations
[00:51:35.010]as shown on the tact milestone number nine.
[00:51:42.030]Okay, so all sessions were conducted in our program,
[00:51:45.630]in our clinic,
[00:51:47.813]the materials we used animals for those nouns,
[00:51:52.730]toy figurines,
[00:51:53.970]and they all had a little accessory item
[00:51:56.470]so that whatever action they were depicting
[00:51:59.390]could be kind of shown a little bit more clearly.
[00:52:03.380]The targets were selected for each participant
[00:52:05.470]based on the individual skills
[00:52:08.350]that they already had in their repertoire.
[00:52:10.100]So they could already tell us those nouns,
[00:52:12.120]they could label the animals
[00:52:13.310]and they could label the actions.
[00:52:14.810]They just couldn't put them together.
[00:52:17.970]So during baseline,
[00:52:20.880]each of the nine combinations
[00:52:22.720]within that matrix were tested,
[00:52:24.850]each trial consisted of the therapist selecting the animal
[00:52:27.930]or the toy figurine.
[00:52:29.280]And then having that figurine kind of act out
[00:52:32.700]what the action was and being asked, what's happening here,
[00:52:36.040]or what do you see?
[00:52:39.030]For each participant in the study,
[00:52:41.190]we developed two matrices, the matrix one,
[00:52:44.740]which we were going to teach the diagonal,
[00:52:47.270]probe the recombination cells,
[00:52:49.350]and then a completely novel generalization matrix
[00:52:53.710]that would never encounter any teaching.
[00:52:58.150]So both of these matrices had,
[00:53:00.910]these were mastered as individual nouns.
[00:53:03.590]These were mastered as individual actions or verbs.
[00:53:08.107]We taught the diagonal targets
[00:53:10.910]and then probed the non-diagonal targets
[00:53:13.620]for recombinative generalization.
[00:53:15.720]And then we probed the novel matrix
[00:53:18.350]that had known components,
[00:53:19.920]but had never been combined in any way.
[00:53:24.750]Okay, so here to be considered correct for a tact response,
[00:53:30.900]we would show the figurine doing the action,
[00:53:36.150]ask what's happening.
[00:53:38.510]The child would then go into their things folder,
[00:53:45.768]then to the particular category.
[00:53:49.050]In this case, animals,
[00:53:50.330]then navigate to the correct animal, dog.
[00:53:54.390]Pick the correct noun,
[00:53:55.430]whichever animal it was for that trial.
[00:53:57.930]Then they navigate back out
[00:53:59.550]to the corresponding actions folder here,
[00:54:03.360]go into there and select the correct verb
[00:54:06.203]of the action that was depicted.
[00:54:09.400]So in this case, dog jumping,
[00:54:11.890]and then they would then press the speech output device
[00:54:17.040]so that the voice output would occur
[00:54:20.210]with both the noun and the verb together.
[00:54:23.420]All right, these graphs show similar results
[00:54:25.610]to the one that I just showed you
[00:54:26.900]with our vocal participants in the first study.
[00:54:29.320]Again, this is a replication
[00:54:30.620]with kids who use speech generating devices.
[00:54:32.871]Again, the black diamonds
[00:54:34.780]are that completely untaught generalization matrix.
[00:54:37.770]That's like our ultimate.
[00:54:38.920]We really are looking for kids
[00:54:40.570]to now just be able to combine all the single nouns
[00:54:44.360]and single verbs in their repertoire,
[00:54:46.060]regardless of whether they've ever been combined together.
[00:54:50.006]Again, the black circles are those diagonal targets,
[00:54:53.690]which are really the only targets in that first matrix
[00:54:56.700]that get direct training or teaching.
[00:54:59.360]And then the open marker here
[00:55:01.200]represents those untaught recombined non=diagonal targets
[00:55:06.500]in that first matrix.
[00:55:08.640]As you can see, none of our kiddos,
[00:55:10.560]despite having those single words in their repertoire,
[00:55:14.330]none of them could engage in word combinations in baseline,
[00:55:19.300]for all three of our participants,
[00:55:20.800]we taught the diagonal targets,
[00:55:22.470]and no surprise, they learned them without any trouble.
[00:55:26.960]And when we tested the recombined non-diagonal targets
[00:55:32.050]and the generalization matrix for two of our participants,
[00:55:37.240]complete generative responding,
[00:55:39.080]which is exactly what we had hoped to see.
[00:55:41.770]Now with our third participant,
[00:55:43.690]we actually didn't see that.
[00:55:45.510]We saw the recombination.
[00:55:47.398]So the non-diagonal targets
[00:55:49.820]that were a part of this first matrix,
[00:55:52.350]but when we tested the novel generalization matrix,
[00:55:55.600]we didn't see that generalization.
[00:55:57.600]And it's a good thing that we tested this
[00:55:59.560]because this tells us we're not done.
[00:56:01.218]Ryan here is not able to sort of apply this
[00:56:04.750]to all the single words in his repertoire
[00:56:07.040]and combine them in novel ways.
[00:56:08.620]And so we don't wanna stop.
[00:56:10.470]So all we did at this point
[00:56:12.050]was engage in multiple exemplar instruction
[00:56:14.540]and we created a whole bunch of more matrices
[00:56:17.787]to just expose him to more teaching along the diagonal.
[00:56:22.410]So here you can see we developed another matrix.
[00:56:25.000]We taught here the diagonal, when we probed,
[00:56:29.130]we got the recombined skills, but again,
[00:56:32.040]the generalization matrix did not pop up.
[00:56:34.760]So we're not done.
[00:56:35.714]We taught another set of three targets,
[00:56:39.220]just down the diagonal.
[00:56:40.830]We saw some nice recombination
[00:56:43.260]but not the generalization matrix.
[00:56:45.860]So we taught one last one and here,
[00:56:48.590]all Ryan needed was a little bit more exposure.
[00:56:51.740]And so, teaching a series of sets here,
[00:56:55.920]and we were able to get that generalization.
[00:56:58.180]And what we love to see is generative flexible use
[00:57:03.090]of their speech generating devices.
[00:57:06.150]Okay, so in general,
[00:57:08.150]all three participants
[00:57:09.220]learned to emit noun-verb combinations when directly taught,
[00:57:12.470]all three emitted recombined responses.
[00:57:15.015]Two of the three showed immediate generalization
[00:57:18.240]to those novel combinations in the generalization matrix.
[00:57:21.800]And one participant eventually did,
[00:57:24.470]he just required multiple exemplars.
[00:57:28.130]So our goals were to replicate procedures
[00:57:31.580]that we've seen to be effective
[00:57:33.390]with children who use vocal language,
[00:57:35.850]children with autism,
[00:57:37.340]to those of a similar skill set,
[00:57:40.010]but who use speech generating devices instead.
[00:57:43.430]All of our studies required multi-step navigation
[00:57:46.360]with the exception of our typer.
[00:57:49.040]And what we see here is that children with autism
[00:57:52.070]presenting as level two,
[00:57:53.410]level three learners on the VB-MAPP who are non-vocal,
[00:57:57.040]can acquire complex skills using their SGDs.
[00:58:02.090]And we really need more research into the development
[00:58:05.120]of advanced verbal behavior
[00:58:06.470]using high tech speech generating devices.
[00:58:10.760]Thank you so much for your time today
[00:58:12.810]and looking forward to any questions that you might have.

The screen size you are trying to search captions on is too small!

You can always jump over to MediaHub and check it out there.

Comments

0 Comments

Teaching Advanced Verbal Behavior Using Speech- Generating Devices

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments