 Hey guys, today marks the fourth in the series I'm doing, all this month on taking lesser talk to go parts of autism and explaining them to a sort of scientifically informed lens, but layman style. So I left yesterday off not really sure on what I wanted to cover, but I had just been reading some general stuff, sort of preparation for everything this month, and something really stuck. It's stuff that I've already been familiar with, but just seeing more evidence for the thing, like, you know, let's let's do that today. So faces for a long, long time now, it's been thought that the autistic struggle with recognizing face, especially the emotions in those faces. And I do understand where people are coming from. You're making eye contact less, how could you possibly be picking up information? You have to be missing it, you're not making eye contact, right? Now, the reports of some of the autistic kind of call it into question. The example I gave in the first video was Carly Fleishman describing it as like taking thousands of pictures of your face every second, how overwhelming that is. And I think that's highly fitting. But understandably, it started out as you have to be deficient at processing emotions. It seems like a great way to explain away the social difficulties, still, you know, I make eye contact, you're missing important parts of communication that leads to social communication. That has to be it. Right? Specifically, two key things there. The autistic have impaired ability to recognize emotions and impaired ability to recognize faces in general. There are some reasons why they were thinking that. Now when we're talking about emotions, regardless of exactly what you're talking about, it's a good idea to understand that there are both simple and complex emotions. The Gottman wheel is a good model to use for analyzing this, although there aren't others. Really anything that sort of breaks it down into the hierarchy that is our emotions will work just fine. An example would be happiness. That's a very basic general emotion. What does it mean to be happy? What specifically are you feeling? But generally speaking, that's a good classifier. More specifically, you could get into something like pride. If you're proud of work that you've done, you are happy, but pride is a much more specific thing than other forms of happiness. You can be proud without being ecstatic. You can be ecstatic without being proud. You could potentially be both, but both are happy. This distinction is important because some of the research that was originally being done on whether or not emotional recognition in the autistic world paired was inconclusive. There was a lot of research that was saying that there were deficiencies. There was a rather good amount that was showing no deficiencies at all, or in some cases slightly better functioning. So what the hell is going on there? You shouldn't be singing conclusive anything. Either you're better or you're worse at it, right? Or it should be relatively the same. No statistical difference, but they're finding both. And that's odd. So they got thinking, maybe it's that the autistic are fine at recognizing these simple emotions, but struggle to identify these more complex emotions. So the idea there that I would be able to recognize that someone's happy, but not necessarily that they are specifically proud, that would lead to a less useful context in the social interaction for all, and would then lead to the social difference, right? That makes sense. A similar problem. Results were inconclusive. A lot of it was showing deficiency is a good amount of it was showing better capacity than normal. And again, how do you have something like that happen? So still bubbling down on the idea that the autistic have to be deficient in this, because can't be something else. They got in the, well maybe it's the result of analyzing faces in a more feature role versus configuration kind of way. You know, what feature role face processing means is like looking at, say, the size of the eyes, whereas configuration facial processing would be more like, okay, well, what's the distance between the eyes? You can use both approaches in sort of analyzing what an emotion is being expressed. I can't explain this without spoiling the findings. So configural processing is what most people do, and it's not what the autistic do. Autistic, much more feature role. But it was believed that, okay, this has to lead to the difficulties. Feature role processing has to be worse than configural processing. And so this says to explain these difficulties. They finally found evidence of something. The autistic are more feature role facial processors, non-autistic are more configuration feature role processors. But this was unable to consistently find a deficiency. So what the hell is going on? Those discrepancies are interesting. And in an effort to try to better understand what's going on, what some researchers started to do was, are we inadequately controlling for things? Because normally when you see variance, that's probably what's going on. And they've been consistently finding variance no matter how they look at it. So we're probably inadequate with control. One of the things people thought to check for is to look at whether or not social intelligence was being analyzed and adequately controlled for us as a result. One of the problems with a lot of autism research is when they get samples. And I'll delve way more into this in a much later video on Simon Barrett Cohen's work because it is horrible. A lot of the samples for the neurotypical children are gotten from school. So they'll take a local school and sample children from there. Whereas the autistic children are sampled from hospitals. They're not necessarily in school. They don't necessarily have the same kind of social exposure. They rarely control for additional confounding factors. In fact, a lot of studies have had problems with taking neurotypical school children and then autistic children that have not been controlled for, like, the literature says mental retardation. We try to say, like, developmental disability now. Obviously that would cause some problems. Whereas if you were to take all non-autistic children with the, not sampling school specifically, but just and actually have some MR children in there, the results would be more balanced. But ideally you should filter out the MR from both as you're trying to look at autism, which is non-autism, not autism plus MR against neither. And then trying to argue that this only applies to the autism and not the MR or both together. So it's a very similar problem with the social intelligence. And so some researchers started to control for this better and do questionnaires on social intelligence and either exclude some, whether they scored too low or whatever from the study on both sides or group them into different categories. And those multimodal ones where they grouped into different categories had one of the more interesting findings. That whether you were autistic or not did not make any difference as far as broad emotional recognition goes. That somebody with the same social intelligence score, regardless of whether they are autistic or not, is going to have the same deficiencies in recognizing emotional expression. And if the social intelligence score is average, there's going to be no difference in determining any broad sense with the emotions in the face, whether you're autistic or not. So then one thing that naturally rises from that are do the autistic score lower on social intelligence scores? Originally they started finding yes, but this is where the MR thing, that conflation winds up biting them in the ass again. They didn't control for that. And so when that was controlled for, so you've got an autistic where there's no additional underlying conditions, there winds up being no difference in the social intelligence distribution. The once things started to be adequately controlled for, like they are supposed to be done in proper science, because you want to compare one discrete thing against a control, not a cluster of things against a control where you then describe the findings to one particular part of that cluster. That's not, no, it's bad stereotyping. One discrete thing. What they were able to find was that if you have autism, autism alone, your ability to recognize simple emotions is just as accurate and just as fast as neurotypical individuals. Now in practice, you don't find this next part all that often. But if you rotate the faces, whether it's a picture that you rotate or you rotate the person, they weren't too clear in the methodology. What was going on here? I don't know, like tilting your head or actually rotating the person overall. But the autistic at that point actually start to form a little bit faster and that's actually an advantage to feature processing. And it's why computer facial recognition systems tend to favor feature processing a little bit more. Although it depends on the exact computer vision system that is being involved there. But that's a whole other topic covering tech really in this. As far as more specific emotions go, so from happiness moving into pride or another one of those more complex specific emotions. This is where things get a little buried and I think is a part of why the original researchers were noticing a lot of variants. You know, part of it was controls, but they did legitimately find some differences. And that's why one of the previous things they referenced. I emphasize that it was general emotions. As far as specifics go, you do actually start to see some differences in the ability to process them. Now, you'll be noticing a theme here. Researchers, many of them, not all, many of them, still had it that the autistic have to be deficient in this. That has to be the result of the problems. Has to be. I don't know why so dead set that it has to be deficient and not a difference. But that's the attitude many have. What they were able to find and should be obvious considering how many times I've said it already in previous videos, I do have links in the down in the video description. The parts I'm talking about right now have much more relevance to the cited literature rather than just historical stuff. So think, wishful thinking or whatever. People will dismiss some of this stuff. These specific claims that I am making all have direct citations down in the video description. Negative emotions, especially fear and anger, are processed just as accurately in the autistic as neurotypicals, but also are processed faster in the autistic than in neurotypicals. So somehow we are better in speed, not accuracy, but not deficient in accuracy at recognizing negative emotions, specifically fear and anger. Now there are some reasons to suspect why this is the case and there are additional studies involved in this that do show atypical responses to fear. So presents differently. Our fear response is different. But the ability to recognize that somebody else is afraid or that ourselves is afraid is just as accurate and similarly anger just as accurate is also faster. Most complex emotions they didn't really notice a difference for. But these researchers that are adamant, there has to be some kind of deficiency. They were right in a remarkably limited time. We did manage to find some very specific deficiencies in the ability to recognize certain emotions in the autistic, specifically whether the other person is embarrassed or not, and in our ability to assess the trustworthiness of another person. Those two actually do have quality studies behind them that confirm an actual deficiency, but those are the only two that I have seen in several dozen articles that have been able to actually firm through well-controlled repeatable research a deficiency. This also fits some findings, descriptions rather from some of the autistic. A common complaint involved in why we socialize less, aside from increased load from processing voice, has to do with difficulty navigating what people's actual intentions are, essentially, trustworthiness. There is a tendency to get burned a lot by other individuals to put yourself out there and have them stab you in the back. This seems to actually happen far more often among the autistic, both in reports from the autistics, but also in some experimentally controlled research. Not only does the evidence say that there's an actual deficiency in the ability to recognize the trustworthiness of another individual, but also it does actually seem to check out in practice. So get out of the lab and look at what's actually going on. It seems to also be there, so the trustworthiness especially, I feel definitely checks out embarrassment. I want to see that be checked outside of the lab as well, but I could see it potentially being the case, so I'm not going to give the heart that a hard time. And again, the research that was able to show that was good, but I didn't notice any problems there. So quick summary of what we've talked about, the autistic process faces differently, but generally speaking, this doesn't really lead to any differences in the accuracy in determining emotions. There are some very minor exceptions, but it goes both ways with certain emotions being recognized quicker and others have some trouble with recognizing, but overwhelmingly, there isn't actually a difference. A lot of, no, I'll get into criticizing the research and other. So hopefully this has been helpful. I'm thinking for next video, we'll actually cover that fear response, why the autistic seem to be better at recognizing fear, but also seem to have a little bit of a different fear response themselves, and also how that ties into why allergies are less common in the autistic and why gastrointestinal problems are more common, because believe it or not, all three of those have a lot more to do than common than you'd probably think. They're actually seem to be caused by the same thing. I promise that's going to make sense. But until then, have a good one, guys.