Microexpressions

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.
B

Bmewriter

Full Member
10+ Year Member
Joined
May 26, 2009
Messages
59
Reaction score
3
Anyone into these? I'm kind of a psych fanboy ever since I've learned to read these. I feel like Cal Lightman!
Members don't see this ad.
 
Ekman himself says that the majority of micro expressions are meaningless and that the character does several hours of work in a few seconds....
 
PSYDR said:
Ekman himself says that the majority of micro expressions are meaningless and that the character does several hours of work in a few seconds....
Click to expand...

They tell you what the emotion is, but not the reason for it. I don't regard that as meaningless.
 
Members don't see this ad :)
WisNeuro said:
I know it's what they try to use when teaching lie detection, although even trained individuals are barely better than chance from the research I've seen. I don't know much about its uses in other areas.
Click to expand...

Not going to lie. I've tried using them to detect lies and I'm terrible at it. All I can say is how a person actually feels. As for the reason....I dunno. I'm still fascinated by them though and the universality of emotion.
 
WisNeuro said:
I don't argue that the concept is interesting, I just don't know too many practical uses for it from the research.
Click to expand...

As a medical student, I use them to figure out how my attendings and residents are feeling and then I act accordingly. I know when I saw teaching, I would know when students didn't understand what I was talking about due to the disgust on their faces. But this is more anecdotal.
 
It's the consistency of the emotion in response to a stimuli, not a single event. Which is why there is a complex coding system that allows for statistical analysis. A single event could be an inconsistent emotion, it could be a fart. Ekman's current online mett and sett training are essentially lay modifications to his facs training.
 
PSYDR said:
It's the consistency of the emotion in response to a stimuli, not a single event. Which is why there is a complex coding system that allows for statistical analysis. A single event could be an inconsistent emotion, it could be a fart. Ekman's current online mett and sett training are essentially lay modifications to his facs training.
Click to expand...

I've done both trainings and yeah, I look for the consistency. I tend to use them to then go ahead and have a conversation until I reach some sort of Duchenne smile. Once I get to that smile, I kinda know how to modify my conduct to suit that person.
 
Just whatever you do, don't let your students know that you are coding their microexpressions! My advisor is a FACS expert which makes for anxiety inducing meetings. :nailbiting:
 
The last time I looked at the research (admittedly 5+ yrs ago) it was mostly junk/fringe psuedo-science. Beware that pushing a fringe topic may reflect poorly upon your judgment of research.
 
  • Like
Reactions: 1 user
Members don't see this ad :)
Therapist4Chnge said:
The last time I looked at the research (admittedly 5+ yrs ago) it was mostly junk/fringe psuedo-science. Beware that pushing a fringe topic may reflect poorly upon your judgment of research.
Click to expand...

Agreed - we've done some FACS training in our lab and it is very important to remember that you are only coding the movement of specific facial muscles. You certainly are not coding for an emotion, which Bmewriter claimed. For example, we had a clip of a woman having a conversation with her partner. Only she was visible and we were unable to hear the content of their conversation. Our RAs all coded this woman as expressing an AU 12+6, a Duchenne smile. However, upon playback with audio, it was clear that she was actually experiencing fear/tension.
 
  • Like
Reactions: 1 users
Therapist4Chnge said:
The last time I looked at the research (admittedly 5+ yrs ago) it was mostly junk/fringe psuedo-science. Beware that pushing a fringe topic may reflect poorly upon your judgment of research.
Click to expand...

Facial Coding broadly or the idea that one can do "live" readings?

The literature behind FACS is actually pretty solid and has been for some time. We're talking about spending hours at a time going frame by frame through video - very different from "reading" live microexpressions. I certainly haven't seen much to suggest the latter is realistic, though you can probably find a half dozen workshops that will tell you otherwise for a couple thousand dollars.
 
  • Like
Reactions: 1 user
Ollie123 said:
Facial Coding broadly or the idea that one can do "live" readings?

The literature behind FACS is actually pretty solid and has been for some time. We're talking about spending hours at a time going frame by frame through video - very different from "reading" live microexpressions. I certainly haven't seen much to suggest the latter is realistic, though you can probably find a half dozen workshops that will tell you otherwise for a couple thousand dollars.
Click to expand...

I believe the expressions can be read reliably, it's just that the reading of those expressions lacks practical value according to what it's purported to do in many instances (e.g., lie detection).
 
  • Like
Reactions: 1 user
Ollie123 said:
Facial Coding broadly or the idea that one can do "live" readings?

The literature behind FACS is actually pretty solid and has been for some time. We're talking about spending hours at a time going frame by frame through video - very different from "reading" live microexpressions. I certainly haven't seen much to suggest the latter is realistic, though you can probably find a half dozen workshops that will tell you otherwise for a couple thousand dollars.
Click to expand...

I was referencing the latter…and the associated "weekend seminars" in it.
 
WisNeuro said:
I believe the expressions can be read reliably, it's just that the reading of those expressions lacks practical value according to what it's purported to do in many instances (e.g., lie detection).
Click to expand...

To a certain degree yes, but its certainly going to be a much noisier/less reliable way to do it vs the incredibly intensive coding process used in FACS. Which is one of several reasons its of limited practical value in most circumstances...more noise = less predictive power. Given we're generally talking about dozens of muscle groups acting at once with enormous numbers of interactions determining meaning....its going to be a messy process when you are trying to recognize a half-second twitch in the corrugator muscle from across a room.

Really though, I think the biggest barrier is that these are signals of emotion...which is generally a pretty poor predictor of any behavior, let alone the extraordinarily specific ones people purport to assess with them.
 
  • Like
Reactions: 1 users
Therapist4Chnge said:
The last time I looked at the research (admittedly 5+ yrs ago) it was mostly junk/fringe psuedo-science. Beware that pushing a fringe topic may reflect poorly upon your judgment of research.
Click to expand...

I kinda want to know more about the pseudoscience aspect of it. What about microexpressions makes it a psuedoscience rather than an actual science, because as I understand it, Ekman compared facial expressions of different emotions across cultures and found them to be the same, even if they came from a tribe in Papua, New Guinea. I don't really find the lie detection part appealing as much as it being a visible signal of emotion. If it is a pseudoscience, then (a) what did Ekman do that made this a pseudoscience rather than something that was an actual science and (b) what other studies have been done on human emotion and the way that it is expressed? Is it at least proven that human emotion as it is expressed on the face is universal? and (c) is perhaps taking a "history" and putting a person in "context" a better predictor of future behavior than using microexpressions? Because in the end, I'm interested in the whole science behind a person's motivations and attitudes, especially if it will help me understand my patients better and give them the treatment that they need in the way that they would like.

Tupelo Honey said:
Agreed - we've done some FACS training in our lab and it is very important to remember that you are only coding the movement of specific facial muscles. You certainly are not coding for an emotion, which Bmewriter claimed. For example, we had a clip of a woman having a conversation with her partner. Only she was visible and we were unable to hear the content of their conversation. Our RAs all coded this woman as expressing an AU 12+6, a Duchenne smile. However, upon playback with audio, it was clear that she was actually experiencing fear/tension.
Click to expand...

Wait...what? Do you have a paper or can you link me to a paper with similar data? I want to read up on this. I thought that something like Duchenne was actually reliable. =/

How did you determine that she was feeling fearful or was in tension? Was it within the context of the conversation?
 
Last edited:
I don't believe that people are saying that the universality of emotion stuff is pseudoscience, rather, what people are doing with it. It doesn't predict behavior all that well, although some people, usually people offering weekend training courses for a couple grand, will say that it does, with little to no empirical backing.
 
WisNeuro said:
I don't believe that people are saying that the universality of emotion stuff is pseudoscience, rather, what people are doing with it. It doesn't predict behavior all that well, although some people, usually people offering weekend training courses for a couple grand, will say that it does, with little to no empirical backing.
Click to expand...

Ohhhhh. That makes way more sense. I know that I can't fully rely on them, so I have to ask questions or comment in a certain way to confirm what I saw is what I saw. I don't use them for lie detection, because even if you can pick out the emotion, you cannot know the reason why. I want to hear more from Tupelo though, because apparently, he found a case in which Duchenne's indicated fear rather than true happiness.

So I had this question: Is it better to take a history of a person, put him in his social "context" and given what you know about a person's motivations, phobias, etc, to then predict what person is going to do next? Somehow, I would like to think that it is that simple, but I have a strong feeling that it isn't.
 
A big part of my job as a highly skilled psychotherapist is reading facial expressions. This is actually the first I have heard of microexpressions as a research area, but we practiced reading facial expressions during grad school. We used a Tony Soprano psychotherapy session to do that and went frame by frame. We also studied attunement and affect regulation in infants. These are things that I use everyday. Can I predict behavior based on it? No. Can I tell if someone is lying? No. Can I tell if someone is on the verge of crying? Yes. Can I see someone fighting to control their anger? Yes. In my context though, I just ask the patient what they are feeling and they tell me and it is either congruent with what I see or incongruent. Pretty straightforward stuff. I also have concerns when we come up with some new theory that could become popular and then over-apply it. Not good for our reputation and what we actually do.
 
Bmewriter said:
So I had this question: Is it better to take a history of a person, put him in his social "context" and given what you know about a person's motivations, phobias, etc, to then predict what person is going to do next? Somehow, I would like to think that it is that simple, but I have a strong feeling that it isn't.
Click to expand...

I'm not quite sure what you are asking here. Can you be more specific in a clinical context?
 
WisNeuro said:
I'm not quite sure what you are asking here. Can you be more specific in a clinical context?
Click to expand...

Taking a patient's socioeconomic status, past history, culture, educational history, etc and then using that to determine what type of patient you have and using that to determine what you should be doing for the patient. If the patient grew up in an immigrant family where he or she is expected to excel in school and go to an Ivy and has been having symptoms of depression from that, for example, then using that background to determine what the appropriate therapy is for the patient. I wondering if the answer of what type of therapy given will change depending on the person as each person is different.

Perhaps, also using microexpresssions to gain more information as well. Microexpressions, from what I hear, are not the be all, end all, but surely, if someone is making an expression of what you think is disgust, then asking questions in a certain way to figure out if the person is truly disgusted or not.

Something like that.
 
WisNeuro said:
Are you implying that using microexpressions, you would change treatment recommendations?
Click to expand...

Oh hell no. Microexpressions aren't the be all or end all, nor do I think they should be used to make decisions. Just that if you think if you see an emotion, you should ask more about it to find out more about your patient. I don't think microexpressions are always accurate, nor should they be used to make clinical recommendations. Just if you think you see something, you should either ask or say something in a way to confirm or deny what your patient is feeling and then go from there.
 
Depends on what you need to do. In my interviews, I just need information (i/ADL functioning, changes in physical status, timeline of cognitive changes, etc), the emotions in the moment are relatively meaningless. Also, is it necessary to constantly have a running narrative of a patient's emotions to treat them for a specific problem? I can see maybe a time or two where this would be helpful in a therapy context, but I could see it also interfering with the session if the therapist is overly focused on trying to accurately read microexpressions.
 
Bmewriter said:
Oh hell no. Microexpressions aren't the be all or end all, nor do I think they should be used to make decisions. Just that if you think if you see an emotion, you should ask more about it to find out more about your patient. I don't think microexpressions are always accurate, nor should they be used to make clinical recommendations. Just if you think you see something, you should either ask or say something in a way to confirm or deny what your patient is feeling and then go from there.
Click to expand...
As I stated before it is already part of what we do. When I see something in session, I either comment on it or don't depending on a variety of factors. Also, in the example you mentioned above about taking into account the various socioeconomic factors. A good psychotherapist does that, as well. I actually am thinking that this whole concept of microexpressions sounds like a reinvention of the wheel. If you want to know about facial expressions and how they relate to neuroanatomy and emotional regulation and the interpersonal and how that can apply to psychotherapy see Alan Schore or Daniel Stern or Lou Cozolino. I think these guys are probably a step ahead of microexpressions.
 
You must log in or register to reply here.
Next unread thread
Top