Fidgeting as Lie Detection

Sophie Van Der Zee and colleagues have a new paper on using body movement as a lie detector:

Abstract: We present a new robust signal for detecting deception: full body motion. Previous work on detecting deception from body movement has relied either on human judges or on specific gestures (such as fidgeting or gaze aversion) that are coded or rated by humans. The results are characterized by inconsistent and often contradictory findings, with small-stakes lies under lab conditions detected at rates only slightly better than guessing. Building on previous work that uses automatic analysis of facial videos and rhythmic body movements to diagnose stress, we set out to see whether a full body motion capture suit, which records the position, velocity and orientation of 23 points in the subject’s body, could yield a better signal of deception. Interviewees of South Asian (n = 60) or White British culture (n = 30) were required to either tell the truth or lie about two experienced tasks while being interviewed by somebody from their own (n = 60) or different culture (n = 30). We discovered that full body motion—the sum of joint displacements—was indicative of lying approximately 75% of the time. Furthermore, movement was guilt-related, and occurred independently of anxiety, cognitive load and cultural background. Further analyses indicate that including individual limb data in our full bodymotion measurements, in combination with appropriate questioning strategies, can increase its discriminatory power to around 82%. This culture-sensitive study provides an objective and inclusive view on how people actually behave when lying. It appears that full body motion can be a robust nonverbal indicator of deceit, and suggests that lying does not cause people to freeze. However, should full body motion capture become a routine investigative technique, liars might freeze in order not to give themselves away; but this in itself should be a telltale.

This is a first research study, and the results might not be robust. But it certainly is interesting.

Blog post. News article. Slashdot thread.

Posted on January 6, 2015 at 2:44 PM46 Comments

Comments

nony January 6, 2015 4:04 PM

Presumably, someone malicious (i.e. double agent) would have some training to throw this technique off by random movements throughout the interview.
I’d be more interested to know what the success rate is when the interviewees are actively trying to mislead this technique.

Darron January 6, 2015 4:24 PM

As someone who drives my friends crazy with my fidgeting, I’m more concerned about the false positive rate.

grepo January 6, 2015 5:01 PM

Another stupid lie detector machine/scheme. Just what we need, more false positives and negatives to get innocent people in trouble with corrupt or zealous government agents.

Sancho_P January 6, 2015 5:31 PM

Did the study include the President of the United States of America?
What does it say then?

Bauke Jan Douma January 6, 2015 5:41 PM

@grepo

You said it. Exactly my sentiments, some more stupid servile Brits, preoccupied with terrorists as they are (and as themselves they have become — though they see themselves as saviors), accompanied by some dumbass wannabee Sixth-Eye to the Five Eyes Dutchies, working their way up into the Big Brother USA-style upper echelon.
People lie, so do you so do I. What else is new? The neuroticism of efforts like this in trying to catch them in the act? Hardly. The zeal, the money that goes into control. Control. Control. Control. Because that is what this is all about — yet again.
Control.
Croud control.
Individual control.
Because all these studies cannot cover up the fact that some things are artifcially scarce and the ones ending up on the short end of the stick time and again, every time, far outnumber those that think up ways to control control control because they fear with great fear, deep fear and angst.

Anura January 6, 2015 5:45 PM

@Sancho_P

Given that people who are able to obtain high political offices, especially those such as POTUS, tend to exhibit psychopathic tendencies, they will probably pass most generalized lie-detecting methods.

Bob S. January 6, 2015 5:55 PM

“Ask me no questions, I’ll tell you no lies.”
(Also see Bill of Rights -5th Amendment.)

No one, sane or not, should ever voluntarily agree to a polygraph. It’s all snake oil, witchcraft and voodoo no matter how scientific they make it sound.

Well, I guess you could say I not especially fond of body scanner lie detectors then.

albert January 6, 2015 7:09 PM

This is a load of bull (to paraphrase Bill Clinton).
.
How big a load depends on how it’s used.
.
Here in the US, any lawyer will tell you this, which is the very best advice you can get, “Say nothing.” You are not required to give a statement, or answer any questions in any interview by law enforcement. The purpose of interrogation is to establish evidence of guilt. That’s it.
.
If LE has evidence of guilt, they will arrest you immediately; if not, they will use interrogation. Even if you are innocent, especially if you’re innocent, answering questions can be very dangerous, no matter how clever you are. There are many examples of this happening to innocent people.
.
LE wants convictions, and they are not too concerned about where they find them.
.
Remember this, if you witness a crime, you are also a suspect in that crime.
.
I gotta go…

v January 6, 2015 10:26 PM

Fast forward a few years…
Some employees are still required to wear uniforms.
Uniforms have built in motion capture capability, bundled with other tech, for ‘health & safety monitoring’.
Your employer calls you in for a ‘chat’.
Interesting future in terms of legal rights.

Marc January 7, 2015 12:31 AM

I’ve got enough free-floating guilt hanging over my head to defeat this, I think. Tell the truth, think about shameful things; tell a lie, and concentrate on the current situation. Should flatten my response curve nicely!

Ray Dillinger January 7, 2015 2:09 AM

Lying affects different liars differently.

There is a minority – more significant in numbers than most realize, maybe as much as seven or eight percent – who have absolutely no sense of shame or guilt, no matter what. Among other things they are completely psychologically unaffected by the awareness that they are lying. They won’t fidget. Their pupil dilation won’t change. They won’t sweat or develop tells. And they are the slimiest assholes in the whole damn world.

And, bluntly speaking, I’ll trust someone whom I know is lying, AND AFFECTED BY LYING, more than I would trust someone who even might possibly be one of those assholes. The devastating harm that can come from trusting one of them is far worse than the harm that comes from trusting an ordinary liar, because an ordinary liar usually at least gives a crap about whether you live or die.

Y’all can worry about the false positives, the people who are telling the truth and get tagged for lying. I’m a whole lot more worried about the possibility that I might be dealing with someone who can elicit false negatives.

Andrew_K January 7, 2015 2:51 AM

The last sentence of the abstract is the most interesting. It’s yet another example of the old “if you don’t have anything to hide, you can talk to the police!”. Talk bias…

Just realize — they are wrong on at least 20% (correct me if I misinterpreted), that’s every fifth case. Who would take a medicine that does not heal but hurt in every fifth case?

On their basic assumption: I had to learn to stand still in situations where my body wants to move and I kind of got used to it. I learned to control body motions up to the point of overcoming the awkwardness of staring down others (which was a rather big step for me). This can be learned, it’s a question of training.

Further critique can easily be deduced from the study setup. I would never confuse test participants in scientific environment which are told to lie (and know that nothing depends on it) with “professional criminals” (who know what depends on their deception or silence).

Winter January 7, 2015 3:13 AM

I think the problems of lie-detectors start with the idea that there are only two options on a one-dimensional scale: Lying or telling the truth.

This belongs to the same class of myths as “The literal meaning of a text”. These myths are not even wrong, they make no sense.

So, I do not have to think long about whether this will scheme will work.

Gerard van Vooren January 7, 2015 3:52 AM

@ Sancho_P

Did the study include the President of the United States of America?

Why do you think that at press conferences they stand behind that large table with the big logo? It’s not only because you can hide a hooker or a bottle of scotch there but also that whenever answering a question the audience doesn’t see the legs shaking as if the guy has severe Parkinson’s disease.

Paul Renault January 7, 2015 6:37 AM

In the future, the interviewers will be forced to regularly ask the interviewee whether they need to go to the bathroom…

Wael January 7, 2015 7:09 AM

This technique is prior art. Body movements, specifically the lips, were used for several years to detect lies. Here is the proof: Q: How can you tell if a politician is lying? A: when his/her lips are moving 🙂

Clive Robinson January 7, 2015 7:10 AM

Hmm any one else read,

… we set out to see whether a full body motion capture suit, which records the position, velocity and orientation of 23 points in the subject’s body, could yield a better signal of deception.

Does anyone think forcing people into “gimp suits” which have lot’s of trailing wires to not be a form of torture?

Another issue is humans actually have very few major symptoms for hundreds of different triggers.

That is the major response to intense stress is the same as travel sickness and quite a few others the primary evolutionary one being “I’m poisoned”. So your brain detects anomalies from normal operation assumes it’s poisoning and responds by making you evacuate food fluids etc…

Thus a reasonable supposition –as found with other lie detectors– is that all manner of inputs will produce the same major responses as telling lies… Which will almost always fail to give valid indicitive responses whilst leaving the door wide open to abuse…

I was once told by a researcher that people had it wrong about “bad cop good cop” it was not the fear of the bad cop or the hope the good cop would be nice. No it realy only works with those who have sufficient empathy that actually feel sorry for the good cop and don’t want him to look bad infront of the bad cop…

That is we instinctivly don’t like Hawks so we help Doves, apparently most of us have a deep seated primative desire to protect “young” from Hawks as a species survival mechanism.

These deep seated survival mechanisms primary purposes may long be gone in normal life, but the brain/body has recycled them for a myriad of other reasons. And as with all such things the edges are ill defined and very very broad and thus close to impossible to diagnose the trigger that gives rise to the symptoms. Such one wayness is ideal not just for those wishing to abuse the questioning of individuals, but also for those trained to resist such questioning.

Then of course there are the eighty percent of the population who are not “average” whose triggers, physiological responses and physical anomalies tend to create more noise than signal with these types of techniques. I often wonder if in twenty or fifty years people will look back on this sort of thing in the same way we now regard the art of phrenology, and similar where it was believed that bumps on your head, or the shape of your ears etc was proof positive you were in “the criminal class” and thus “must be guilty”.

Any way time to eat lunch, hopefully my body won’t think it’s being poisoned 😉

paul January 7, 2015 9:31 AM

Even if this kind of tech got way better than it is now, and you could get past the issue of false positives and false negatives in general, the biggest issue (as several other commenters have said implicitly) is the distribution of those false positives and false negatives. If errors are randomly distributed then using the technique can be better than not using it, but if they’re not randomly distributed (and the user doesn’t know the factors governing the distribution) then using the technique can be worse than not using it. This is exactly the argument about profiling versus random screening, just in a different context.

Meanwhile, a couple of notes about the tech. On the one hand, I don’t put that much faith in studies where subjects are instructed to lie or not lie (even with incentives) because the conditions are so far from real attempts to distinguish between honest answers and deliberate falsehoods. On the other hand, this kind of thing isn’t going to require complex gadgetry if put in practice. Even today you could do most of the processing with a few cameras (especially if the subject is wearing distinctive garb) or with a sensor array in the room and a bunch of locatable tags sewn almost invisibly into a prison jumpsuit.

Oliver January 7, 2015 10:53 AM

Lie detection is bullshit, plain and simple.
There is no scientific evidence that any of that supposed “lie detection” is anything more than snake oil selling.

Any “research” into those things needs to be mocked senselessly, and repeatedly!

Matt January 7, 2015 11:17 AM

Your molecules were moving! Liar/Guilty by default!

I think it is then safe to assume that all people are lying at all times.

Herman January 7, 2015 11:20 AM

So, someone who smoked a doobie, or a cigarette, or used a mega dose of vitamin B8, or used any other anti-anxiety medication before the test, will do better and anyone who suffers from ‘restless legs’ or Parkinson’s disease, or who has diabetic (or hunger) tremors from low blood sugar (or substance withdrawal), will do worse.

As Marilyn Manson put it: We’re all stars now, in the dope show…

Obviously, this technology is just what we need in the 21st century.

Bob Paddock January 7, 2015 12:14 PM

“Who would take a medicine that does not heal but hurt in every fifth case?”

Far to many actually.

The antibiotics Levofloxacin (Levaquin), Ciprofloxacin (Cipro), Moxifloxacin (Avelox), Norfloxacin (Noroxin), Ofloxacin (Floxin), Gemifloxacin (Factive) and Finafloxacin (Xtoro) should be taken off the market, all members of the Fluoroquinolone class. Instead the FDA puts on a Black Box warning that the patient never knows anything about.

This class was meant to be the line of last resort for things like Anthrax. Sadly doctors are giving out Levaquin and Cipro like candy, maiming many people for life. They cause DNA damage and tendon ruptures in many individuals, such as my late wife. I watched Karen crawl around the house for a year to let her tendons heal so she could walk again. Sadly the damage is not reversible and contributed to Karen’s suicide.

Everything in this picture is the result of Karen taking Levaquin. <a hrf=”http://www.kpaddock.com>Karen’s Journal is now required reading at Duke School of Medicine. Everyone that I hear from say they are in tears by the end of reading it.

Fred P January 7, 2015 3:56 PM

It is unfortunate that they don’t appear to test attacks on their system in this paper.

The most obvious one to me is someone who attempts to move in roughly the same way regardless of the veracity of their statements. An obvious upgrade would be someone training themselves to do so (instead of just thinking about it).

Another obvious attack would be to think of something else – say, a lie – whenever telling the truth.

asparagus January 7, 2015 8:13 PM

These are known tells. Except when the subject is not neurotypical. As a youth, I played bridge most successfully. I was, apparently, exeptionally hard to ‘read’ by people who didn’t know me because my body language is abnormal. I was, however, using the game to learn how to read implicit signals people who were prohibited from communicating in a way other than that proscribed. There are a significant fraction of people who are wired differently and present ‘lying’ signals constantly, or whenever under social pressure.

Fast forward many years and it is apparent my tweenage daughter also presents differently. I can read her easily though and it drives her nuts. Won’t meet anyone’s gaze, but has tricks for appearing to do so, has tricks to enable (almost) unseen fidgeting.

This is quite prevalent amongst my friendship group as similar mindss congregate (insert favoured idiom here) and apparently breed. It’s great fun dealing with people who aren’t aware of autistic spectrum disorders and diagnosis deniers as their ‘intuition’ says ‘liar’.

I suspect what has been created here is not so much a lie detector but a neurotypical detector.

Andrew_K January 8, 2015 1:14 AM

@ Clive Robinson, good cop vs. bad cop scheme; being forced into a gimp suit

You’re right, that’s why the “good cop” interrogator tends to say things as “I have not much time, I will be removed from your case but I want to help you”. Notice how it is never the bad cop that is about to be given another case… moreover the bad cop horasses the good cop in front of the suspect for his “warm and cuddly moron methods” which are not working to his standards. Works even better if the bad cop is higher ranked than the good cop and the suspect knows about, so he can support poor little David against growling Goliath.

Whether being forced into a wired gimp suit can be considered torture may depend highly on how it is presented to the suspect and how he is literally put into it. When officers present it as mind-reading trick and use it to threaten the suspect — yes. If the test can be denied by the suspect without any negative consequences (pause here for laughs), no.

@ Bob Patrick
My candid sympathies. I did not intend to hurt with what I tought was a nice metaphor. Sadly it is none.

Paula Thomas January 8, 2015 2:46 AM

Not for the first time I am left wondering how this test would work with me. I have athetoid cerebral palsy which means that my body moves randomly all the time!! Also no method is perfect but this one, on their own data, appears to be wrong in nearly one in five cases. That is not good,

Andrew_K January 8, 2015 3:43 AM

@ Bob Paddock — please excuse my error in your name. One should not start typeing while still being somewhere else mentally. Won’t happen again.

GreenSquirrel January 8, 2015 6:48 AM

@Clive “I was once told by a researcher that people had it wrong about “bad cop good cop” it was not the fear of the bad cop or the hope the good cop would be nice. No it realy only works with those who have sufficient empathy that actually feel sorry for the good cop and don’t want him to look bad infront of the bad cop…”

This is what I was taught in the late 90s and even then it was generally regarded as something people had known since Tin Eye Stephen’s time.

@Oliver

“Any “research” into those things needs to be mocked senselessly, and repeatedly!”

Well put.

jim January 8, 2015 9:57 AM

First time a paper? Don’t think so unless it is ment as a security paper. As a psychology paper? It goes back to Frudo, durn you remember the 18 the century sex guy. No sex with mom dittohead. I remember his having proposed such a theory and another, I believe it was Skinner, who wrote a follow-up paper, but this was in the same time period that maslov was teasing his dogs with bells, and noted what the assistants were doing when punishing the dogs for inappropriate response, and others were doing the same experiments on people, and reporting the results.

albert January 8, 2015 11:47 AM

No response to my comment about not talking to the police? I hope that means you all know that, and I’m just stating the obvious. But if that’s the case, why are you all prattling on about interrogation?
.
“…the accused refused to answer any questions, but when we asked if he did it, his body movements indicated he was reacting…”;
“his body movements indicated he was lying”;
“His body movements indicated that he was guilty”. Yeah, that’s it.
.
So much for social ‘science’..
.
I gotta go…

Stranger January 8, 2015 3:05 PM

I’m glad this study is culturally-sensitive. I’d sure hate for lie detecting to be racist 🙂

Clive Robinson January 8, 2015 7:37 PM

@ GreenSquirrel,

This is what I was taught in the late 90s and even then it was generally regarded as something people had known since Tin Eye Stephen’s time.

Hmm I wonder if it might have been the same person that “told” me, as I’ve wondered befor, have our paths crossed…

@ All,

Col Robin “Tin Eye” Stephens methods are often described as original, but I suspect the seeds of them came from his non academic learning whilst away at school. Even today some of the schools that date back to that time can be fairly evil on children via fagging / bullying.

Tin Eye honed his skills on POWs in Latchmear House in what was then Surrey, just a few hundred yards south west of Richmond Park, a place I used to pass very frequently when young. However by that time it was not full of POWs but “delinquents” as post war it became a Borstal / Correctional facility for minors. The fences around it were around three stories high and fairly imposing in the choice of wire etc they were made of. Oddly although it’s only a couple of miles up the road, I’ve not been back that way since going to a party in the prison officers social club twenty years ago, so I’ve no idea if it’s even still there.

For those that are wondering why he was called Tin Eye it was because of his monacal, which looked like it had been made from the bottom of a “shot glass”. Apparently it was also joked that he had “a key in his back” because his behaviour was often like that of a “clock work soldier”. As they say “appearances can be deceiving” because behind the martinet exterior was a very acute mind. His methods were apparently rather more successful than you would imagine from hearing them described, and worked by getting inside the other persons head one way or another. I have a book or two in my dead tree cave where he and his groups activities get mentioned indirectly along with the work of Sefton Delmar’s black propaganda which used the aspidistra transmitter Bruce has blogged about in the past.

From what I can gather Tin Eye was very averse to the “rough em up” interrogation technique and all forms of physical torture, beliving that you only get to hear any old nonsense from the prisoner that will stop the pain for even a short while.

The results he got if reported historicaly correctly proved his point, thus with recent revelations from the CIA further indicating physical interrogation does not produce real intel, you would have thought the tortures trade would be on the wane. But apparently not, which suggests there is something not right in the heads of those who advocate it. It has been suggested that they get some kind of gratification power trip similar to that of rapists, which might account for some of the degradation methods used in Iraq that ended up with photos etc and eventual conviction of those in the photos, but not those who had devised the methods and ordered them to be carried out.

Clive Robinson January 8, 2015 7:47 PM

@ Albert,

No response to my comment about not talking to the police? I hope that means you all know that, and I’m just stating the obvious. But if that’s the case, why are you all prattling on about interrogation?

Possibly because we also know that –rather more than we would care to admit,– “not talking to the authorities” is not an option…

GreenSquirrel January 9, 2015 4:49 AM

@Clive

Re: Paths. Possibly crossed, especially if you were ever around the manor or associated buildings in the 90s or over the water. Either that or your peers taught me 😉

Tin Eye’s lessons and the resultant protocols were heavily covered on a lot of courses I took around this time and it has stuck with me ever since.

From what I can gather Tin Eye was very averse to the “rough em up” interrogation technique and all forms of physical torture, beliving that you only get to hear any old nonsense from the prisoner that will stop the pain for even a short while.

Very true and interesting that it was at a time when bombs were falling across London and the country was in “total war.”

Just to be clear, Tin Eye wasn’t a proponent of tucking up the enemy in bed at night and gentle massages with a happy ending – however he did strongly believe that one day the war would be over and everyone would have to live with each other and the things they had done.

At the risk of being massively off-topic now, I also want to point out that, having read the SOPs produced by Tin Eye, I dont agree with Ian Cobain’s generalisations – which are based on Col Scotland’s supposed memoirs anyway (although I have no knowledge of the Bad Nenndorf saga so that may be true).

Nicole Bienfang January 9, 2015 4:57 AM

I love reading about research like this. I wonder if there is any pattern that the test subjects fell into with their “full body motion” when compared to each other. I am however a bit confused when it comes to this line “Furthermore, movement was guilt-related, and occurred independently of anxiety”. Did the test subjects have a exit interview to determine if their movements was indeed because of guilt or was this just supposed by those running the experiment? How can feelings of guilt happen independently from anxiety, one would think they come hand in hand, no?

albert January 9, 2015 11:26 AM

@Clive
.
Here in the US, you are not required to answer questions by law enforcement(LE), without your lawyer present, and your lawyer will indicate what questions you may safely answer. You are not required to give a statement, either.
.
We ARE required to answer questions in a court trial, Senate hearing, or grand jury proceeding. IANAL, but ask any US lawyer.
.
It’s a fools errand to think you can fool a skilled interogator. If they want you (guilty of not), they’ll try anything to get you.
.
I gotta go…

vas pup January 9, 2015 12:18 PM

Many good postings, in particular I like:
Bob S. • January 6, 2015 5:55 PM
Ray Dillinger • January 7, 2015 2:09 AM
Oliver • January 7, 2015 10:53 AM.
General remarks:
(1) any such equipment is measuring stress level, not lie telling.
(2) for each person there is own base line to be established for comparison answers to neutral questions and ‘loaded’ questions.
(3) any lie detection could be utilized as intel seeking tool only to direct further LEAS actions (like red flags), but NEVER as evidence.

Alan January 9, 2015 7:05 PM

@albert

you’re missing the point entirely. that sh*t won’t fly if a terrorist tag is placed on a detainee.

Jorge January 11, 2015 5:08 AM

I have a weird neurotic trait – I have to shake my leg all the time, at least when sitting. I wonder how the system would perform on people like me.

Andrew_K January 12, 2015 1:42 AM

Regarding interrogation techniques (and indirectly Tin Eye): I wouldn’t leave any doubt that there is not much of a difference between physical and psychological roughness in interrogation (aside psychological force having the advantage of not leaving physical evidence).
I can understand those trying to differentiate it. It may provide peace of mind after giving order to keep a detainee awake day and night. I consider such behavior psychological self-protection.
Interrogators may have suffered the same experience in training without being damaged (or without realizing being damaged; allowing a “if I can bear with it, so will they!” mindset) — forgetting that training was in a pro-forma hostile environment. Malicious acting friends stay friends, your subconscious mind will keep that belief up and it will feed you through almost anything (and so does knowing how much time such training usually takes, in other words, that there will be a [happy] end).
That’s an important psychological difference to a detainee — who will be given a piece of “no one knows where you are and we keep you here as long as we want”. I think, that’s rough, even if those words are never said.

Oh well, another conversation that went from somewhere to the dark sides of HUMINT. Sorry for getting off-topic.

GreenSquirrel January 12, 2015 5:44 AM

@Andrew_K

I appreciate the off-topic is getting far enough here, and I am at least partly to blame, but I wanted to say that I actually agree with you.

I certainly don’t have a reasoned response as to why psychological torture is more acceptable than physical torture and I am reasonably convinced that most people would see the psychological techniques used in an interrogation session as torture to at least some degree.

You are also correct that the training does implant a mindset that “I didnt mind therefore..” and there is little to support this.

However, in defence of this, I would say that the really hard techniques frequently publicised (such as sleep deprivation, white noise etc) were originally aimed for the front line prisoner handling / tactical questioning teams who needed a fast, short duration, way to determine which prisoners went on to the interrogation centres. Sadly, as with all things, human laziness meant that the supposedly more specialised & skilled interrogators ended up doing things the PH/TQ way.

I was taught that the skilled interrogators didnt need sleep deprivation, hunger, fear to work their magic, they did it by understanding the captor, engaging with them, keeping control of the facts and building enough of a rapport that the subject let their guard down.

None of this ever happens when you are waterboarding them.

None of this removes the cognitive dissonance I have around the topic though.

vas pup January 12, 2015 3:38 PM

@GreenSquirrel • January 12, 2015 5:44 AM:
“I was taught that the skilled interrogators didn’t need sleep deprivation, hunger, fear to work their magic, they did it by understanding the captor, engaging with them, keeping control of the facts and building enough of a rapport that the subject let their guard down.” Just out of curiosity: Is this country where you were taught is Israel. I am asking because they understood importance of valid information obtained out of interrogation, and not just building cases to prove budgeting.

Leave a comment

Login

Allowed HTML <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre> Markdown Extra syntax via https://michelf.ca/projects/php-markdown/extra/

Sidebar photo of Bruce Schneier by Joe MacInnis.