CNN: polygraph not a lie detector

[hgc]

At least that is the position of CNN's security analyst, Ken Robinson. He was being questioned yesterday by Anderson Cooper (I think) about the what might happen with Saddam's interrogation, and asked about the lie detector. Robinson said, and I'll have to paraphrase, "it is not effective," and "the polygraph does not detect lies; it detects [something about physiological changes]." He then suggested that analysis of body language, such as eye movements, is probably more effective.

This is surprising. Is the mass media's unquestioning acceptance of the polygraph-as-lie-detector starting to crumble? Or is this just one rogue CNN employee (or contractor)?

[pupdog]

Maybe they ought to borrow a scientologist's E-Meter.

[bug_girl]

This is correct. in most US jurisdictions polygraphs are not accepted as evidence in court. There are lots of ways to knowingly (and unknowingly) control the output.

It is however effective at use in interrogation ("the machine says you did!") and it makes for great television, since their evidentiary rules are.......looser


I'll let someone with a T1 connection provide the citation.

[bug_girl]

oh wait--i found a quick loading one.
Cornell Law Briefs

this is a case that occured after a major ruling byt the supreme court that polygraphs didn't meet the Frye test of "general acceptance of the procedures and methodology as reliable within the scientific community"

i think that's the frye test. Council?

[Jeff Corey]

There's no doubt that measures like the GSR, heart rate, BP, and all can detect physiological responses mediated by the sympathetic NS.
In many people, these measures can detect emotional responses to key words or phrases.
But others are harder to read because they either under- or over-react. So you get a lot of false positives and negatives.
I use a cheap GSR to let a class decide which number between 1 and 10 a volunteer is thinking of.
Their hit rate is pretty high.

These devices can be used in the treatment of anxiety disorders by providing biofeedback.

[Iamme]

What is GSR and NS, if I may ask?

[SteveGrenard]

Quote:

Originally posted by Iamme
What is GSR and NS, if I may ask?

Since you didnt direct your question to anyone in particular, Ill tell you:

GSR = Galvanic Skin Response

NS = Nervous System

[SteveGrenard]

The statement by CNN that polygraphs are not lie detectors is true. Polygraphs were never lie detectors. This is not news.

Polygraphs are multi-channel devices designed to collect, amplify and record physiological signals.

ECGs that simulatenously record multiple leads, up to 12 in fact, are polygraphs.

EEGs are polygraphs.

EMGs are polygraphs.

Up to 32 and more simultaneously collected channel polygraphs are used in medical diagnostics for a variety of purposes. They can be set up to record breathing, ECG, pulse, aterial oxygen saturation, multiple EEG channels, EMGs of limbs, and the movement of the eyes under the lids during REM sleep (EOG).
etc etc etc. There is nothing inherent in the definition of the term polygraph that makes them "lie detectors." This term was invented by people who claim they can use polygraphs, some particular simulatenously recorded channels, to detect deception.

[Iamme]

Thanks Steve. That enabled me to go back and reread Jeff'as post and get a handle on what he was talking about.

(Jeff)

I use a cheap GSR to let a class decide which number between 1 and 10 a volunteer is thinking of.
Their hit rate is pretty high.

-----------------------------------------------------------------------------
Jeff---can you explain the process, and why the hit rate would be high?

[hgc]

Quote:

Originally posted by SteveGrenard
The statement by CNN that polygraphs are not lie detectors is true. Polygraphs were never lie detectors. This is not news.

Polygraphs are multi-channel devices designed to collect, amplify and record physiological signals.

ECGs that simulatenously record multiple leads, up to 12 in fact, are polygraphs.

EEGs are polygraphs.

EMGs are polygraphs.

Up to 32 and more simultaneously collected channel polygraphs are used in medical diagnostics for a variety of purposes. They can be set up to record breathing, ECG, pulse, aterial oxygen saturation, multiple EEG channels, EMGs of limbs, and the movement of the eyes under the lids during REM sleep (EOG).
etc etc etc. There is nothing inherent in the definition of the term polygraph that makes them "lie detectors." This term was invented by people who claim they can use polygraphs, some particular simulatenously recorded channels, to detect deception.

Thanks for the detailed explanation. My purpose was not to point out the obvious. Maybe you know and I know that polygraphs are not lie detectors, but the general public impression in this country is that the polygraph is a lie detector. Hence the common interchangability of the terms.

My purpose was to celebrate that at least one on-air employee of a major media outlet has stated such. He did it in response to a question from the anchor, no less, asking him if Saddam would get a lie detector test (meaning polygraph test).

It's a good thing that courts won't accept polygraph tests as evidence, but they are still heavily used in criminal justice (police interrogations), in government security screenings and in private hiring practices. Perhaps if more of the media starts to speak the plain truth about polygraphs, then we won't have people like Aldrich Ames escaping scrutiny merely by passing a "lie detector" test.

[SteveGrenard]

Quote:

Originally posted by hgc
Thanks for the detailed explanation. My purpose was not to point out the obvious. Maybe you know and I know that polygraphs are not lie detectors, but the general public impression in this country is that the polygraph is a lie detector. Hence the common interchangability of the terms.

An unfortunate and erroneous interchangeability that gives polygraph devices an undeserved black eye.
Polygraph evidence:

medical/phsyiological data obtained by polygraph, including EEG, ECG and EMG are accepted in court cases all the time. But not for the detection of deception. Everytime one mentions to a patient or someone asking about such devices, they have to go into a lengthy explanation why the polygraph is not, per se, a lie detector and that we are not using a lie detector on them to record their physiological data, we're using a polygraph! Polygraph does not equal lie detector.

Quote:

It's a good thing that courts won't accept polygraph tests as evidence, but they are still heavily used in criminal justice (police interrogations), in government security screenings and in private hiring practices. Perhaps if more of the media starts to speak the plain truth about polygraphs, then we won't have people like Aldrich Ames escaping scrutiny merely by passing a "lie detector" test.

Agreed. The courts won't except them because they are not 100% reliable which is sort of ironic since courts accept other types of evidence which is not 100% reliable either. Heck people are convicted of murder without a body, w/o a weapon and w/o clear cut motive. I think its called cirucmstantial evidence but it is allowed.

The common 4 channel polygraph aka lie detector (heart rate, breathing rate, blood pressure and GSR) is subject to too many countermeasures to be reliable on the positive side and because some people under stress don't do well (even after test questions which validate the aberrations to be expected if someone is lying)and are made to look like their being deceptive. This can throw off an investigation rather than be useful for it. And yes, accepting them as evidence that an Aldrich Ames type is not spying for the enemy is ridiculous as well.

Researchers are working on other physiological means for detecting deception. Check out

www.brainwavesciences.com

for one such interesting effort.

[SteveGrenard]

Quote:

Originally posted by hgc
Thanks for the detailed explanation. My purpose was not to point out the obvious. Maybe you know and I know that polygraphs are not lie detectors, but the general public impression in this country is that the polygraph is a lie detector. Hence the common interchangability of the terms.

An unfortunate and erroneous interchangeability that gives polygraph devices an undeserved black eye.
Polygraph evidence:

medical/phsyiological data obtained by polygraph, including EEG, ECG and EMG are accepted in court cases all the time. But not for the detection of deception. Everytime one mentions to a patient or someone asking about such devices, they have to go into a lengthy explanation why the polygraph is not, per se, a lie detector and that we are not using a lie detector on them to record their physiological data, we're using a polygraph! Polygraph does not equal lie detector.

Quote:

It's a good thing that courts won't accept polygraph tests as evidence, but they are still heavily used in criminal justice (police interrogations), in government security screenings and in private hiring practices. Perhaps if more of the media starts to speak the plain truth about polygraphs, then we won't have people like Aldrich Ames escaping scrutiny merely by passing a "lie detector" test.

Agreed. The courts won't except them because they are not 100% reliable which is sort of ironic since courts accept other types of evidence which is not 100% reliable either. Heck people are convicted of murder without a body, w/o a weapon and w/o clear cut motive. I think its called cirucmstantial evidence but it is allowed.

The common 4 channel polygraph aka lie detector (heart rate, breathing rate, blood pressure and GSR) is subject to too many countermeasures to be reliable on the positive side and because some people under stress don't do well (even after test questions which validate the aberrations to be expected if someone is lying)and are made to look like their being deceptive. This can throw off an investigation rather than be useful for it. And yes, accepting them as evidence that an Aldrich Ames type is not spying for the enemy is ridiculous as well.

Researchers are working on other physiological means for detecting deception. Check out

www.brainwavescience.com

for one such interesting effort.

[hgc]

Quote:

Originally posted by SteveGrenard

...
Agreed. The courts won't except them because they are not 100% reliable which is sort of ironic since courts accept other types of evidence which is not 100% reliable either. Heck people are convicted of murder without a body, w/o a weapon and w/o clear cut motive. I think its called cirucmstantial evidence but it is allowed.
...

This is incorrect. It's NOT because they are not 100% reliable. It's because there is insufficient scientific basis to the claim that they are lie detectors. You need read no further than bug_girl's 2nd post (and the link therein) in this very thread.

[SteveGrenard]

Bug Girl's Post#1: "This is correct. in most US jurisdictions polygraphs are not accepted as evidence in court. There are lots of ways to knowingly (and unknowingly) control the output."

Bug Girl's Link #2: Inadmissible based on defendent's state of mind when he took the lie detector test. The state of mind of a subject can result in #2 or #1 result below. The device cannot control for this. If you want to say this is a scientific basis fine.

hgc's post: "This is incorrect. It's NOT because they are not 100% reliable. It's because there is insufficient scientific basis to the claim that they are lie detectors. You need read no further than bug_girl's 2nd post (and the link therein) in this very thread."

Now my reply:

First of all 100% reliability is not obtainable with anything. Secondly, the lack of even near 100% reliability is partly due to the reason clearly stated by bug girl but in different language, to wit: "There are lots of ways to knowingly (and unknowingly) control the output."

People are convicted of crimes including capital crimes based on less than 100% reliable evidence. But not lie detectors.
I agree it would be ridiculous to convict or exonerate someone solely on the basis of this single "test" so therefore it is a good thing that they are not admissible. They are an investigative tool and a poor one if the results are erroneous.

Frankly hgc I didn't see anything in her post about scientific proof but that's okay. Scientific proof is important and it
would be based on proof that would not be subject to controlling the output which is another way of saying that the 4 channels
now used in polygraph testing are subject to countermeasures which would knowingly control the output. People taking
lie detectors can take medications, can learn to control their responses and pathological liars could (unknowingly?) easily
subvert or control the output. On the other hand stressed out, nervous people who do well on control question tetsing
(a technique known as CQT) may still "fail" or be made out to be deceptive. There is scientific proof that the output can
be knowingly or unknowingly controlled, thus making the device not 100% reliable. There are three types of results that
reuslt from a polygraph examination:

1. People who lie and the machine doesn't catch it (e.g Aldrich Ames)
2. People who tell the truth and the machine makes it look like their are being deceptive (e.g. innocent parties)
3. People who lie and the machine does catch it (e.g. truly guilty parties)

hence it is not 100% reliable. But number 3 does occur.

[hgc]

Ay carumba!

You say polygraphs are not admissible because their not 100% reliable.

Quote:

Steve Grenard:

The courts won't except them because they are not 100% reliable which is sort of ironic since courts accept other types of evidence which is not 100% reliable either.

I say, not correct, it's because there's no scientific basis, and that bug_girl already pointed that out.

Quote:

hgc:

This is incorrect. It's NOT because they are not 100% reliable. It's because there is insufficient scientific basis to the claim that they are lie detectors. You need read no further than bug_girl's 2nd post (and the link therein) in this very thread.

Now you want to turn around and lecture to me that 100% reliability isn't obtainable? Well, no friggin' duh. That's why I corrected you in the first place.

Now look at bug_girl's post again.

Quote:

bug_girl:

this is a case that occured after a major ruling byt the supreme court that polygraphs didn't meet the Frye test of "general acceptance of the procedures and methodology as reliable within the scientific community"

That line about "[not meeting ... a test of] general acceptance of the procedures and methodology as reliable within the scientific community;" that's what I mean by "no scientific basis." Gee, the word "scientific" appears in both places.

All of this is really minor quibbling, but that you are starting to get on my nerves.

[SteveGrenard]

Polygraphs are 100% reliable; lie detectors are not 100% reliable. I suggest that they do not meet the Frye test of because they are not 100% reliable. In fact different lie detector operators have gotten different results with the same subject. The irony, once again, is that there is no Frye test for an absent corpse, a missing murder weapon or an indeterminate motive or circumstantial evidence. There are many people on death row and jailed for life who may be innocent because of this.

There is a scientific basis for polygraphs, but not for lie detectors. If you doubt this you are throwing out every facet of electroneurodiagnostic, cardiological and electromyographic technology there ever was. Every EEG, ECG, EMG, EOG, apnea recorder, etc etc etc collect and record data by means of a polygraph.

[hgc]

Quote:

Originally posted by SteveGrenard
Polygraphs are 100% reliable; lie detectors are not 100% reliable. I suggest that they do not meet the Frye test of because they are not 100% reliable. In fact different lie detector operators have gotten different results with the same subject. The irony, once again, is that there is no Frye test for an absent corpse, a missing murder weapon or an indeterminate motive or circumstantial evidence. There are many people on death row and jailed for life who may be innocent because of this.

You should have had your Wheaties this morning; you are tripping all over yourself. Compare the first line of the above quote with another quote from you in this thread.

Quote:

First of all 100% reliability is not obtainable with anything.

Now, you're not telling me that all the stuff that the polygraph is measuring is really 100% reliable, are you? Because that's just ... unlikey and unproven.

Frankly I don't have a lot of interest in what the polygraph does (not that I necessarily think it's worthless), but that it would be great if people realized it's not a lie detector. If it does have some other practical use, then groovy.

Once again, the Frye test is not testing for 100% reliability, nor is any sensible measure of sufficient scientific evidence. So all your bemoaning of the use of circumstantial evidence is really beside the point. Circumstantial evidence can be very good and very compelling. Not all evidence is scientific evidence. A jury can reasonably infer that a person is dead, absent a body.

[Jeff Corey]

Quote:

Originally posted by Iamme
I use a cheap GSR to let a class decide which number between 1 and 10 a volunteer is thinking of.
Their hit rate is pretty high.
-----------------------------------------------------------------------------
Jeff---can you explain the process, and why the hit rate would be high?

I show them how the galvinometer works by putting on the finger electrodes and showing how the needle responds about 3 sec after I slap myself in the thigh. Then I recruit a volunteer, ask her to think of a number between 1 and 10, write it down privately and seal it into an envelope so no one knows what it is.
Then she sits in a desk in front of the class with the electrodes on, with the meter facing the class. She can't see the meter, but I can. I have to continually adjust zero and sensitivity. (Expensive meters are self-adjusting).
The class is told to remain silent, as is the volunteer.
After the readings settle down, I say the numbers in a random order and note the magnitude of the GSR, repeating the process until I think there is a clear outcome.
Then I ask the class. If we agree, the envelope is opened. In about 18 out of 20 times the class is right.

[hgc]

Quote:

Originally posted by Jeff Corey

....
Then I ask the class. If we agree, the envelope is opened. In about 18 out of 20 tihes the class is right.

That's a very cool experiment. Question though: who's more reliable - the meter or the class? Doesn't seem to me that you're measuring that, but your opinion?

[Jeff Corey]

Quote:

Originally posted by hgc
Question though: who's more reliable - the meter or the class? Doesn't seem to me that you're measuring that, but your opinion?

That isn't the question. The class and I are judging which digit elicts the greatest GSR from the meter. That one is usually the secret number.The meter is just sitting there and reading skin resistance.

[Iamme]

Jeff---What is going on in (or on) the finger, for the meter to know?

[Jeff Corey]

Quote:

Originally posted by Iamme
Jeff---What is going on in (or on) the finger, for the meter to know?

The sympathetic branch of the autonomic branch of the peripheral system reacts to stressors, like pain. This gives rise to a variety of measurable physiological behaviors. Like sweating.

So you go to Radio Shack and get an ohmmeter, figure out how to amplify the output, and attach two eletrodes with contacts made of aluminium foil to 2 adjacent fingers.

As sweat is secreted, the resistance (since we're talkin' DC here) decreases. Salt water conducts electricity better than dry skin.

(All material copywrote 2003 by Jeff Corey. Any quotes used for educational purposes must not exceed 500 words and must include a proper citation.)

Edited to add, don't forget the duct tape used to make sure the eletrodes adhere tightly and to hold the hand down on the desk.

[BillHoyt]

Quote:

Originally posted by SteveGrenard
Polygraphs are 100% reliable; lie detectors are not 100% reliable.

Fabulous claim, Steve! 100% reliable? No systematic error? No random error? No + / - figures for the measurement accuracy? The perfect measuring device? Please provide evidence for these specs, Steve. Give us the links to all the manufacturer's specifications. I mean, all of them, after all: "100% reliable." I really want to see them. They should look like this:

Operating temperature: any
Measurement range: - INF to + INF
Precision: infinite
Accuracy + / - 0.00...
Pen nib width: 0.00...

I'm looking forward to reviewing these specs.

[Jeff Corey]

Quote:

Originally posted by BillHoyt
Fabulous claim, Steve! 100% reliable? No systematic error? No random error? ...

Stop being picky, Bill. He meant to say "sort of reliable", but he didn't edit his post in time.

[SteveGrenard]

You caught me. I should have said 100% reliable when used according to specs such as between +5 C to +40 C. Or RH at 30% to 75%.

Polygraphs used as lie detectors are NOT reliable even when used according to specs. Are you saying that they are if they are used according to specification. If so, I disagree. Polygraphs are reliable for the purpose of collecting and recording physiological signals if used according to spec. If a subject is in V-Tach while hooked up to a polygraph measuring the ECG you can be 100% certain that the patient needs to be resuscitated. If his heart rate jumps from 90 to 100 when asked a question you cannot be certain at all he is lying.

(PS: Pen nib witdths can be adjusted virtually now since most polygraphs are now computer based.) I havent used a paper chart recorder in close to ten years but they are around. Luddites just can't throw them out in favor of new technology.

[hgc]

Quote:

Originally posted by SteveGrenard
You caught me. I should have said 100% reliable when used according to specs such as between +5 C to +40 C. Or RH at 30% to 75%.

Is that your final answer? Are you sure you want to stick with that?

Quote:

Polygraphs used as lie detectors are NOT reliable even when used according to specs. Are you saying that they are if they are used according to specification. If so, I disagree. Polygraphs are reliable for the purpose of collecting and recording physiological signals if used according to spec. If a subject is in V-Tach while hooked up to a polygraph measuring the ECG you can be 100% certain that the patient needs to be resuscitated. If his heart rate jumps from 90 to 100 when asked a question you cannot be certain at all he is lying.

You must be kidding. No one in this thread has averred that polygraphs are reliable lie detectors.

[Ed]

Quote:

Originally posted by SteveGrenard


First of all 100% reliability is not obtainable with anything.


Polygraphs are 100% reliable; lie detectors are not 100% reliable

You can't make this stuff up.

[Ed]

Quote:

Originally posted by Jeff Corey


As sweat is secreted, the resistance (since we're talkin' DC here) decreases.

[BillHoyt]

Quote:

Originally posted by SteveGrenard
You caught me. I should have said 100% reliable when used according to specs such as between +5 C to +40 C. Or RH at 30% to 75%.

Okay, so those specs you still haven't provided a link for should give an operating temp and RH range other than any. But they still claim infinite precision and +/- 0.00... accuracy? Where are those specs?

Quote:

Polygraphs used as lie detectors are NOT reliable even when used according to specs. Are you saying that they are if they are used according to specification. If so, I disagree. Polygraphs are reliable for the purpose of collecting and recording physiological signals if used according to spec. If a subject is in V-Tach while hooked up to a polygraph measuring the ECG you can be 100% certain that the patient needs to be resuscitated. If his heart rate jumps from 90 to 100 when asked a question you cannot be certain at all he is lying.

They never break down? Never need calibration? Are perfectly precise and perfectly accurate? They have an infinite MTBF? and a 0 MTTR? Where are those specs?

Quote:

(PS: Pen nib witdths can be adjusted virtually now since most polygraphs are now computer based.) I havent used a paper chart recorder in close to ten years but they are around. Luddites just can't throw them out in favor of new technology.

Kewl! So this magical computer-generated line can be adjusted to 0 width and still be visible? The A/D converter on the input end has an infinite number of bits and a 0.00... conversion time? Wow! Where are those specs?

[Cynical]

Take it easy, Billy Boy. You're beginning to sound like Claus Larsen with your demands for evidence. No sense in having two people here going over the edge.


Chill out, Billy.

[BillHoyt]

Quote:

Originally posted by Cynical
Take it easy, Billy Boy. You're beginning to sound like Claus Larsen with your demands for evidence. No sense in having two people here going over the edge.


Chill out, Billy.

If you make a ridiculous claim, you need to provide evidence. "100% reliable" was ridiculous. If he has the evidence to marshall, he should marshall it.

Now, ten-watt, do you have any substantive to add?

[Brummbar]

All "Lie Detectors" are fraudulent because they do not, in fact, detect lies.

They detect anxiety in the subject (to use simple terms).

Whether the testing subject's anxiety is actual, inferred, assumed and to what degree it can be quantified are all irrelevent.

In order for Lie Detectors to actually work, there would have to be a distinct, quantifiable physical manifestation of conscious dishonesty that ONLY occurs when a person is lying and NEVER occurs for other reasons.

I suggest calling this non-extant phenomenon the "Pinocchio Response."

Since there is no Pinocchio Response, the only way Lie Detectors could work would be to literally read the subject's mind.

[SteveGrenard]

First of all 100% reliability is not obtainable with anything.


Polygraphs are 100% reliable; lie detectors are not 100% reliable




--------------------------------------------------------------------------------




Ed: You can't make this stuff up.


Response: I amended the above statement to say if used according to spec, say for ECG or EEG or EMG they are 100% reliable and if thery are not broken. Obviously. They are collecting an electromagentic physiologic signal. The signal is amplified, filtered if nec and recorded and then can be read by the human eye. If a human does not use proper filters, proper gains or makes an improper diagnosis they are 100% doomed to failure. However, if they are set-up correctly, used according to spec and if the electrodes are placed according to international standards such as the International 10-20 System then they are 100% reliable. There are lots of factors that infleunce such devices which is why they need to be used according to spec.

Also think hard about the term 100% reliable where physiological measurements are concerned and where the use of polygraphs as lie detectors is concerned. The ECG or brain wave trace you see, if collected properlyand at proper (adjustable) settings is 100% reliable. What you see is what you get. You cant get a flat line on a functioning, properly set-up and electroded (patient) unless they are in cardiac arrest (asystole). You can't get V-tach or VF or see PVCs unless they are there. There is no + or - accuracy factor. If used as a lie detector it is far from this.

[hgc]

Quote:

Originally posted by SteveGrenard
blah blah blah ... What you see is what you get ... blah blah blah

You might be on to something here. What you see is what you get, but that ain't the same as what is.

[SteveGrenard]

Quote:

Originally posted by hgc
You might be on to something here. What you see is what you get, but that ain't the same as what is.

This statement just proves you dont know squat about the collection of phsyiological signals using a polygraph. A signal is a signal is a signal and it is what it is. What it isn't, however, is a means to detect deception. I am sorry if you feel otherwise. So many people defending this concept
while castigating the true and valid use of polygraphs is amazing.

[TLN]

Quote:

Originally posted by SteveGrenard
So many people defending this concept while castigating the true and valid use of polygraphs is amazing.

And just what is the "true and valid" use for a polygraph?

[SteveGrenard]

Quote:

Originally posted by TLN


And just what is the "true and valid" use for a polygraph?

excuse me, but geezus:

12 lead electrocardiography
electroencephalography - multi lead
electromyography - many muscles
electrooculography-up to 4 points (eyes)
continuous oxygen saturation
continuous rate and depth of respirations
chest wall motion - rate and depth
abdominal wall - rate and depth
continuous end tidal CO2

and about anything else that produces an analog signal that can be converted to a digital one

[TLN]

Quote:

Originally posted by SteveGrenard


excuse me, but geezus:

12 lead electrocardiography
electroencephalography - multi lead
electromyography - many muscles
electrooculography-up to 4 points (eyes)
continuous oxygen saturation
continuous rate and depth of respirations
chest wall motion - rate and depth
abdominal wall - rate and depth
continuous end tidal CO2

and about anything else that produces an analog signal that can be converted to a digital one

Well excuse me, but that doesn't answer my question at all. Try to get some verbs in there and try again.

What's the true and vaild use for a polygraph? What does it detect and what usefullness does it have?

[TLN]

Steve, just say you don't know. It's okay, really.

[SteveGrenard]

I really don't suffer fools easily. Excuse the baby talk:

it detects brain waves -- this includes diagnostic indicators that determine sleep stage, seizure activity and even the presence of tumours.

it detects heart waves; this includes normal heart waves, abnormal heart waves also known as arrhythmias and
it can also diagnose a number of different types of conditions under the umbrella of heart block. It detects potentially fatal arrhythmias such as ventricular tachycardia and ventricular standstill (also known as asystole).
In short: cardiac arrest but people can have 10 second or less asystoles which are anormal but are prognostic of problems.

it detects defects in muscle activity which means it can be used to diagnose a number of neuromuscular diseases

using evoked potential, auditory, visual and even small electric charges, it can determine if there is a normal or abnormal response to these. Preprogrammed flashing strobe like lights can elicit epileptiform seizure activity seen on the polygraph measuring brain waves in people with epilepsy. Some kinds of epilepsy does not manifest with outward convlusive seizures, only internalized seizure activity in the brain. It is very damaging and dangerous and needs to be treated and this is the only way to diagnose it.
Particularly in children.


when electrodes are placed near the eyes, it can record and diagnose rapid eye movements while asleep (REM sleep) as well as abnormal eye movements anytime including while awake

if your breathing becomes obstructed partially it can detect the presence of an abnormal respiration known as a hypopnea

if your breathing becomes completely obstructed, even for short periods while asleep (sleep apnea), it can detect the presence of an abnormal respiration which is a flat line and no respiration at all -- its called an obstructive apnea

if the chest wall and abdomen also decrease or go flatline, the apnea is known as a central apnea: these are due to drugs and to some kinds of neuromuscular disorders

if your oxygen and carbon dioxide levels in the blood become abnormally low, even transiently, it records that so that measures can be taken to correct them if the swings are wide, sustained and potentialy harmful

I hope the above helps. These are all valid and legitimate medical uses for polygraphs.