posted 09-07-2008 11:48 AM            

I'm waiting to see research on both of them. I'm still not convinced they're nothing more than occlusive plethysmographs displaying a few things going on. (That doesn't mean they don't work. I just think it still leaves too many questions.)

The farther one gets from the heart, the weaker the signal, and by the time you get to the finger tip, you're not getting much by way of blood pressure.

posted 09-08-2008 11:26 PM            

Firstly, although being very sensitive has its pluses it also poses some problems. The sensitivity sometimes causes the tracing to be unstable due to examinee hand tremors (e.g. shaking in their boots). Also, the slightest movements can askew the tracing.

Secondly, I have seen very infrequent divergences in the arm and finger cuff tracing. This could be due to several reasons (e.g. the distal placement of the finger cuff and differences in the two cuffs pressure. Again, this is really infrequent.

Lastly, it is not a real problem but my own concern with the finger cuff use along with the arm cuff, because I place it on the hand opposite of the arm cuff. So, when using the EDA, pulse oximeter, and finger cuff all opposite hand of the arm cuff, there are quite a bit of components on that hand of the examinee.

posted 09-09-2008 07:36 AM            

Which system do you use?

------------------
Ben

posted 09-10-2008 08:01 PM            

The publication bias is important to remember.

One way to investigate the validity of the finger cuff would be to use samples of confirmed cases and evaluate he concurrence of decisions or numerical scores. The methodological limiation of this approach is that it requires confirmation for those cases conducted with both cuffs.

It may sound odd to us field examiners, but a better way would be to evaluate the correlation between the raw data streams. This is easy with the Limestone system, because the raw data are readily available to any science-minded person. On the Lafayette, it requires manually turning on the raw data feature which is buried somewhere in the admin or user options. Stoelting also stores data in ways that are readily accessible to researchers. Axciton does not. This is unfortunate, because we have an archive of confirmed cases that is at risk for slipping into the dark ages if we never access the raw data. (Tracings on the screen are not the data. Tracings on the screen are simply pictures of the data. Data are numbers that can be entered into a spreadsheet or statistical software for analysis.)

Analyzing the correspondence between the thumb cuff and finger cuff does not require anything but simultaneously recorded data under live conditions. Calculate the difference of every successive data point, then calculate the difference between those difference scores for the Arm cuff and finger cuff. Use a simple ANOVA for the data from an entire chart. If you don't like ANOVA, you can even use a Pearson Correlation, using the difference scores within each tracing. At 25 or 30 cps, you'll have about 15,000 data points for each tracing, for a 5 or 6 minute chart. Do this for a sample of charts - say 30 nice looking charts, 30 moderately messy charts, and 30 ugly charts. First calculate the mean and standard deviation, and then the confidence interval for your estimates.

The statistical estimates and confidence interval will represent the correlation or correspondence between the arm and finger cuffs, under normal, difficult and ugly condictions. With a sufficiently high correlation, you can reasonably replace one cuff with the other. Weak or poor correspondence would indicate the need for caution surrounding any attempt to replace the arm cuff with the finger cuff. Of course, later replication would always be recommended.

You can see something like this approach in the work of Vernon Quinsey, a rearcher at Queens University, who developed two actuarial risk prediction instruments - the VRAG (violence risk assessment guide), and SORAG (sex offense risk assessment guide). VRAG and SORAG both depend on the Hare Psychopathy Checklist-Revised (PCL-R) (now in version 2) score as data point. This is a test within a test - its not that uncommon. The PCL-R requires time and expertise of its own. Quinsey et al. evaluated VRAG/SORAG score using an alternate scale, the Child and Adolescent Taxon (CATS) as an alternative to the PCL-R, and reported a Pearson correlation (r) score of .975. That's darn near perfect - meaning you can score the VRAG and SORAG with the CATS instead of the PCL-R-2. Why would you do that, you ask? Because the CATS is quick and dirty, after reading a case history, which you would do anyway, you can complete the CATS in about 2 minutes.

Done.

Of course, this does not mean the CATS is equivalent to the PCL-R, and the CATS is not a substitute for the PCL-R-2 in other uses. This only applies to the VRAG and SORAG.

r

------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)

posted 09-12-2008 08:15 AM            

------------------
Ex scientia veritas

posted 09-12-2008 09:35 PM            

I like your method of analysis better. My suggestion, above, pertains to evaluating correlation between the actual raw data. There is a good possibility they won't. However, your method references the measured transformation of the raw data, in ways directly related to polygraph scoring. It is possible the raw data don't concur well, while the measured data concur more effectively.

It would be wise to be careful with terms such as "blood volume." The term was in vogue for a while, among polygraph examiners, but lacks a clear construct definition. The actual volume of blood does not change. Any concept of volume in proximal or distal organs, or vessels and capilaries is nowhere near adequately operationalized - for example, how do we actually measure "volume." We don't. Blood volume is a fancy sounding term of convenience. Polyscore documents describe the use of a blood volume derivative, which sounds to me like a statement of description, meaning the cardio measurement is probably a composite of all or some of the things we humans consider interpretable or observable cardio features.

The risk is that we humans also have a tendency to fill in the blanks with mythology when we don't completely understand something. Hence the confusion about blood volume increase while the tracing decreases. It becomes tempting to narrate our own hypothetical understanding of the actual physiology. The problem gets worse when we neglect to actively account for the fact that we are engaging in conjecture about the physiological mechanisms. With P02 sensors, thumb-cuffs, and "occlusive plethysmograph" statements, we have to remember that we may not even be measuring blood pressure with those devices. Its better to go to the physiology textbooks when attempting to understand the physiological mechanisms. Do a google search on the term and you'll see articles related to veinous occlusive plethysmography. They seem to refer more to "blood flow," and not to "blood volume." Not surprisingly, they describe sometimes occluding (cutting-off) circulation. We don't do that. But yes its a plethysmography - a thumb plethysmograph. I don't think it's an occlusive plethysmograph. The arm cuff is an arm plethysmograph (occlusive when checking blood pressure, not in polygraphy). The pneumo tubes are plethysmographs (non-occlusive) for the upper and lower body areas. A plethysmograph is simply a device intended to measure volume changes via circumference.

All of this is simply yet-another example of the need for complete and accurate documentation about what our scientific tools actually do.

Now I want a thumb cuff.

r

------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)

posted 09-13-2008 07:53 AM            

Point taken.

I just don't like the term, because I think it's vague and prompts inaccurate assumptions.

Andreassi does note that BV is measured in proportional changes against a baseline. That is for the very reason I stated. BV is a vague construct.

We still have to be careful filling in the blanks with our imaginations. We still need documentation on the hardware, because with some signal transformations we could simply draw the data differently - with larger tracings indicative of reactions. At present, it is the reduction of tracing amplitude that tell us about a reaction. Unless we know exactly how the signal is obtained and transformed, our desire for understanding can easily lead us into some fictional representation.

r

------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)

posted 10-03-2008 08:39 AM            

That is the kind of stuff I look forward to on this forum. That truly is an eye-opener for me. Thanks, Barry!

------------------
Ben

posted 10-03-2008 08:57 AM            

------------------
Ben

posted 10-03-2008 10:10 AM         

That is one of the charts I was referring to in the thread on "Finger Cuff Pressure". I think Limestone is way out front on this one! My only concern is that the Limestone finger cuff seems a little fragile however, I have not managed to break it yet!

posted 10-03-2008 06:31 PM            

This is a "good" segment.

It seems intuitive that the arm and thumb cuffs might have differing vulnerabilities to adverse circumstances, as JB pointed out. Shaking hands is one of the potential issues. We might expect the data to vary differently under sub-optimal conditions.

This, of course, begs the question around whether or not we should be attempting to interpret sub-optimal data in the first place.

Today at the Utah Polygraph Association Conference, sponsored by the APA, Chris Faucett shows a prototype for a new-and-improved Lafayette thumb cuff, which looks like it has good potential. It is smaller and stiffer than the infant cuff, and was still comfortable at over 100mmHg.

Still, I would refrain from replacing the arm cuff until we know more about the performance of the thumb cuff when compared to the arm cuff.

.012

r

------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)