The Journal of Credibility Assessment and Witness Psychology 2006, Vol. 7, No. 2, 108-115
The Journal of Credibility Assessment
and Witness Psychology
2006, Vol. 7, No. 2, pp. 108-115
Published by Boise State University
Emerging Methods and Measures for
Detecting Stress and Deception:
Thermal Imaging
Dean Pollina
Department of Defense Polygraph Institute,
Ft. Jackson, South Carolina
Copyright 2006 Boise State University and the Authors. Permission for non-profit
electronic dissemination of this article is granted. Reproduction in hardcopy/print format
for educational purposes or by non-profit organizations such as libraries and schools is
permitted. For all other uses of this article, prior advance written permission is required.
Send inquiries by hardcopy to: Charles R. Honts, Ph. D., Editor, The Journal of Credibility
Assessment and Witness Psychology, Department of Psychology, Boise State University,
1910 University Drive, Boise, Idaho 83725, USA.
108
The Journal of Credibility Assessment and Witness Psychology 2006, Vol. 7, No. 2, 108-115
Emerging Methods and Measures for
Detecting Stress and Deception:
Thermal Imaging
Major Points




Description of thermography
Data processing and algorithm development issues to be resolved
Mock crime studies – Some findings
Theoretical discussion and conclusions
Description of Thermography
109
The Journal of Credibility Assessment and Witness Psychology 2006, Vol. 7, No. 2, 108-115
Issues to be Resolved: Image Analysis




Which data transformations and Filters should we
use?
How should head movement be tracked?
How to determine rate of change (blood flow) in
visual images?
How best to standardize measurement locations,
sampling rates, radiation detected?
Issues to be Resolved: “Deception” Algorithms
Some Problems:
1. Individual variability and subpopulations

Differences in physiology

No standardized test formats
2. Ill-defined psychological construct
3. Little understanding of the physiological process

Can’t develop theoretical framework

Can’t constrain the infinite number of transformations possible
4. Either Ground Truth Unknown or Little Jeopardy
Issues to be Resolved: Methodological Problems
1. Comparison Question Test Format (Pavlidis et al., 2002; Pollina & Ryan, 2002)

Head movement not adequately tracked/controlled

Stimulus presentation times variable

Algorithm sensitive to study conditions
2. Study 2. Concealed Information Test Format (Pollina et al., submitted)

Camera Sensitivity (.10 C) not adequate

Head movement not adequately tracked/controlled
Issues to be Resolved: Test Question Sets
New technology will have to be tested

Mock Crime / Field Study

Ground Truth

Question Sets Used
110
The Journal of Credibility Assessment and Witness Psychology 2006, Vol. 7, No. 2, 108-115
Issues to be Resolved: What is the psychophysiological process?
Periorbital Region 1s after the presentation of a crimerelevant polygraph question.
Two broad categories
Core temperature (TC).
Regulated temperature (TREG ).
Issues to be Resolved: Head Movement
Algorithm Types:

General Purpose

Specific Purpose
Machine Vision:
Pattern Match Template
Head Movement:
1. In plane vertical
2. In plane horizontal
3. In plane rotational
4. Out of plane
111
The Journal of Credibility Assessment and Witness Psychology 2006, Vol. 7, No. 2, 108-115
Early Thermal Imaging Studies: Overview



Mock Crime: Murder
32 Participants, U.S. Army basic trainees
Simultaneous polygraph and thermal imaging of the face
Grand Averages: Skin Temperature
Ear
Mouth
10
5
10
1
5
0
0
-5
-5
-10
-10
Ne ck
Scalp
10
10
2
5
5
5
5
0
0
-5
-5
-10
-10
Eye
Nose
10
6
10
4
3
5
5
0
0
-5
-10
4
3
-5
-10
Deceptive
Nondeceptive
112
6
The Journal of Credibility Assessment and Witness Psychology 2006, Vol. 7, No. 2, 108-115
Regions Sampled
Mouth
1. Buccinator.
2. Depressor labii
inferioris.
3. Levator labii
superioris.
4. Mentalis
5. Orbicularis oris.
6. Depressor anguli
oris.
7. Zygomaticus
major and
minor.
9
8
11
1213
14
3
7
5
6
1
2
4
+ 60
Ear
8. Temporoparietalis.
Scalp
9. Frontalis.
Neck
10. Platysma.
Eye
11. Corrugator
supercilii
12. Orbicularis oculi.
Nose
13. Procerus.
14. Nasalis.
10
0
- 60
Mock Crime Study, ZCT: Grand Average Frame Means: Eye
Deceptive
Nondeceptive
Temperature Difference
4
2
0
-2
-4
-6
-8
1
2
-10
1
150
Frame Number
5
1
113
300
The Journal of Credibility Assessment and Witness Psychology 2006, Vol. 7, No. 2, 108-115
ZCT Data: Combining Polygraph and SST
Area Under the ROC Curve Derived from Binary Logistic Regression
R2 (Cox
& Snell)
.41
ROC
Area
.88
Sig. *
SST: Nose, Mouth, Eye,
Scalp, Neck, Ear
.09
.70
.09
(N.S.)
Polygraph and SST
Amplitude: Nose
BV, EDA, AR, TR, SST:
Nose
.49
.90
.001
Polygraph and SST
Amplitude: Eye
BV, EDA, AR, TR, Eye
.46
.90
.001
Polygraph and SST
Amplitude: Eye, Nose
BV, EDA, AR, TR, SST:
Eye, Nose
.52
.92
.001
Regression Analysis
Predictor Variables
Polygraph Measures
BV, EDA, AR, TR
SST Amplitude
Measures
*
.002
Null hypothesis: true area = .50
Irrelevant Items
100
CIT Results
80
Deceptive Group. Temperature change to
critical item prior to verbal response.
60
40
20
Maximum Temperature Difference
0
Nondeceptive
Deceptive
Relevant Item
100
Nondeceptive Group. Gradual increase in
temperature throughout the response
interval. Similar for critical and non-critical
items.
80
60
40
20
Maximum Temperature Difference
0
Nondeceptive
Response
Response
Response
Response
Response
Response
Deceptive
- 1.00 Sec to Response - .66 Sec
- .66 Sec to Response - .33 Sec
- .33 Sec to Response
to Response + .33 Sec
+ .33 Sec to Response + .66 Sec
+ .6 Sec to Response + 1 Sec
114
The Journal of Credibility Assessment and Witness Psychology 2006, Vol. 7, No. 2, 108-115
CIT Critical Item
-30 to - 20
-20 to - 10
Resp to +10
+10 to + 20
-10 to Resp
+20 to +30
Conclusions



Thermal imaging shows some promise, especially when combined with traditional polygraph
measures.
Questions such as optimal measurement sites, sampling rates, transformation algorithms, and
combination (with polygraph) strategies need to be developed and tested.
The long range goal is to use these measures to determine the specific emotions experienced by
examinees on-line, and to use this feedback as an aid in credibility assessment.
Published: 1 June 2006
115