PHENOMENOLOGICAL RESEARCH AND ANALYSIS (INTERIM TECHNICAL REPORT) (23 JULY - 30 NOVEMBER 1992)

• Document Type: CREST
• Collection: STARGATE
• Document Number (FOIA) /ESDN (CREST): CIA-RDP96-00789R003100230001-2
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 25
• Document Creation Date: November 4, 2016
• Document Release Date: May 22, 2000
• Sequence Number: 1
• Publication Date: November 30, 1992
• Content Type: RP

li?Aff1 R@se 2000/08/08: CIA-RDP96-00789R003100200I6?j-07-406-010
Phenomenological Research
and
Analysis
Edwin C. May, Ph.D., Wanda L. W. Luke, and Nevin D. Lantz, Ph.D.
0
Science Applications International Corporation
An Employee-Owned Company
Contract MDA908 - 91 - C - 0037
(Client Private)
Submitted by:
Science Applications International Corporation
Cognitive Sciences Laboratory
1010 El Camino Real, Suite 330, P.O. Box 1412, Menlo Park, CA 94025 ? [phone]
OthRr .91C O/ficF,.s;lllbdq~p~'q~p&.~p~t~rSCnlor~eh.Sw'jggq~Qaufon. FiM[H: ills-ltts.l(flaas_Las.Asy~ka,110c4ea 16O2106rdf2 falo Alto, Seattle, Tucson
 Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
Phenomenological Research and Analyhsis: Interim Report
TABLE OF CONTENTS
LIST OF FIGURES .................................................................. ii
LIST OF TABLES ..................................
I OBJECTIVE ................................................................ 1
II INTRODUCTION .......................................................... 2
III PROGRESS TO DATE ................................................. 3
6.1 Basic Research -. ........................................................ 3
6.2 Data Patterns/Correlations ................................................ 4
6.3 Applied Research ........................................................ 4
6.4 Theoretical Issues ......................................................... 9
6.5 Research Methodology ....................................... ........ 9
IV BIBLIOGRAPHY .......................................................... 12
V GLOSSARY ............................................................... 16
APPENDIX ........................................................................ 17
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2 i
 Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
Phenomenological Research and Analyhsis: Interim Report
LIST OF FIGURES
1. Response threshold (1.196) for target pack 1, coding bit 3.............................. s
2. A two-by-five bit, error correcting block code . ............................. .
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2 ii
 SRXr%jkogogica~l esearc7~ an/dBAnalyhsis Interim Repo 03100230001-2
LIST OF TABLES
1. Visual memabership values for target pack 1 ......................................... 6
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2 iii
 Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
Phenomenological Research and Analysis: Interim Report
1. OBJECTIVE
The objective of this document is to provide an interim technical report on tasks 6.2, "Basic Research,"
6.3, "Applied Research," and 6.4, "Research Methodology," as listed in the 1991 Statement of Work.
This report covers the time period from 23 July to 30 November 1992, and includes all subtasks.+
* This report constitutes the fifth and sixth deliverable DI-MISC-80508 under contract number MDA908-91-C-0037.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2 1
 Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
Phenomenological Research and Analysis: Interim Report
II. INTRODUCTION
Under Statement of Work items 6.2, 6.3, and 6.4 in contract MDA908-91-C-0037, the Cognitive Sciences
Laboratory of Science Applications International Corporation is tasked to conduct phenomenological re-
search and analysis. This document details the activity accomplished under these items and constitutes the
interim report covering the period from 23 July to 30 November 1992.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2 2
 Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
Phenomenological Research and Analysis: Interim Report
III. PROGRESS TO DATE
This section describes the progress to date for each of the subtasks listed under items 6.2, 6.3, and 6.4
(basic research, applied research, and research methodology, respectively). The section numbering
corresponds to the numbering in the Statement of Work.
6.1 Bastc Research
6.1.1 Biophysical Measurements
6.1.1.1 MEG and EEG Correlates
We have identified a considerable literature that describes the correlation between MEG and EEG
measurements. Since each of these techniques measure different aspects of brain activity, we would
expect correlations of some types of measurements and not for other types. Currently, a neuroscientist
on our staff is examining the pertinent literature and formulating a comparison for our specific types of
MEG/EEG measurements.
6.1.1.2 MEG Data Analysis
We are examining over 800 megabytes of brainwave data from two different analytical perspectives.
(1) Traditional Event-Related Desynchronization (ERD). A brain indicates that it is inattentive by
the production of alpha waves. It is well-known that the brain can be aroused by motor action (e.g.,
moving a body part such as a finger), cognitive process (e.g., mentally reviewing a word list), and
responding to an external stimulus (e.g., direct light flash). One of the brain's indicators of arousal
from these activities is called ERD (i.e., alpha power either vanishes or is sharply reduced). Thus,
if we can assure ourselves that the stimuli that we used in the MEG experiment constituted a AC
stimulus, it would be likely that they also would show an ERD.
To search for ERDs in the previous data, we filtered each 2-minute run for alpha activity (i.e.,
7-13Hz). As is published in the general ERD literature, we computed the power by squaring each
sample point. We performed an ensemble average for 500 milliseconds of prestimulus time and for
1,500 milliseconds of poststimulus period.
We are currently examining a variety of normalization techniques so that we will be able to combine
the data from run-to-run and across blocks. A combination across blocks of data will allow a very
sensitive determination of Whether ERDs are seen in the previous data.
(2) Wavelet Transform. In a recent article, Schiff demonstrated the use of wavelet mathematics to
identify transients in non-stationary time series data.* This technique appears to be straight-for-
ward, and we are currently adapting it to our MEG data.
Steven. J. Schiff, "ResolvingTime-Series Structure with a Controlled Wavelet'hansform," OplicalEngineering, Vol.31, No.11,
pp. 2492-2495 (November 1992).
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2 3
 Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
Phenomenological Research and Analysis: Interim Report
6.2 Data Patterns/Correlations
6.2.1 Sender/No Sender Analysis
We are very sad to report the recent death of Mr. Charles Honorton of the Psychophysical Research
Laboratories (PRL). Through the PRL, Mr. Honorton had been subcontracted to conduct a detailed
analysis of sender/no sender issues. This analysis was to include statistical evaluation methods and a
discussion of the implications regarding target-individual matches. Mr. Honorton was an invaluable
contributor to the field of anomalous mental phenomena research and will be greatly missed.
At the time of his death, we were working closely with Mr. Honorton in the analysis of the Ganzfeld
data; however, he had not conducted any experimental activity on his subcontract. We are currently
discussing the possibility that Dr. Robert Morris of the psychology department at University of Edin-
burgh will assume the tasks under a separate subcontract.
6.2.2 MEG and EEG Correlates with Anomalous Cognition
We have conduced a detailed literature search for MEG and/or EEG correlates with anomalous mental
phenomena, and identified 53 relevant papers, which can be found in the bibliography in Section IV.
We are currently examining these papers to provide a mini-meta-analysis on the topic area.
6.3 Applied Research
6.3.1 Long Distance Experiment
The objective of this experiment is to adapt a standard anomalous cognition (AC) experiment to a
forced-choice situation. An additional objective is to incorporate fuzzy set technology into a "crisp,"
two-by-five, error correcting block code to improve. AC detection.
6.3.1.1 Background
In the Spring of 1992, SAIC conducted a pilot experiment that was designed to explore the potential for
maximizing the reliability of AC responses through objective and rapid analysis. In this study, we re-
verted to using a dichotomous binary procedure as opposed to a fuzzy set technique. By carefully select-
ing the dichotomous elements, we could use standard block coding techniques to incorporate complete
single-error correction and a few two-fold error corrections as well. We used a message sending motif as
a test-bed for this kind of analysis.
Unfortunately, only one receiver demonstrated an effect size larger than 0.20 (i.e., 0.22) for evidence of
AC and no evidence of enhanced detection was seen.
One of the primary problems, which may have contributed to this result, was the nature of the target
pool. In an attempt to make the targets dichotomous within packs and at the same time inherently in-
teresting, targets within the pool ranged in scale from a panoramic scene of a cityscape to a photograph
of three chairs and thus possessed a large variety of potential target elements. Since receivers were told
in advance that the targets could contain absolutely any material, they were unable to censor their inter-
nal experiences, which may have resulted in enhanced intrinsic receiver noise. We define such a target
pool as possessing a large target-pool bandwidth.
Another problem was that each encoding bit in the block code was linked to only one percept (e.g., the
single target element of water). This exaggerated the importance of the chosen dichotomous elements.
For example, if a receiver failed to sense water in the target but managed to sense most other aspects of
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2 4
 Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
Phenomenological Research and Analysis: Interim Report
the target regardless of whether they were part of the bit structure, then the block code was not particu-
larly applicable.
In an AC application, a fundamental imbalance exists in this type of bit structure. The block coding
assumes that binary zero is "assertive." For example, when water is not indicated in the response, it is
equivalent to indicating that water is definitely not in the target. Unfortunately, it is possible or even
likely that unless a receiver specifies explicitly that water is not present, then the presence or absence is
indeterminate. The net effect is to render this type of discrete block coding invalid for AC applications.
Tb achieve our second objective (i.e., improved coding procedures), our new experimental protocol at-
tempts to correct the problems discussed above, so that potential enhancement of the detection of AC
may be optimized. The following modifications have been made.
? We used the National Geographic static target pool, which has been successful for many AC experi-
ments in the past. This pool appears to include enough basic elements to keep a receiver from gues-
sing, yet allow for some internal self editing to decrease the receiver intrinsic noise (i.e., intermediate
target-pool bandwidth).
? The sensitivity to discrete block coding has been reduced by using a number of fuzzy-set elements to
define each block coding bit. Thus, each bit will not rely upon a single percept, but rather represent
classes of different percepts.
We anticipate that these improvements will allow for a much improved detection of AC, and provide a
more sensitive test of whether error-correction can be successfully applied to AC detection.
Tb achieve our primary objective (i.e., apply AC techniques to a forced-choice circumstance), we used
an associative anomalous cognition (AAC) procedure. In this technique, AC targets are assigned one
each to a limited set of alternatives. For example, in message sending each of four possible messages are
assigned to a different National Geographic magazine photograph. In principle, a receiver provides his/
her impressions as in a standard AC trial, and if an on-site analyst, who is blind to the target choice, is
able to correctly match the response to the intended target, then the message, which was previously
associated with that photograph is "received" correctly.
6.3.1.2 Target Construction
To accommodate the error correcting portion of the protocol, we assigned each of the block coding bits
to five individual fuzzy set descriptors. These descriptors were developed under another program, but
are included here in the Appendix. Each target in the National Geographic pool had been consensus
encoded with regard to each descriptor's visual importance to the scene. For example, the element wa-
terfall would most likely be coded as a numerical value of one in a close-up photograph of Yellowstone
Falls, whereas it might only receive 0.2 for a distant view of Yosemite Falls. These assignments are called
membership values in fuzzy set parlance. A collect of membership values for a photograph constitute the
fuzzy set representation of the visual content of the scene.
Since the error-correction code that we are using requires us to construct the following four binary num-
bers: 00000, 01110, 10101, and 11011 (see Figure 2, on page 9), we identified five sets of descriptors for
each bit position in the block code word. Each target pack of four photographs possessed their own
unique set of 25 descriptors.
Consulting the historical fuzzy set target database, we summed the membership values for the descrip-
tors for each bit position in each of the 20 target packs. Let the sum of the membership values for a given
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
5
 Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
Phenomenological Research and Analysis: Interim Report
bit position be I. By inspection, we determined a threshold of 1.5 such that if I > 1.5, then the bit was
defined as one. If E < 1.5 then the bit was defined as zero. We illustrate the technique for target pack
one. Table 1 shows the fuzzy set descriptors and the result of the threshold analysis for the four targets in
a pack. For example, target three corresponds to the binary number 10101 (i.e., the shaded cells in Table 1).
That is, the sum of the membership values exceed the threshold for bit positions zero, two, and four.
Visual Membership Value
Block Bit Position
Descriptor
Target 1
Thrget 2
Target 3
Target 4
Wilderness
0.3
1.0
0.5
Urban
0.5
Hills
0.1
0.7
0.1
0
Rocky
0.3
0.5
Aligned
0.9
0.5
TOTAL:
0.3
1.0
2.0
BINARY:
0
0
X K,
#?:`?.zi?`E
Rise
0.1
0.7
Boundaries
0.7
M
Closed
0.3
0.2
Settlement
0.8
Vegetation
0.9
TOTAL:
0
2.8
0.9
2.1
BINARY:
0
Parallel lines
0.2
0.9
Repeat Motif
0.2
0.8
0.2
0.3
Striated
0.1
0.6
0.4
0.4
2
Weathered
0.7
Grainy
0.3
0.7
0.1
TOTAL:
0.8
2.3
2.0
0.8
BINARY:
0
1

1
Table 1.
Visual membership values for target pack 1.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2 6
 Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
Phenomenological Research and Analysis: Interim Report
Visual Membership Value
Block Bit Position
Descriptor
Target 1
Target 2
Target 3
Target 4
Valley
0.5
0.1
Mountains
0.4
0.6
Diagonal Lines
0.2
0.6
0.5
4
Vertical Lines
0.2
0.1
0.4
Horizontal Line
0.3
0.9
0.3
0.2
TOTAL:
0.5
1.2
1.8
1.8
t7BINARY:
0
0
' O C v 4)? 0~p
CL
yEo o C
a
N UE E O 3 'mc ?.-8
00000000
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
 Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
c
N
E
4) PL
CL 0
w ~- CL V >
U E
13 m I
a CL
"?0 a~ c
Ow i~ ?o 0
Z3 0> 0 'v
000000
m
E
III E
v 0
?y-
TI 0
m
E
aE o 1m'
000000000
m C m Y
Q~ U
7 ~'C 2.8 0
>? 0 co ag -0 , ,0 3
D00~000000
oa
0000
8 s g 8
DD
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
00
 Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2
C
E
? a c
Q V >
f-
!LL
U)
t7~
W
ON
a
000
on
IS ZAC,L GA g
000
N N N
yy y a
c
TA > .C .-
> O
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100230001-2