CIA-RDP96-00787R000300100001-9

n
SRI International
. Interrationci :
Approved For Release 2000/08/10 : CIA-RDP96-00
7R000300100001-9
1
I
L
Final Technical Report
Covering the Period 1 October 1986 to 30 September 1987
December 1987
ENHANCED HUMAN PERFORMANCE
INVESTIGATION (U)
1
L
SRI Project 1291
C
Copy
4 of
10. Copies.
This document consists of 53 pages.
SRT/GF-0308
Copy 1083
SFRobroved For Release 2000/08/10 :NCIA-RDP96-00787R650366100001-9
(415) 326-6200 . Cable:
SRI INTL MPK . TWX:
910-373-2046
-----
T
L

1
Approved For Release 2000108/10 &'s SPFPEP9787R000300100001-9
TABLE OF CONTENTS (U)
LIST OF TABLES
iv
LIST OF FIGURES
iv
I
INTRODUCTION
1
A.
Overview
1
B.
Definitions
1
C.
Program Scope
1
D.
Program Objectives
2
E.
Program Resources
2
II
PROGRESS TO DATE
4
A.
Status of Subcontracts
4
B.
Status of Consultants
5
C.
Progress to Date for Each Objective/Task
5
III
PROBLEM AREAS
45
IV
ADMINISTRATIVE COMMENTS
46
V
PROJECT MILESTONE CHART
48
1
VI
COST SUMMARY
50
REFERENCES
53
APPENDIX A - A POSTERIORI ASSESSMENTS OF THE SCIENTIFIC
OVERSIGHT COMMITTEE
A-1
APPENDIX B - PHYSIOLOGY CONFERENCE LETTERS
B-1
-
iii
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

1
1
Approved For Release 2000/08/12 &'s SPF?EPS787R000300100001-9
LIST OF TABLES (U)
1.
Status of Subcontracts for FY 1987
4
2.
Status of Consultants for FY 1987
5
3.
Priority/Deliverable Assignments for FY 1987
46
4.
Authorized Task Changes to FY 1987 Statement Of Work
47
5.
Authorized Interpretations of FY 1987 Statement Of Work
47
6.
Enhanced Human Performance Investigation -- FY 1987
48
7.
Cost Analysis By Objective/Task -- FY 1987
51
N
LIST OF FIGURES (U)
1.
The Attribute Set for the DACOS System
8
2.
DACOS Hierarchy for Water-Present Targets
9
3.
DACOS Hierarchy for Water-Absent Targets
10
4.
Hypnosis-RV Protocol
22
1
iv
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

-
1
Approved For Release 2000/08/10 .: CIA-RDP86-00787R000300100001-9
UNCLASSIFIED
(U)
University of Delaware, plus the consultants having expertise in specific areas of interest to the
program.
1
3
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

Approved For Release 2009/08
UNICLASSIFIEDS 787R000300100001-9
II PROGRESS TO DATE (U)
(U) For this reporting period (1 October 1986 to 30 September 1987), our primary
progress was made in the areas of pilot and formal experimentation in RV and RA.
A. (U) Status of Subcontracts
(U) Table 1 shows the current status of the subcontracts for FY 1987. For administrative
purposes, it was convenient to use a number of different types of contractual agreements:
· Purchase Order - Used for non-research-oriented deliverables.
· Consultant Agreement - Used for a single individual within a large
organization.
· Services Contract - Used for contracts having total funding of less than
$100,000.
SGFOIA2
. Full Subcontract - Used for contracts let by prime contractor having total
funding of greater than $100,000.
4
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

Approved For Release 2000/08/12 &''s STPPEP787R000300100001-9
SGFOIA2
C. (U) Progress to Date for Each Objective/Task
1
(U) The progress to date for each Objective and Task in the Statement of Work is
described below.
1. (U) Objective A, Task 1 -- Statistical Protocols and Research Design
-
(U) All of the formal experiments have been reviewed by the Scientific Oversight
Committee (SOC) during FY 1987.
2. (U) Objective A, Task 2 -- A Posteriori Assessments
(U) On November 12-13, 1987, the available members of the SOC met to review the
FY 1987 work. Their comments and, where appropriate, SRI's responses are included in
Appendix A.
5
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

Approved For Release 2000/08/10 .: (
UNCLASSPEPE £9787R000300100001-9
3. (U) Objective A, Task 3 -- Improve RV Evaluation
a. (U) Fuzzy Set Applications in Remote Viewing Analysis
(U) In FY 1987, fuzzy set mathematical techniques were applied to the
problem of remote viewing analysis. Two analytical methods were developed: the first was
designed to be sensitive to the verbal content of the RV response; the second was designed to
account for the visual/spatial arrangements of response elements. A definition of "ground
truth," against which these new analytical techniques could be tested, was also devised.
(U) The verbal method is predicated on the application of fuzzy set
mathematics to the figure of merit (FM) technology .* The method also features a new
descriptor list, which was introduced to provide a richer vocabulary for analysis. The list's
hierarchical structuring in levels, ranging from very abstract to very concrete, affords
considerable flexibility for analytical manipulation of descriptor elements. A pilot application of
the verbal analysis was shown to correlate highly with ground truth.
(U) The combination of fuzzy set technology and the new descriptor list also
proved effective for the visual/spatial approach. The implementation of these techniques -- in
conjunction with a third technique known as "cluster analysis" -- has resulted in an algorithm for
the production of orthogonal target sets. This has resulted in a significantly more effective
rank-order analysis procedure.
(U) The visual and verbal analyses were each determined to have certain
strengths and weaknesses. The verbal analysis is statistically more powerful and provides a more
comprehensive breakdown of the verbal information in an RV response. It is quite
labor-intensive to apply, however, and it appears to be relatively insensitive to noisy RV data.
"Noisy," in this context, can be defined as a preponderance of incorrectly identified response
elements. The visual analysis system is statistically much less powerful and is less capable of
providing systematic objectification of the true RV signal content. It can be rapidly applied,
however, and is sensitive to the primary manifestation of true RV signal in noisy data -- namely,
the visual arrangement of RV response elements, regardless of their verbal labels. Potential
applications of these techniques in their current states have been suggested; areas of future
research for their refinement have also been identified.
*
(U) The FM analysis has continued to undergo refinement since its inception in FY 1984.
6
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
b. (U) An Expert System Approach to Remote Viewing Analysis
(1) (U) Motivations to Explore Expert Systems
(U) The judging of RV transcripts has proven to be a difficult task for
the experienced as well as the uninitiated analyst. Judging can be both analyst and viewer
dependent, in that the combination of a viewer's response style and an analysts's interpretive
style may enhance or hinder the analytic task. The process is, to a large extent, a subjective task
that does not lend itself to a literal or procedural quantification. In an effort to render the
judging more transferable, if not more uniform, we decided that, if development were feasible,
an expert system to assist the analyst would prove invaluable.
(U) The ultimate task would be to develop an expert system that could
ask an analyst a series of questions about a given RV transcript, arrive at a composite description
of the response, and map the response to a group (possibly with only a single member) of targets
within a known target pool. The system would have standard data about a number of common
abstract and concrete objects. Furthermore, the system would maintain a data base of the
ideograms and idioms commonly used by a particular viewer and their possible/probable
correspondences. By prompting the analyst for information about the concrete or unambiguous
elements of a response (i.e., are there parallel lines, or are there elements labeled as structures),
the system would combine the user-supplied data with data collected in previous experiments
with the same viewer, to piece together composite hypotheses about the transcript. In an
interactive exchange, the system would attempt to present the analyst with possible transcript
interpretations of increasing complexity and/or concreteness until some kind of composite picture
could be drawn.
(U) Clearly, such an undertaking is very ambitious and well beyond our
current expertise. Acting as a consultant, Dr. Jacques Vallee was to undertake the initial steps
towards the development of such a system. As requested, we supplied him with the. NExpert®*
development system, an expert system shell which exploits non-monotonic reasoning (i.e.,
-
simultaneous forward and backward chaining of, analogously, simultaneous inductive and
deductive reasoning). As the analyst's task is by no means a clearly hierarchical or linear
process, this feature of the NExpert® system is a clear necessity.
(2) (U) Initial Goals
(U) The delivery of an expert system matching our specifications, if
indeed such a system can be devised, would obviously require numerous development cycles, and
*
(U) NExpert® is a product of Neuron Data, Inc., 444 High Street, Palo Alto, California.
7
Approved For ReleaseUPPCLASSIFIET36-00787R000300100001-9
- .
-
:
:

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
was therefore not an immediate expectation. After discussions with Vallee, it was decided that'
the thrust of the FY 1987 effort should be focused upon the mapping of response to target.
Rather than assuming that the target possibilities are effectively infinite, the system should be
designed and programmed to have prior knowledge of the structure and contents of the 200
targets in our current target pool. This design decision renders the problem much more
tractable, and certainly does not prevent further efforts from focusing upon response analysis
with little or no knowledge of the target universe. Once a system was in place, we would supply
Vallee with transcripts from 1987 experiments to serve as test data for the system.
(U) Even with a limited universe of targets, mapping a response to a
single target is not a practicable goal. Many targets are visually similar, and the information
contained in a typical response transcript is not sufficient to distinguish, for example, the Gobi
desert from the Sahara. Rather, a more reasonable task is to break up the target pool into similar
groups and map responses to a target type. As we had not yet determined these groupings,
Vallee undertook the description and classification of the 200 targets and used the resulting
target types as the basis for his work.
(3) (U) Description and Classification of Sites (DACOS)
(U) The system Vallee developed for categorizing the target pool,
DACOS (Description and Classification of Sites), contains 40 distinct target categories made up
of specific combinations of 27 possible target attributes. The set of attributes used is broken up
into six types: Water, Structures, Interfaces, Contours, Land, and Features (Figure 1).
-
Water
Structures
Interfaces
Contours
Land
Features
Present
Absent
Absent
Non-Urban
Urban
Waterfall
Island
Peninsula
Coastline
Lake
River
Hilly
Flat
Mountainous
Rugged
Fertile
Arid
Dense City
Open City
Rocky
Volcanic
Snowy (Glacial)
Green (Verdant)
Plowed
Monuments
Ruins
Dunes
1
1
Applicable
if water
is present
Applicable
if water
is absent
UNCLASSIFIED
FIGURE 1 (U) THE ATTRIBUTE SET FOR THE DACOS SYSTEM
8
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

-
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U) DACOS is a hierarchical system which is best described by a tree
whose interior nodes each represent a decision point and whose levels each represent an attribute
type. By selecting one of the possible paths at each node, one traverses the tree until a leaf node
is reached, at which point a target category has been selected. As is clear by inspecting the
hierarchy, it is possible to determine a target category by answering a maximum of five questions
(Figures 2 and 3).
Water
Structures
Interfaces
Contours
Features
Volcanic
Island
Non-Volcanic
Waterfall
Structures
Absent
Glacial (Snowy)
Coastline
Non-Glacial (Snowy)
Lake
Flat
River
Volcanic
Hilly
Non-Volcanic
Island
Water
Present
Structures
Non-Urban
Coastline
Lake
Flat
. River
Hilly
1
Island
3
-
Peninsula
Structures
Urban
Coastline
Monuments
Flat
River
Non-Monuments
Hilly
UNCLASSIFIED
FIGURE 2 (U) DACOS HIERARCHY FOR WATER-PRESENT TARGETS
9
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
Water
Structures
Contours
Land
Features
Volcanic
Mountainous
Non-Volcanic
Structures
Absent
Rocky
Rugged
Dunes
Snowy
Plowed
Fertile
Flat
Verdant (Green)
Arid
Plowed
Water
Absent
Structures
Non-Urban
Verdant (Green)
Fertile
Monuments
Hilly
Arid .
Snowy
Monuments
Dense City
Flat
Non-Monuments
Ruins
Open City
Structures
Urban
Non-Ruins
Monuments
Dense City
Hilly
Non-Monuments
Open City
UNCLASSIFIED
FIGURE 3, (U) DACOS HIERARCHY FOR WATER-ABSENT TARGETS
(U) The computer system Dr. Vallee delivered was developed using the
NExpert® system, but was later transferred and coded in BASIC for efficiency reasons. The
DACOS program initially prompts the analyst for an answer to the question "Is there water
present?" and, depending upon the response, continues to traverse the appropriate DACOS
decision sub-tree asking further questions. The final output is a list of pairs, the category type
with its corresponding confidence factor. Thesexconfidence factors directly reflect the number of
attributes that correspond to any given category. For example, if the final attributes were Water,
Urban, River, and Hilly, the categories under the Water-Absent node would have a factor of
zero, the categories under the Water-Present, Non-Urban node and the Water-Present,
No-Structures node would have a factor of one, etc. With these confidence factors, one can
then determine a hierarchy of possible correspondences for the given transcript and propose a
best match category.
10
UNCLASSIFIED
Approved For Release 2000/08/10 .CIA-RDP96-00787R000300100001-9

·
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(4)
(U) Results of Testing DACOS with Actual Remote Viewing Data
(U) The data given to Vallee for testing consisted of 30 RV transcripts
from one 1987 experiment. He then used the DACOS program to assign confidence factors to
the different target categories. Overall, the DACOS performed only marginally better than
chance. Out of 30 transcripts, the water attribute was correctly identified 17 times, the structures
attribute 13 times, and the full target classification twice. Unfortunately, this series of remote
viewings, when judged, did not show a significant RV effect, and thus proved a poor test case.
Nonetheless, this exercise allowed us to evaluate the progress and direction of this work.
(5)
(U) Shortcomings of DACOS
(U) The categories within the DACOS system were constructed to
produce visually distinct or "orthogonal" target types. The first two attribute levels of the
DACOS hierarchy, Water and Structures, are by far the most clear, simple, and symmetric; most
important, they correspond to common elements of RV transcripts. Nonetheless, some of the
target classifications are inappropriate for the kind of RV response data typically seen.
Experience has shown that the visual content is the most important aspect of a target; the minute
details of a target are often missed and thus should not overpower the overall description of the
target. For example, although several of the water targets do picture water, the water is confined
to such a visually insignificant region as to be either unnoticed or ambiguous.
(U) The deeper levels of the DACOS hierarchy do not maintain the
symmetric nature of the first two levels because they inherit properties from the preceding levels.
-
Furthermore, the attributes chosen for the deeper levels do not necessarily reflect the actual
visual nature of the targets. The attribute Monument discriminates targets on a very high social
and cognitive level; identifying a monument requires a significant amount of conceptual or
functional knowledge about a particular site, and is not necessarily evident from the visual
contents of a target. The attribute Non-Monuments, representing the explicit absence of
monuments, is even more abstract and visually ambiguous. Although Vallee's hierarchy yields a
target category with a maximum of five questions, the choice of attributes is not ideal for the task.
(U) Clearly, the most significant problem with the DACOS system is its
strict hierarchical structure. The system, by its tree traversal method, makes each decision
strictly binding; once water has been determined to be absent, the system does not have any
means of reconsidering water as an attribute. Effectively, by completely isolating sub-sections of
the decision tree, the DACOS system renders the attribute Hilly for Water-Absent targets to be
distinct from the attribute Hilly for Water-Present targets because the categories on one side of
the tree will receive credit and those on the other side will receive none. This occurs because the
11
Approved For Release 200UNCLASSIFIED787R000300100001-9

1
.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
system does not allow for any network or global attribute semantics. Ideally, the system would
not force a decision about the presence or absence of an attribute if that information is not
available; rather, it would consider the data already acquired, and prompt for alternative data.
Further efforts in the development of an expert system must allow for a broad and dynamic
evaluation of all of the data the analyst presents.
(U) The DACOS system implicitly assumes that all the information it
receives is certain. Clearly, a tool for the novice or uninitiated analyst must not expect that the
analyst will be correct in 100% of his decisions. Furthermore, the system must not expect that an
analyst will be able to render a meaningful decision about every possible attribute; in the absence
of data about any given attribute, the analyst cannot necessarily assume that the attribute in
question is in fact not present. For example, the particular attributes chosen for the lower levels,
optimized for the minimality constraint Vallee imposed upon the system, do not best represent
the elements typically contained in an RV response. The attributes Fertile, Arid, and Plowed are
rarely seen; assuming that data pertaining to these three attributes cannot be discerned, eight
categories are reduced to two, and the system has no way of resolving the analyst's uncertainty.
As all RV response data are, by their very nature uncertain, the expert system we envision must
deal with uncertainty from the very start.
(6)
(U) Future Directions
(U) The NExpert® development system offers many capabilities tailored
to dealing with uncertain reasoning. Unfortunately, the power of NExpert® was by no means
fully tapped by Vallee's initial effort. For this reason, we cannot make a meaningful assessment
of the potential utility of an expert system approach to RV analysis. Further development should
continue in this area, but the development will clearly need to focus upon the rectification of two
specific shortcomings of this year's effort: (1) the integration of uncertain reasoning into the data
analysis, and (2) the redefinition of an attribute set and network that allows for a more flexible
and comprehensive evaluation of the response data.
12
Approved For Release 2000/08/10 .
UNCLASSIEJER
0787R000300100001-9

1
.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
implemented the first working version of PsiLine. Bibliographic information and abstracts of the
major parapsychological journals from 1970 to the present were entered. During the second
year, FY 1987, the main purpose was to add as much material as possible. PsiLine now contains
bibliographic information and abstracts of the major parapsychological journals from 1940 to
date, complete sets of some of the minor parapsychological journals and several
parapsychological newsletters, over two thirds of the major English-language books on
parapsychology from 1880 to date, articles on parapsychology originally published in a language
other than English, and 1,000 relevant articles published in nonparapsychological journals.
6. (U) Objective C, Task 1 -- Personality and Health Assessments
(U) There was no activity on this task during FY 1987 because no new subjects were
added to our in-house subject pool.
.
7. (U) Objective C, Task 2 -- Analyze Personality Data
(U) A purchase order was let for this work to Dr. David R. Saunders of MARS
Measurement Associates during the first quarter of FY 1987. Specifically, Dr. Saunders was
asked to continue adding new cases to the PAS/psychoenergetic data base both from SRI
International and from subcontractor sources, to continue his study of the relationship between
the Personality Assessment System (PAS) and the Myers-Briggs Type Indicator (MBTI), and to
add known good hypnotic subjects to the data base as a potential basis for selecting hypnotic
subjects for psychoenergetic research and hypnosis.
(U) No new cases were added from SRI during the year but Dr. Saunders added PAS
data on four subjects from Psychophysical Research Laboratories, two subjects from Princeton
Engineering Anomalies Research Laboratory, and nine subjects from John F. Kennedy
University to the data base.
(U) In his work comparing the PAS with the MBTI, Dr. Saunders concluded that
MBTI scores could be predicted from the PAS but that predicting PAS scores from MBTI data
was not feasible at present. He suggested that predicting potential psychoenergetic function
directly from the MBTI and then using the PAS to confirm was a better procedure at this time.
(U) On the basis of accumulating PAS data on known good hypnotic subjects, Dr.
Saunders identified several potential subjects in our current data pool who would be both good
hypnotic subjects and good RV subjects.
14
Approved For Release 2009/08-0-61550595-60787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
8. (U) Objective C, Task 3 -- PAS Review
(U) During FY 1987 we conducted a thorough review of the (PAS) to gauge its
continued usefulness as a screening and selection instrument and a personality descriptor for
subjects in the psychoenergetics project. Data for this review came from published articles where
the PAS was reviewed or used as a research tool, attendance at the annual PAS conference, and
extensive interviews with several of the principal developers.
(U) The PAS is a multifactored personality assessment instrument that has been
evolving over the past 30 years using behavioral measures as raw data for making inferences and
predictions about personality and behavior. The early development work was conducted by John
Gittinger and his associates in a private firm that served clients in business and government.
During the last 20 years, the test has begun to make small inroads into the academic environment
but it remains obscure and controversial.
(U) It was concluded that although the PAS appears valid and is receiving growing
attention in academic circles, the instrument is currently not useful as a screening and selection
device either by itself or in conjunction with self-report measures. It is much too labor intensive
to be used alone and it has not been found possible to predict PAS profiles from MBTI data. In
addition, any type of screening use would require the testing of more high-quality remote viewers
than are available to the project at present. Use of the PAS as a descriptive tool has continuing
merit and it is recommended that we continue to test persons who show psychoenergetic abilities
on laboratory psychic tasks.
9. (U) Objective D, Task 1 (see Objective G, Task 1)
1
10. (U) Objective D, Task 2 (see Objective H, Task 3a)
11. (U) Objective E, Task 1 -- RA Effects on Marine Algae
$ (U) In FY 1986, SRI International awarded a subcontract to the College of Marine
Studies of the University of Delaware to conduct remote action experiments using marine algae
as target elements. Protocols were developed during that year that would enable SRI to test, with
a living system, the Intuitive Data Sorting model. During FY 1987, significant improvement was
made to stabilize the data so that standard analysis techniques (e.g., ANOVA) might be used.
While much progress was made toward that end, significant auto-correlations persist.
Regardless, an attempt was made to generate successful RA. SRI analyzed the data of four
participants and found no evidence of RA.
15
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
12. (U) Objective E, Task 2 -- RA Effects on Human Blood
(U) An experiment was conducted by the Mind Science Foundation to study the
possible relationship between intent to remotely influence a biological system and actual changes
in the system. Three phases of the investigation were conducted, including a pilot study, an
intermediate study, and a confirmation study. The first two were used to test and refine the
protocol for the third and final study. As a result of these preliminary studies and further input
from various experts, the confirmation study appears to have been extremely well conducted.
(U) Thirty-two subjects participated in the confirmation study. Their task was to
attempt to retard the rate of hemolysis (destruction) of red blood cells that had been placed into
a tube of distilled water and saline in a distant room. Each subject participated for one hour,
broken into four fifteen-minute periods. Of these four periods, two were identified as control
periods and two as protect periods. The experimenter measuring the rate of hemolysis was blind
to this condition. During the protect periods, subjects used visualization and other intention
strategies to try to protect the blood cells. During the control periods, subjects were to try to
think of other matters. In one control and one protect period, eight tubes of blood were
processed, and in the other periods two tubes were processed. Subjects were blind to this
condition. It was used to attempt to ascertain whether observed effects could be attributed to
causal relationships, or to intuitive data sorting. To see whether or not blood source was
important, fourteen of the subjects were trying to protect their own blood, and eighteen were
trying to protect that of another. Both subject and experimenter were blind as to the source of
blood.
(U) Results showed that 9 of the 32 subjects were able to achieve a significant
difference in the rate of hemolysis for the control periods versus the protect periods. The
probability of such an extreme result by chance alone is 1.9 x 10-5. There was no significant
difference between those trying to protect their own blood and those trying to protect that of
another.
, (U) The study was designed to try to determine whether causal forces or intuitive
data sorting were responsible for any observed psi results. The extreme heterogeneity in the data
made it impossible to make that determination. It is recommended that future studies of this
type be designed in such a way that data from each subject can be analyzed separately. It
appears that level of psychic functioning, whatever the underlying mechanism, is highly
individualized, so that it is difficult to test a specific theory using data combined across subjects.
16
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

.
.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U) Random access to any single target was programme'd via the computer, and a
tutorial program was constructed that walked the practicing viewer through an RV session. Using
these tools, an individual can randomly select a target from the pool, do a remote viewing
session, and receive immediate feedback.
(U) The first step in exploring the use of video disk RV training technology was to
establish a suitable video target pool. It was discovered that because of some deterioration in
picture quality inherent in the recording and playback process, a different set of visual criteria
had to be applied to obtain targets that, when copied to the video disk, retained acceptable levels
of feedback information. Additional considerations included, for example, finding appropriate
target materials to fit within frame parameter constraints, achieving acceptable color, granularity,
and focus, etc. After some experimentation, 243 National Geographic Magazine targets were
photographed frame by frame onto the video disk to serve as a pilot target pool.
(U) One of the best novice viewers from the FY 1986 training group was used to
demonstrate the capability of the system. Viewer 137 produced two sets of eight RV responses to
each of 16 targets selected randomly by computer. The responses were judged by comparing
each response to the eight targets in the set and ranking the response according to the visual
correspondence between the response and the eight targets. Analysis of these rankings showed
that significant RV occurred in one of the two sets. We concluded that significant RV functioning
could be obtained using the video disk format and propose that in coming years this device be
employed in conjunction with any proposed training program. In addition to using the video disk
technology as a training device, we also formulated a way of applying it to a screening and
selection task (see 'Objective F, Task 7).
15. (U) Objective F, Task 3 -- Develop and Test RV Training Hypotheses
(U) During FY 1987, an informal group of advanced remote viewer trainees and
researchers was organized to: (1) discuss variables that may affect the quality limits of RV, (2)
conduct practice sessions to maintain the in-house viewers in a state of readiness for formal RV
experiments, (3) provide a setting for reinforcing the positive psychological set necessary for
consistency of viewers' effort, and (4) develop experimental protocols designed to test
hypotheses generated during discussions. This effort was exploratory in nature and it was agreed
that any formal experiment proposals generated would be reviewed by the appropriate primary
investigator and the SOC before any formal experimental trials were conducted.
(U) The group met weekly during the first half of the year. On the basis of
discussions and informally conducted RV sessions, three experiments were proposed. The first
18
Approved For Release 2000008 CLASSIFIED.
0787R000300100001-9

.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
was designed to test the hypothesis that several RV sessions could be conducted over a period of
time without feedback after each session. The purpose of this proposed experiment was to
examine what effect lack of feedback would have on subsequent viewing attempts, given the
possibility that post-session feedback may not always be available.
(U) A second proposed experiment dealt with the issue of different types of target
material. Some viewers have reported subjectively different impressions when the task is to view
an actual outdoor scene (after which the viewer goes to the scene for feedback) than when the
task is to view a photograph of an outdoor scene (following which the viewer is shown the
photograph). The specific opinion is that the RV impressions are richer, more varied, and not as
limited to visual for actual scenes than for the photographs. It follows that viewer responses
might also be more detailed and not as limited to visual impressions. This hypothesis could be
tested by conducting a series of viewings where targets are randomly chosen from a pool of
outdoor sites and photographs.
(U) A third proposal reached the stage of a formal written protocol. It was to test the
idea that experienced viewers could perform just as well without a monitor as with a monitor
present in the RV session. This experiment was rated as a high priority because if it were
successful it would (1) reduce the resources necessary to conduct an RV experiment, (2)
eliminate potential monitor cues, and (3) allow multiple viewers to work on the same target at
once.
(U) Formal , work on experiments previously proposed and approved preempted
further efforts on these proposals. Advanced viewers spent the second half of the year
participating in several experiments with large numbers of remote viewings. Weekly meetings of
the group ceased for the remainder of the year.
16. (U) Objective F, Task 4 -- Develop RV Training Hypothesis
(U)'This task, originally intended to be fulfilled through a subcontract with
Consultants International (CI), was converted to a consulting relationship with the founder of CI,
Mr. Gary Langford. During FY 1987, Mr. Langford provided consulting services in two areas:
Remote Viewing
-
During the year Mr. Langford served as a viewer in approximately 100
RV sessions including practice, exploratory, and pilot work, and served
as subject in two major experiments:
(1) the Real-Time versus
19
Approved For Release /206. DASSIFIEDP96-00787R000300100001-9

-
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
Precognition experiment and (2) the Feedback experiment (see
Objective F, Tasks 1a and 1b).
. Hypothesis Generation
-
Mr. Langford participated in the meetings with other viewers and
researchers contributing to the formulation of testable RV hypotheses.
17. (U) Objective F, Task 5 -- Investigate RV Stimulus-Response Times
(U) Experienced monitors of remote viewing sessions have often come up with
hunches as to how to tell when a particular session might be more successful than another. One
such hunch has to do with the length of the response latency following the writing of the stimulus
word "target." The hypotheses tested in this pilot study were (1) shorter response latencies
produce relatively better RV responses, and (2) better responses are produced when less time is
spent producing them.
(U) Twenty-four RV sessions from a separate experiment were videotaped. An
independent analyst viewed the tapes and measured the response latency following each
presentation of the stimulus word "target" with a stopwatch. In addition, the total time elapsed
from when the viewer began his response to when he stopped to take a break was recorded. The
RV responses were analyzed by figure of merit analysis. The average response latency and the
average production time for each presentation of the stimulus word were calculated for each
session. Results showed a significant tendency for higher quality viewings both when response
1
latencies were relatively short and when production times were relatively brief.
(U) These results also add confirmatory evidence that RV impressions are relatively
brief and easily subject to modification by the associational processes of memory and experience.
18. (U) Objective F, Task 6 -- Investigate Hypnosis as an RV Debriefing Tool
(U) It has been assumed that remote viewing information is mediated through
subconscious processes and is therefore not readily available to conscious retrieval. Hypnosis has
been found to increase the ability of observers to recall information acquired in a variety of
circumstances where conscious recall has been blocked (i.e., material was presented
subliminally, trauma was associated with the initial perception, or information overload
occurred). In the present study, hypnosis was used in an attempt to enhance the data of an RV
session.
20
Approved For Release 2000/08
UNIGLASSJEJE
0787R000300100001-9

1
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U) The specific hypothesis examined in this study was that hypnotizing a subject
following an RV session and giving instructions to recall all the information associated with the
just completed RV experience would facilitate the recall of subconscious information blocked
from awareness during an RV session. It was hypothesized that hypnosis could provide a
significantly better aid in the recovery of unconscious, target-related material following a
standard RV session than only a second try at the same target. To test the hypnosis hypothesis
(hypnosis condition), a subject was hypnotized following a standard RV session (before feedback
was given) and given instructions to remember everything about the target from the just
completed session. A second RV session followed. The hypnosis condition was compared to a
control condition (proofread condition) where the subject was asked to proofread technical
report material following a standard RV session. A second RV session followed the proofreading
period.
(U) An SRI employee with previous remote viewing and hypnosis experience was
used as a viewer in the demonstration. Remote viewing experience included more than one
hundred monitored experimental sessions. In addition, the viewer had received certified formal
training in the practice of hypnosis. In preliminary testing, the viewer was found to rank in the
92nd percentile equivalent on the Standford Hypnotic Susceptibility Scales suggesting high
hypnotizability.
(U) Targets were individually selected just before an experimental RV session and,
while aware of the general nature of the pool, the viewer and experimenter remained blind to the
target until after each trial was completed. Twelve targets were randomly selected for 12
1
experimental trials from a group of 200 National Geographic photographs of natural scenes
previously chosen as a pool of potential targets for RV experiments.
(U) RV sessions were conducted in the standard way with a monitor present. After
the conclusion of the RV session, a computer randomly assigned the session to one of two
experimental conditions. In the hypnosis condition, the viewer was assisted into trance by an
experienced hypnotist. When appropriate trance depth was achieved, the viewer was guided
through a re-experience of the just completed RV session and given post hypnotic suggestions to
recall all the information acquired during the session. The trance was terminated after 30
minutes and a second RV session using the same target was conducted. In the proofread
condition the viewer was given a technical report to proofread for 30 minutes before a second RV
was conducted. Feedback followed the second RV session for each condition. This protocol is
shown in Figure 4.
21
Approved For Release 2000CLASSIFIED0787R000300100001-9

.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
RV 1
Hypnosis
RV 2
Target
randomly
selected
15 minutes
(30 minutes)
15 minutes
Feedback
Proofread
UNCLASSIFIED
FIGURE 4 (U) HYPNOSIS-RV PROTOCOL
(U) Two analyses of the responses were performed. The first involved a visual
judging of the 24 responses with each response blindly ranked by independent judges against a
subset composed of six randomly generated targets from the pool and the actual target.1* Exact
p-values from the sum of ranks were then calculated. Under the pre-treatment condition
(RV1), the RV quality failed to reach significance. However, the post-treatment sessions (RV2)
were independently significant (p < 0.029). Further analysis showed that all the significance was
due to the 6 trials in the hypnosis condition ( p < 0.025; n = 6). There was not a significant
difference between the proofread and hypnosis conditions.
(U) A second analysis of the 24 responses was conducted by another judge to
compare the calculation based on a 133-item descriptor list with the results of the visual ranking
analysis. The FM for each pre-treatment session (FM1) was subtracted from the FM for the
corresponding post-treatment session (FM2) and the resulting difference, AM, was plotted
versus FM1. A regression line was computed for both conditions. An F test was performed
comparing the "full" model which allows two separate lines, to the "reduced" model in which the
lines are the same.2 The general linear test comparing the two models showed no significant
difference between the two conditions -- likely due to the small sample size.
1
(U) The results confirm previous findings that hypnosis can facilitate the acquisition
of information not available to sensory processes. Its efficacy may be due in part to the general
state of relaxation produced by the process or to the greater right hemisphere involvement
thought by some to be a part of the hypnotic experience. These questions should be addressed
by continued research in this important area.
19. (U) Objective F, Task 7 -- Develop Mass Screening Protocol
a. (U) Introduction
(U) Current efforts for establishing a core group of talented remote viewers
have focused primarily on two major approaches: (1) enhancing RV abilities through the use of
*
(U) References may be found at the end of this report.
22
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
specified training procedures, and (2) screening the population for RV abilities using
performance-based psychological instruments (e.g., PAS and the current research into
neuropsychological testing).
(U) In the first approach, the emphasis has been primarily on enhancing
whatever latent RV abilities might be extant in a given subject pool. Talented performers in this
context have been largely defined as those who continue over time to demonstrate stable
accuracy and reliability in remote viewing within the confines of a single, highly specific RV
technology. Therefore, selection of talented subjects is relative to the specific training procedure
being employed and may not be related to identifying those individuals who are best on an
absolute scale.
(U) In the second approach, psychological profiles for known talented viewers
are obtained using two methods of psychological screening technology -- i.e., the PAS and a
battery of neuropsychological tests. In principle, these star subject psychological profiles can
then be used as templates for future subject selection. The major limitation of the psychological
screening approach centers on the labor-intensive nature of test administration. Unless a
meaningful second-order correlation with self-report tests can be effected, the potential for using
these methods for screening large populations appears circumscribed.
(U) Therefore, a third approach for locating talented individuals is suggested,
through the deployment of a standardized and automated procedure that would screen directly
for RV abilities in a large population. The following discussion advances some preliminary ideas
as to how such a pilot mass screening technology might be optimally designed and deployed.
b. (U) Hardware Configuration
(U) Initial design considerations for mass screening hardware would include
the following:
(1) Portability, i.e., the screening unit should be easily deployable in a
variety of settings; it must also be durable enough to withstand
frequent relocation;
(2) Efficiency, i.e., a variety of RV target materials should be rapidly
accessible, in order to exercise the range of a given subject's
abilities as efficiently as possible,
23
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

.
Approved For Release 2000/08/10 : CIA-RDP96, 00787R000300100001-9
UNCLASSIFIED
(U) Analysis of the results showed evidence that the factor of productive ideation was
partially related to measures of psychoenergetic function. Subjects who showed significant
computer dowsing ability in the time condition of the Search/Dowsing study showed a tendency
to have higher scores for productive ideation while subjects scoring significantly in the space
condition showed a tendency to have lower productive ideation scores.
(U) Two groups of remote viewers were tested: (1) a group consisting of four
experienced viewers who had shown significant remote viewing ability in previous experiments,
and (2) a group of novice viewers from the FY 1986 training program. Two of the experienced
viewers (009 and 372) received the highest productive ideation scores of all the individuals
tested. The nine novice viewers were ranked in order of performance on the last six sessions of
the novice training. The best novice viewer had one of the lowest productive ideation scores of
all the persons tested. The other eight showed a pattern of increasing productive ideation scores
as average measures of RV function increased.
(U) The scores on two of the tests, Sketches and Possible Jobs, showed high
correlation with the total scores on all ten tests. Since these tests require about 15 minutes to
perform and can be done in a group setting, it may be possible to use them as part of a screening
effort.
22. (U) Objective F, Task 10 -- Investigate RV of Analytical Information
(U) In the pilot phase of the exploratory analytics program, we have continued to
research some of the fundamental mechanisms of RV. The goal of these analytic experiments is
to identify the internal mental processes and other variables that enhance and/or inhibit psychic
functioning in forced-choice RV. We used one viewer. As during FY 1986, the FY 1987
experiments have been long distance; Viewer 002 was in New York City, and the experimenters
were at SRI International in Menlo Park, California.
(U) During the first half of FY 1987, we conducted a series of approximately 300
trials of the forced-choice format where, before declaring his response, Viewer 002 stated how
he felt about his contact with the target. Specifically, for each trial, he declared one of three
conditions: (1) "yes," he had contact with the target, (2) "no," he did not have contact with the
target, or (3) "?," he was unsure whether or not he had contact with the target. For these trials,
the viewer and an experimenter communicated by telephone. Targets were objects, Zener cards,
or words or numerals written on 3" x 5" cards. The experimenter, who worked in an office with
a computer, chose two possible targets and described them to the viewer. Using a random
number generator (RNG), the experimenter selected one of the two possible targets for the trial,
28
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
off-the-shelf components whenever possible. In particular, we chose a Hamamatsu silicon
microstrip radiation detector as the key element. Although charge-sensitive preamplifiers are
usually selected when using a semiconductor detector, such preamps are awkward to use in large
numbers. Because we required 48 active strips, a compact LeCroy current-sensitive preamplifier
and discriminator unit (2735B) was purchased. The 2735B cards were originally designed for
use with wire chamber detectors but we were assured by the manufacturers that the cards could
be mated with a semiconductor device. As we have learned, a substantial development effort
was required to reduce the noise of the system and create an interface between the detector and
the 2735B. The balance of this note describes the work which was necessary to eliminate
sufficient noise to observe the alpha particles.
a. (U) Initial System
(1) (U) Connector Noise
(U) The discriminator card is a current-sensitive preamplifier that
allows the noise floor to be adjusted using the threshold control. The system uses this threshold
control to calibrate the magnitude of the current pulses. One volt on the threshold- line will
discriminate against 2 LA of signal. If all 48 channels are high until the threshold voltage is
increased to 10 volts, then the noise floor will be 20 LA. As was determined later, the alpha
particles produce 40 LA pulses, not visible in the original system which exhibited 48 LA of noise.
The discriminator was on the outside of the vacuum chamber and was connected to the strip
detector by twisted pair ribbon cables and two vacuum feedthroughs. The contacts on the
feedthrough connectors contributed 10 LA of noise each. Because there was a connector on
each side of the feedthroughs, the noise contribution was 20 LA. This was determined by
unplugging the connectors on each side of the feedthroughs, one at a time, and watching the
threshold voltage go down 5 volts per connector (10 LA). This noise was reduced by relocating
the discriminator cards inside the vacuum chamber and eliminating the feedthroughs in this part
of the circuit. Now, only logic signals pass through the feedthroughs in the base plate, not the
low-level, current-sensitive lines.
(2)
(U) Hybrid Preamplifiers
(U) The wiring between the strip detector and the 2735B was originally
coaxial cables below the base plate and twisted pair ribbon cable inside the chamber. To reduce
the noise, the twisted pair was replaced by coax inside the chamber. When it became evident
that the wire was not the problem but rather that connectors were making the noise, we decided
33
Approved For Release 2dd @bA.S.SAPPED0787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
that the 2735B cards had to be placed inside the chamber. We resisted doing this initially for
several reasons:
The 2735B cards may be damaged in a high vacuum.
.
Overheating may occur.
· The system may be contaminated through outgassing.
(U) The heart of the 2735B is four custom hybrids called HILs. We were
concerned that these sealed hybrids would not work under high vacuum, but LeCroy guaranteed
us that they could even be used in deep space. The HILs require 1.5 amps at -5 volts and thus
produce a great deal of heat. Without convection cooling in the vacuum chamber, overheating
could make long runs impossible. To solve this, we used the strip-detector mounting plate as a
heat sink and mounted the 2735B cards on it backwards, with the HILs sandwiched in between
the PC board and the plate. A leaf of indium foil was inserted between the HIL and the plate to
provide better heat conduction. Using the mounting plate as a heat sink, we stabilized the
temperature of the HILs to 50℃. Our last reservation about mounting the HILs in a vacuum
was our concern for outgassing. This would lower our vacuum pressure and distort the path of
the alpha particles. However, no outgassing has yet been detected.
b. (U) Pulse Processing
(U) The 2735B performs to its specifications and the system noise is only 2 A
with no input. Our next task was to find out why the alpha particles were not visible at this time.
Detailed analysis of the detector preamplifier circuit revealed a subtle problem not addressed by
any vendors.
(U) We have shown the bulk silicon of the device to be N-type, as it is in the
Hamamatsu detector. When such a device is reverse-biased, a positive dc voltage is applied to
the highly doped N+ contact. An alpha particle that enters the depletion volume (the bulk of the
silicon) will generate a cloud of holes (positive charges) and electrons (negative charges) in its
path. Fundamental semiconductor physics dictates that the holes will be collected at the
negative-biased contact and the electrons will collect at the positive-biased contact.
(U) In our microstrip detector, the fabrication of the device determined that
we bias the entire unit through the N+ substrate and make individual connection to the 48 signal
inputs from the P+ contacts. Therefore, the 2735B input signal was in effect biased negatively
with respect to the substrate. However, as described above, the holes (i.e., positive charge) are
34
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

1
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
collected at that contact. As a consequence, a positive current pulse was being transmitted to the.
input of the 2735B.
(U) LeCroy's specifications state that the 2735B card is designed for negative
current pulses and will reject opposite polarity pulses up to 50 LA.
(U) Using a single-channel charge-sensitive preamplifier and a nuclear
spectroscopy shaping amplifier, we were able to carefully examine the alpha particle pulses
generated by the microstrip detector. We determined that the charge pulse should be equivalent
to a 40 pA positive current pulse. Given the rejection characteristics of the 2735B, it was clear
that our positive alpha particle pulse would not be detected.
(U) We elected to design a pulse-inverting circuit. Two custom PC boards
with 24 pulse-inverting transformers were made and installed on the output of the strip detector.
Because the current pulse's duration is 10 ns, we selected an RF pulse transformer so the signal
would not be attenuated.
(U) These transformers have an output impedance of 75 ohms, which
effectively short-circuited the input of the 2735B. It was necessary to add a 0.1-uF ac coupling
capacitor to the circuit. At this point, the alpha particles became visible as a normal distribution
on the computer screen for the first time.
c.
(U) Present System Performance
1
(U) After these modifications were made to the system, the noise floor was
brought down to 24 uA and discriminated away. Only signals larger than 24 LA will be seen by
the computer. The system can now see the alpha particles in real time, and is sensitive enough to
use as a tool to see any system noise. The noise has been greatly reduced; however, noiseless
performance outside of the alpha particle beam has not yet been attained. Two types of noise
have been found and need to be eliminated for an infinite signal-to-noise ratio. These types are
as follows:
1
.
Spurious, intermittent, random noise.
.
Parallel noise pulses.
(1)
(U) Random Noise
(U) There is a rare, random event that shows up intermittently in
random places. In the data presented at the end of this paper, the alpha particles are very visible
35
Approved For Release 200J/NJCLASSBFb5 -0787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
in the middle with a few single noise pulses randomly dispersed. This random noise is produced
by defects in the silicon strip detector. Detectors can be specially selected for their low noise
characteristics and a new detector could be purchased to eliminate these sporadic, individual
noise pulses.
(2)
(U) Parallel Noise
1. DC Line Noise
(U) Another anomaly in the system is a dc noise pulse that
intermittently pulses every channel simultaneously. This is referred to as a parallel noise pulse
and it can be caused only by an event that affects all 48 channels in parallel. The discriminator
power supplies, threshold voltage supply, and the strip detector high voltage supply are all
connected to the 48 lines and could be a common source of noise. A noise pulse on one of the
voltage lines could induce this type of noise pattern. Such noise is found when the telephone is
used. If the telephone is lifted off the hook, it induces several parallel noise pulse's in the system
and increases the count on each channel. A ringing telephone will not affect the experiment but,
as a precaution, the telephones were forwarded whenever data was collected.
2. AC Line Noise
(U) The power line was the next suspect point for ac line noise getting
into the system. All electronic equipment was plugged into a single power line filter/conditioner,
including the computer, CAMAC crate, and the discriminator power supply. When overloading
the conditioner reduced its effectiveness, the computer and CAMAC crate were removed from
the conditioner and plugged into the wall. The discriminator power supply and voltage threshold
supply were left on the power-line conditioner to reduce any noise to the charge-sensitive
electronics. To help filter low frequency noise on the high-voltage bias line of the strip detector,
an RC filter was used with a time constant of 100 ms.
4
d. (U) Electro-Magnetic Noise
(U) The sensitive inputs of the discriminator make it vulnerable to
electro-magnetic pickup. The presence of a large electro-magnetic field could account for the
induction of a parallel noise pulse being induced in the output wire that connects to the
discriminator input and acts as a receiving antenna. Shielded coaxial cables were used at first,
but their capacitance affected the charge-sensitive inputs of the discriminator. Shielding the bell
jar of the vacuum system, where all the charge sensitive electronics are held, helped shield the
36
Approved For Release 2000/RICLASSIFIED9787R000300100001-9

1
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
discriminator inputs from the environmental noise. Using the metal mesh cage around the glass
bell jar as a Faraday shield and covering it with aluminum foil helped to increase the noise
immunity to high-frequency noise injection inside the chamber. To reduce emitted EMI, the
computer and printer were removed from the CAMAC rack and placed 15 feet away from the
experiment. This did not change the system noise and the computer was put back in the rack.
e. ECL Logic Levels
(U) Another problem that could be causing the parallel noise pulse is a weak
ECL logic level on the discriminator outputs. The ECL output voltage is lower than specified but
is just within operating range. There is a 200-mV, 60-Hz ac sine wave riding on the logic
output. Combining the low logic level with the 60-Hz noise puts the logic level right on the edge
of the threshold between a logic 1 and 0. A very small signal on this line, such as an
environmental event, could be just enough to lower all logic levels. The weak logic level on the
discriminator output is not being caused by the coincidence register or the parallel OR gate
loading it down. The logic levels do not change when the registers are disconnected; loading the
2735B has no effect.
f.
(U) Ground Loops
(U) As the system noise diminished, it became apparent that ground loops
were being created by a grounding strap connecting all pieces of the system together. Eliminating
this strap reduced the noise floor another 4 pA.
g. (U) Alpha Particle Distribution
(U) The uneven distribution of the alpha particles is not inherent in the
system, but rather in the curium 244 source. Tests were performed to verify that the detector
strip numbers 1-48 are the same as computer channels 1-48. This indicates that the lines are
properly matched. When the source is moved a couple of millimeters, the alpha particle pattern
moves as well. This indicates that the strip detector is still functioning correctly. The two
collimating screens inside the curium source, or the curium itself, could have shifted slightly
inside the housing. This could explain the uneven distribution of alpha particles shown in the
data.
37
Approved For Release 2000 CLASS+FIE 787R000300100001-9

1
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U) The following data represent a sample of each type of noise. The data
were taken over a period of one-half hour with threshold voltage set at 15 volts and the curium
source set at a distance of 1 cm.
Spurious, intermittent, random noise
1-16
0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
17-32
0
0 0 0 0 0 0
30 36 833 3 4 78
0
14 0
33-48
0
0
0
0
0
0
0
0 0 0
0 0 1
0
1
0
Parallel noise pulses
1-16
0
1
1
1
1
1
1
1
1
1
0
1 1
1
1
1
17-32
1
1
1
0
1
2
0
40
22
848
0 8 71
1
13
1
33-48
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
The following is a chart of eight, one-half-hour sessions compiled with all the random
noise pulses displayed on the same histogram, and with the curium source present but subtracted
from the data:
1-16
0
0
1 3 0.
2
17-32
1
1
0 0 0 0 0 0 0 0 0 0 0 0 0
1
0 0 0 0
6
0
33-48
0 0 0 1
1 0 1 0 0 0 1 0 0 1
0 0 0 1
0
h. (U) Proposed System Testing And Modifications
(U) We are now able to demonstrate the presence of the alpha particle beam
well above,the noise floor. Three different software packages have been debugged and can be
used to give us maximum flexibility in the way we collect and display the data. Both types of
noise still exist but we now have a better understanding of the noise and the limitations of the
system. In order to meet the requirements of the RA experiments, the alpha particle system
must be noiseless except for the beam. In order to examine and, if possible, exclude all
remaining noise, the following tests and modifications are suggested:
38
Approved For Release 2ddNGLASSkbPED. 0787R000300100001-9
.
4

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
4
(U)
.
Separate vacuum system and CAMAC crate/electronics by 10-20 feet.
Monitor outlying random noise channel with charge-sensitive (pulse
shape) electronics.
· Retest, examine, and, if necessary, modify 2735B to increase the ECL
logic to -2 volts.
· Purchase
a new, microstrip
detector
selected
for
low-noise
characteristics. Price $3,000, with delivery in 3-6 months.
· Slowly and carefully move detector connections on present microstrip
device to examine the noise performance of all 48 strip combinations.
· Using EMI equipment and appropriate spectrum analyzers, retest the
ambient electro-magnetic noise environment.
(1)
(U) Detector Instrumentation
(U) During FY 1986, all necessary detector apparatus was specified and
purchased, or was fabricated. During the first half of FY 1987, this equipment was tested for use
in the proposed RA experiment and modified or improved where necessary. The radioactive
source has been verified using a separate detector system, and was found to be in the range of ~
100 counts per second, which will be adequate for the experiment. The operation of the
multiple-strip detector system has also been tested using a single-channel preamplifier unit and
found to be functioning properly when the system noise has been filtered by an appropriate
amplifier time constant. Principal difficulties encountered in the pilot work with the detector
centered on proper, alignment of 48 parallel channels and suppression of electronic noise pickup
from the environment. The first problem was solved by careful checking and rerouting all wires,
vacuum feedthrough connectors, and computer register inputs. Noise interference has been
suppressed by several techniques, including identifying and removing ground loops, establishing a
substantial ground plane next to the detector, supplying a shielded connector inside the vacuum
chamber, and replacing twisted pair wires with coaxial cable.
(2)
(U) Data Display
(U) With the help of an SRI Geoscience and Engineering Center
specialist in real-time computer systems, the LSI 11/23 computer, the Computer Automated
Measurement and Control (CAMAC) interface, and data inputs have been made operational.
We now are able to identify which of 48 possible detector strips have been activated, save those
data, and rapidly reinitialize the system for another cycle. The information is then transmitted
via the CAMAC interface to the LSI 11/23 computer, where it is stored in a memory buffer, then
39
Approved For Release 2009/0a GLASSDEJE00787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
shown on a video-display screen. In parallel, the pattern of activated strips for all 48 locations is
printed for later inspection. At present, this process is repeated each time any detector strip has
been enabled.
(3) (U) Artifactual Influences
(U) We have initiated environmental measurements of potential sources
of artifact, which may influence the flight of the particles or may add extra noise to the
electronics. The room in which the apparatus is located also contains a transmission electron
microscope (TEM), with its attendant high-voltage power supply and typical industrial
fluorescent lighting -- as well as the video-display terminal and other computer equipment.
Measurements of magnetic-field transients indicated a need for some shielding near the detector
apparatus.
This shield will served to suppress ambient electric fields. A TOPAZ power
conditioner was purchased and installed to suppress or eliminate power line surges resulting from
switching of other nearby apparatus such as the TEM mentioned above. The unit meets severe
IEEE and Mil-Std specifications for noise and transient suppression, and appears to have
successfully eliminated such problems.
26. (U) Objective H, Task 2 -- RA Effects on a Few-Photon Quantum System
(U) We have used a single-photon interferometer to examine the role of
consciousness in the state vector collapse. The result was that an "irreversible act of
amplification" does not require consciousness. The implication is that RA is, at least, not a
necessary condition in; nature.
27. (U) Objective H, Task 3 and 3a -- RA Effects on Strain Gauges
(U) In FY 1986, a joint venture was begun to examine possible remote action (RA)
effects on piezoelectric transducers. Participants were recruited, evaluated, and trained, by
,
researchers from John F. Kennedy University. SRI International developed an experimental RA
system, and prepared a well-characterized environment for formal experimental sessions.
(U) During the pilot phase, transducer signals were observed under sufficiently
controlled conditions to warrant continued investigation. During FY 1987, significant
improvements were made to the protocol, system hardware and software, and control
environments. A separate report reviews the FY 1986 pilot study and details the elaborate and
necessary precautions undertaken during FY 1987 to prevent or understand the sources of
artifact. No evidence for RA was observed in this experiment.
40
Approved For Release 200/NBCLASSIFIED0787R000300100001-9

.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
28. (U) Objective I, Task 1 -- Meta-Analysis of RNG Data Base
(U) This work is being conducted under subcontract with Psychophysical Research
Laboratories (PRL) in Princeton, New Jersey. The PRL Purchase Order Contract was let August
4, 1987 with the first deliverable due at the end of the first quarter of FY 1988.
29. (U) Objective I, Task 2 -- Test of IDS Model with "Dynamic" Systems
(U) A computer test of the Intuitive Data Sorting (IDS) model has been carried
forward from the FY 1986 tasking. The primary reason for this delay was that only one out of
the 100 individuals tested was able to demonstrate psychoenergetic ability during the FY 1986
screening phase. Because the IDS model is such an important model for the program at large,
we will continue to screen for talented participants.
.
(U) A modification to the computer program was made in order to provide
information about the details of the button-press timing. Because of the nature of
pseudorandom number generators, adjacent seeds do not produce nearly identical sequences.
Thus, the remarkable 1-ms timing reported by Radin and May3 appears to be a methodological
artifact. We incorporated a simple seed transformation in order to have the significant seeds be
evenly spaced in time. Thus, the IDS experiment is expected to yield results with regard to the
model, as well as with regard to human timing ability under psychoenergetic conditions.
30. (U) Objective I, Task 3 -- Host Theory Conferences
(U) No theory conference was held during FY 1987.
3
31. (U) Objective I, Task 4 -- Princeton Conference
(U) In FY 1987, SRI International awarded a subcontract to the Princeton University
Engineering Anomalies Research Laboratory (PEAR). The purpose of that subcontract was to
organize and host a conference of SRI Cognitive Science Program staff, subcontractors, and
designated consultants. The conference was held at Princeton's Scanticon Conference Center on
April 9-10, 1987, to discuss the topic "What constitutes proof of a controversial claim?"
Thirty-one persons attended the conference. There was also an after dinner speaker each
evening. A separate report contains an SRI assessment of the conference. Following the
conference, the PEAR staff prepared a conference proceedings.4
41
Approved For Release 2000/ CLASSIFIED0787R000300100001-9

.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
cost of the contract will be shared by those requesting coverage, and thus the cost to the project
is not yet known. Nonetheless, the cost will be at least 50% less than previously paid.
(U) Since the beginning of FY 1987, approximately $7,000 has been spent for
hardware repairs to monitors, video controller boards, and power supplies. In the past, it has
been more expensive to buy hardware maintenance coverage than to pay for individual repairs.
Currently, SRI is negotiating a hardware maintenance contract with Sun to cover all Sun systems
at SRI; this contract is expected to be as advantageous as the software contract. Once this
agreement is in place, key nodes in the system will be placed on a hardware maintenance
contract.
(U) Of the 13 Sun Microsystems workstations used by the group, all are in working
order, and only one is off line -- pending the completion of software and hardware modifications
to the tachistiscope experiment.
(U) When the Sun 3/280 file server was brought on line in mid-Match, the most
recent version of the UNIX operating system (Version 3.2) was installed. The most recent
version of the Unify data base program (Version 2.0), along with a new window-based interface
(SunSimplify), will be installed in mid-April. Unlike the old data base system, which was slow
and complex, the Unify system will provide rapid and easy access to data from any workstation
on the network.
36. (U) Objective J, Task 4 -- Upgrade Computer Hardware
(U) No new computer upgrades were made during the second half of FY 1987.
37. (U) Objective J, Task 5 -- Additional RA Experimental Hardware
(U) No new substantial RA hardware was purchased during the second half of FY
1987.
38. (U) Objective J, Task 6 -- Travel
(U) There are no deliverables required for this task.
43
Approved For Release 20UNCLASSIFIED0787R000300100001-9

·
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
39. (U) Objective J, Task 7 -- Additional SRI Internal Technical Support
(U) This task provides the allocation required to support ongoing research in a
variety of ways. Those areas are:
· Administration - Full-time Program Secretary
· Participant Charges
-
Consultant viewer fees
-
Time allocation for SRI participants in all experimentation
· Piezoelectric RA System Development
-
Labor charges for Senior Research Engineer (hardware development)
-
Labor charges for Research Engineer (software development)
· Alpha Particle RA Equipment Development
-
Labor charges for two Senior Electronics Technicians (hardware)
Labor charges for Real-Time Computer Specialist (software).
1
-
.
44
Approved For Release 2000/03/CLASSIFIED787R000300100001-9

.
.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
III PROBLEM AREAS (U)
(U) The FY 1987 Cognitive Sciences Program sadly lost the services of one of its Research
Analysts, as a result of the untimely passing of Martha J. Thomson on February 9, 1987.
Interviews were conducted to identify a replacement, and a new Research Assistant was hired on
June 1, 1987.
1
:
45
Approved For Release 200 UNCLASSIFIED787R000300100001-9

1
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
IV ADMINISTRATIVE COMMENTS (U)
(U) Peter J. McNelis and Edwin C. May separately set priorities for the Statement of
Work for FY 1987. McNelis set priorities on an Objective/Task basis, while May prioritized on a
level-of-effort basis for deliverables. The assignment definitions, which were formally agreed
upon, are as follows:
McNELIS
MAY
A
Must do
1
Major formal report
B
Should do
2
Pilot, exploratory, approximately
3 to 5 pages
C
Postpone
3
Wild guess, few paragraphs
0
Postpone.
(U) Table 3 summarizes the assignments on a task-by-task basis.
Table 3
(U)
PRIORITY/DELIVERABLE ASSIGNMENTS FOR FY 1987
RV
RA
Rating Task
Name
Rating Task
Name
B3
A4
Physical Correlation
A1
E1
Delaware
B3
D1
Dowsing
B2
E2
Mind Science
A1
Fla
RV/Precognition
A2
H1
Alpha-Particle
A2
F2
Video Disk
A1
H3
JFK
A2
F3
Advanced Training
A1
I1
IDS-Dynamic
B2
F4
Training Concepts
A2
J2
RA-Hardware
B3
F5
Stimulus/Response Correlation
A1
H3a
SRI Part of JFK
B2
F6
Hypnosis
A3
F7
Special Targets
B3
F9
Neuropsychology
B2
F10
Analytics
A1
G1
Computer Search
TECH SUPPORT
GENERAL
Rating Task
Name
Rating Task
Name
A1
A1
SOC Design
A3
13
Theory Conference
A1
A2
SOC Assessment
A1
14
Subs Conference
A1
A3
RV Analysis
A1
J1
Administration
B2
B1
Library
A3
J2
Publications
A2
C1
Med/Psych Baselines
A3
J6
Travel
A1
C2
MARS
A3
J7
Additional SRI Staff
C3
C3
PAS Review
A2
F11
MDS
A1
I1
PRL
A3
J3
Computer Maintenance
A3
J4
Computer Hardware
UNCLASSIFIED
46
Approved For Release 2000 MAICOL AIS-SIPPED 787R000300100001-9
4
A1
F1b
Feedback + SL Tachistoscope
A1
H2
QM Photon
A1
F8
Physiology Conference

.
1
,
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U) In a separate memorandum, several FY 1987 task changes and task interpretations,
were formally authorized. These changes are summarized in Tables 4 and 5.
Table 4
(U) AUTHORIZED TASK CHANGES TO FY 1987 STATEMENT OF WORK
TASK CHANGE
ORIGINAL SOW
CHANGED SOW
Old
Task
$ K
Activity
New
Task
Activity
Justification
D-2
35.0
E&M correlates to
dowsing
H-3a
SRI portion of
JFK
Provide recognition
of significant SRI
participation in the
JFK project
F-7
30.0
"Applications" targets
for novice viewers
F-7
Develop video
"mass" screening
Augment talented
viewer pool
F-8
16.2
MEG with
Los Alamos
F-8
Physiology
Conference
More effective search
for physiology
correlates
UNCLASSIFIED
Table 5
(U) AUTHORIZED INTERPRETATIONS OF FY 1987 STATEMENT OF WORK
€
TASK INTERPRETATION
Task
$ K
Activity
Interpretation
Justification
F-1a
87.5
Precognitive RV
To include real-time
RV
Balanced protocol
F-1b
S
87.5
Subliminal Perception
Continue FY 1986
feedback experiment
Experiment
contains SL
G-1
100.0
Abstract to real-world
target link
Continue FY 1986
computer search
activity
Necessary pilot
phase for link
investigation
UNCLASSIFIED
47
Approved For Release 2000108CLASSIFIED9787R000300100001-9

.
A
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
V PROJECT MILESTONE CHART (U)
(U) Table 6 is the overall project milestone chart for FY 1987.
Table 6
(U) ENHANCED HUMAN PERFORMANCE INVESTIGATION -- FY 1987
QUARTER
Phase
1
2
3
4
Objective A -- Protocols:
Task 1
Design
Task 2
Task 3
Task 4
Objective B -- Library
Task 1
Task 1
Objective C -- Psychophysical
Profiling
Task 2
Task 3
Objective D -- Field "Dowsing" Task 1
Objective E -- Continue RA
Experiments
Task 1
Task 2
Objective F -- RV Parameters
Task 1
Task 2
Task 3
Task 4
Task 5
Begin
End With Deliverable
Task 6
.
Deliverable
Task 7
0
End w/o Deliverable
Task 8
Task 9
Task 10
Task 11
UNCLASSIFIED
* Tasks with no beginning indicator (
) are continuations of an FY 1986 effort.
48
Approved For Release
UNCLASSIFIED
-00787R000300100001-9
Key*

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
Table 6 (Continued)
(U)
ENHANCED HUMAN PERFORMANCE INVESTIGATION -- FY 1987
QUARTER
Phase I
1
2
3
4
Objective G -- Computer
"Search"
Task 1
Objective H -- RA Parameters
Task 1
Task 2
Task 3a
Task 3
Objective I -- IDS Model /
Task 1
Conferences
Task 2
Task 3
Task 4
Objective J -- Administrative
Hardware
Task 1
Task 2
Begin
Task 3
End With Deliverable
Task 4
·
Deliverable
Task 5
0
End w/o Deliverable
Task 6
1
Task 7
UNCLASSIFIED
*
Tasks with no beginning indicator (
) are continuations of an FY 1986 effort.
-
49
Approved For Release 2000/08/10
00787R000300100001-9
Key*

SGFOIA2
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
Next 2 Page(s) In Document Exempt
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9

-
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
REFERENCES (U)
1.
Humphrey, B. S., May, E. C., Utts, J. M., Frivold, T. J., Luke, W. W., and Trask, V.
V., Fuzzy Set Applications in Remote Viewing Analysis, Final Report, Objective A, Task 3,
SRI Project 1291, SRI International, Menlo Park, California (December 1986)
UNCLASSIFIED.
2.
Neter, J., Wasserman, W., and Kutner, M. H. (1985), pp. 94-96, Applied Linear
Statistical Models, 2nd Edition, Richard D. Irwin, Inc., UNCLASSIFIED.
3.
Radin, D. I., and May, E. C., "Testing the Intuitive Data Sorting Model with
Pseudorandom Number Generators: A Proposed Method," Proceedings of the 29th
Annual Convention of the Parapsychological Association, pp. 537-535, Sonoma State
University, Rohnert Park, California (August 1986) UNCLASSIFIED.
4.
Hubbard, G. S., The SRI International Cognitive Sciences Conference at Princeton
University, Final Report, Objective I, Task 4, Project 1291, SRI International, Menlo
Park, California (December 1987) UNCLASSIFIED.
-
3
-
53
Approved For Release 200g/N/CLASSIFIED787R000300100001-9

.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
A
APPENDIX A
1
A POSTERIORI ASSESSMENTS OF
THE SCIENTIFIC OVERSIGHT COMMITTEE*
(This Appendix is Unclassified)
.
-
*
The SOC members were requested to complete a "Reviewer's Comments" sheet (see example on next page)
for each task that they had elected to review. This Appendix provides a verbatim, unedited transcription of
the reviewers' (mostly hand-written) comments on a task-by-task basis. SRI responses have been
appended to the reviewers' comments where appropriate.
A-1
Approved For Release 2000NBCLASSIFIED 0787R000300100001-9

.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
REVIEWER'S COMMENTS
The attached report titled:
has been reviewed by the undersigned.
My assessment of the research design, statistical protocols employed, the analyses of the
data, and conclusions reached in this report is as follows:
N
4
Additional comments:
.
1
SIGNED
DATE
A-2
Approved For Release 20bbpbCLASSIFIED 0787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
SOC Reviewers' Comments, Objective A, Task 3
(Fuzzy Set Applications In Remote Viewing Analysis)
(verbatim transcription -- not edited)
NAME:
S. James Press
Comments:
1.
The research is very interesting and is clearly very fruitful. It is so promising it
should continue to be pursued with vigor.
2.
Page 12, line 10. "Lowest p-values" - should be deleted. It's not necessary for
the argument -- "figures of merit" is enough. Then you're not involved in the
issue of interpreting p-values.
3.
Page 13, line 1. You should include 0 < p ≤ 1. I would put the 5th paragraph
about "the assigned u's, ... " up with line 2.
4.
Page 12, bottom line. Explain how the consensus decision was made. What
happened when there was disagreement? Were they ever at an impasse? How
far apart were they to start? Did you record this information? Would 5 analysts
make a difference? Would 2 analysts make a difference? What was the variance
in assessments?
5.
Page 14. In the equations, Rr and Ts are not defined; fuzzy set intersection is not
defined; how would we set the weights, W? Are the equations correct?
6.
Page 14, last paragraph. It would be substantially clearer if you gave a numerical
example, with a threshold, to show how o - cuts are used, and how Accuracy and
Reliability are actually computed.
7.
Page 16, line 8. How stable is the value "37," the average number of non-zero
values? If you did the evaluation many more times what would the variance in
the 37 be?
8.
Page 17. Formula is not correct.
9.
Page 19, paragraph F. "Ground truth" is a term appropriate in "remote
sensing." Since RV is a kind of remote sensing, the term seems entirely
appropriate. This might be explained.
10.
Page 21. I would not know how to record my answer to the "Belief in ESP"
question, nor do I know how to interpret other's responses to that question. To
me, the biggest issue is binary belief: it it possible at all? I don't know what
degree-of-belief means here? What does "complete" mean?
11.
Page 22, line 6. How do you know the scores are normally distributed? Do you
need them to be normally distributed?
A-3
Approved For Release 2009/03 /01 : 49SIFIEE90787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
12.
Page 22, Figure 5. Aren't the 6 subjective evaluations by a single analyst
correlated? All biases of a given analyst are bound to influence each of the 6
assessments
13.
Page 23, line 4. What are the definitions of "experienced," "expert," and
"novice" analysts?
14.
Page 23, Figure 6. I am confused by this figure. Are the entries correlations? If
they are correlations, why are some numbers greater than 1.0 and less than -1.0?
I am not clear what the entries in Figure 6 are.
Line 7 asserts that 15 differences were computed. I'm not sure, but I guess these
are differences between figures of merit for each response-target combination
and degrees-of-correspondence obtained from the PMT series.
If so, the
numbers aren't really comparable.
15.
Page 24, line 3. The correlation coefficient computed is only meaningful for
significance computations if the data are normally distributed. But are they? It is
not likely.
16.
Page 28, References. Why not delete "unclassified" for references 1, 2, 4, 6 and
put "classified" for the remainder? It calls less attention to the classified nature
of this work.
Recommendation:
Yes
11-24-87
NAME:
Brian Skyrms
Comments:
1.
This is a careful and well-thought out use of fuzzy set theory. I agree that the
important point is the identification of "orthogonal" target sets.
2.
I doubt if use of fuzzy sets with fuzzy elements (page 26) would really lead to
greater accuracy.
3.
The suggestion made in session that RV subjects use descriptors to do their own
classification after drawing, is a good to isolate fewer experimenter effects.
11-13-87
Recommendation:
Yes
NAME:
Mike A. Wartell
Comments:
1.
This method* seems to be an extremely powerful means by which to analyze the
results of experiments. It is the most promising formalism I have seen for use on
ambiguous data of the type provided by these experiments. I am slightly
*
Fuzzy sets and cluster analysis and appropriate universe of descriptors and "ground truth"
assessment.
A-4
Approved For Release 200PPPCLASSIFIED 0787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
concerned (but only slightly) by the small size of the sample used for determining
"ground truth."
2.
The described methodology must be published in some legitimate outlet so that it
gains external validation and is made available for other uses. (But I can't
overemphasize the need for external validation of every aspect of this project.)
3.
Suggest future research directions are appropriate.
4.
Also, choice of orthogonal targets as decoys makes analysis more straightforward.
5.
This may be the most significant secondary achievement coming from the project.
Response To S. James Press' Comments:
NAME:
Edwin C. May
1.
Most all of Dr. Press' comments have been attended to in the published final
report for Objective A, Task 3.
-
A-5
Approved For Release 20ddNCLASSIFIED00787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
SOC Reviewers' Comments on Objective C, Tasks 2 and 3 .
(Review of the Personality Assessment System)
(verbatim transcription -- not edited)
NAME:
Brian Skyrms
Comments:
1.
This careful review concludes that the PAS will not be useful for the purposes
envisioned. The conclusion appears warranted.
Recommend:
Yes
11-18-87
NAME:
Mike Wartell
Comments:
1.
Good descriptive report on PAS methodology. Does not really detail usefulness
or lack-of same to the project. Does not really report correlation of results, and
is, therefore, functionally different from other reports.
11-11-87
Recommendation:
Yes
NAME:
Philip Zimbardo
Comments:
-
1.
This is an excellent review which could be published in a professional journal with
minor revision.
2.
Need to mention that the PAS has also been rejected by psychologists because it
is a type theory at a time when typologies are not in vogue.
3.
I accept the conclusions reached that there is limited utility of PAS for RV
screening.
It should also be noted that a reason for the obscure status of PAS is the fact that
early research was probably classified, since it was part of the CIA's program of
identifying target individuals for various mind/behavior modification attempts
4.
Five characteristics of individual's that might correlate with RV ability:
1. Openness to new experiments (paper and pencil test).
2. Remote associates test (paper and pencil test).
3.
Tolerance for ambiguity.
A-6
Approved For Release 2000 d Gb A S S kBH D00787R000300100001-9

,
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
4. Tolerance for frustration.
5. High self esteem - high ego strength.
6. p. 4 Bem is Daryl not Darrel
Recommendation:
Yes
11-13-87
1
A-7
Approved For Release 2009'RECL:ASSIMEL0787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
SOC Reviewers' Comments on Objectives D and G, Task 1
(Computer-Assisted Search)
(verbatim transcription --- not edited)
NAME:
S. James Press
1.
Results of this work are interesting and encouraging, but much more work needs
to be carried out. The underlying phenomena are not yet well understood, and
replicability is problematic.
2.
Page ii, paragraph 3. Last sentence is an overdrawn conclusion based upon very
small sample.
3.
Page ii, last paragraph. I don't find that this "effect" is sufficiently well
understood as to it being ready to be proposed to the military. For me, we would
need large samples, replicated many times with the same subjects - the time and
space differential effects are not yet understood. The conclusion is overdrawn.
4.
Page 3, line 2. The formula 1-(1 -. 0001)72 is based upon independence of trials.
But they weren't independent; there was learning.
.
5.
Page 9, last paragraph. This approach to a p-value for an experiment is hokey,
ad hoc, and not founded in a scientific basis-or is it? If so, I would like to see a
proof of why this procedure is appropriate. It's not obvious.
6.
Page 13, middle paragraph. Results in 1987 with completely chance results
emphasize more than ever that these dowsing results are nowhere near ready for
military application.
7.
Page iii, last paragraph. Placing the subject directly over the target (the wreck) is
a location that could have been guessed by the subject (just as the combination
lock on a suitcase is usually a birth date, a marriage date, etc., and so with a bit
of prior effort, the result can be guessed). Strangely, there was no significant
result.
Recommendation:
Yes
11-13-87
Response To S. James Press' Comments:
NAME:
Jessica Utts
1.
Response to comment 2: It is true that the sample size upon which this statement
is based was small for the FY 1987 experiment, but the surprising thing is that this
same trend has survived three years of experiments. Also, the sentence was
worded as a suggestion of a trend, not as a conclusion.
1
A-8
Approved For Release 200PRPCLASSIPTED0787R000300100001-9
1

-
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
2.
Response to comment 3: Agreed. The paragraph in question has been reworded
to reflect a more cautious interpretation of the results.
3.
Response to comment 4: Learning would not change the fact that the trials are
independent under the null hypothesis, and p-values are always computed under
the null hypothesis. The game is set up so that even complete knowledge about
how it works will not help one's chance for success without the use of psi.
Response to comment 5: This approach to a combined p-value was the first one
used historically for combining results. It is powerful when one or more subjects
can produce a large effect, whether or not others can. This seems to be the case
in psi experiments, so it seems to be a good measure of the overall significance of
a psi experiment. An explanation to this effect has been added to the report.
Response to comment 6: Agreed. See note (3) and response.
Response to comment 7: This comment reflects a misunderstanding of the
experiment; the report has been reworded to try to clarify that issue. The subject
knew that they were anchored over the wreck, but did not know which segment of
the (unmarked) map corresponded to that spot. The grid used for dowsing was
keyed to the map in a random fashion.
NAME:
Michael Wartell
1.
Research design and analysis are acceptable, where CAS is involved. Some
aspects of the design of the Atocha experiment are unclear. Analysis of both
experiments is sensible.
2.
Some parts of preliminary draft are confusing to this reader as marked on report.
Writing needs clarification.
Recommendation: ,
Yes
11-13-87
4
NAME:
Philip Zimbardo
Comments:
1.
The computer-generated Task 1 target procedure is sound.
2.
This application of psi has clear operational implications -- and should be
continued.
3.
The results are disappointing, if the prime findings are those on page 11, last
sentence. However, operationally, if you are looking for anyone who can
enhance search operations for your client, then the individual data for s # 837 is
promising. But to be operational, there would have to be more stability of the
"dowser" over time, i.e., test situations.
4.
Q: Why not use as Ss known "dowser" with some established track record in a
laboratory controlled setting?
A-9
Approved For Release 200M3LA SISIPPEL9-00787R000300100001-9
.

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
5.
Page 13, paragraph 1. Conclusion: "once again produced significant results."
This conclusion again is vastly overstated, and should be tempered by greater
scientific cautiousness.
Recommendation:
Yes
11-13-87
Response To Philip Zimbardo's Comments:
NAME:
Jessica Utts
1.
Response to comment 4: We have worked with one known dowser (Subject
198), with mixed results.
Some of these are reported with the FY 1984
experiments.
Response to comment 5: While it is true that the experiment produced a
significant result using one particular criterion, the overall results were not as
good as they were in past years. Thus, the statement in question has been
tempered in the report.
1
A-10
Approved For Release 200UNCLASSIFIED00787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
SOC Reviewers' Comments on Objective E, Task 1
(An RA Investigation With Marine Microorganisms)
(verbatim transcription -- not edited)
NAME:
Michael Wartell
Comments:
1.
I have concern about quality and quantity of data contained within this report.
As stated, no conclusions can be drawn. I do feel that the proposed experimental
design and data analysis were acceptable.
.
2.
Report should simply identify this experience as a "busted contract" and go on
from there or drop the subject.
11-11-87
Recommendation:
Yes
A-11
Approved For Release 200RfCLASSIPTED0787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
SOC Reviewers' Comments on Objective E, Task 1
(October, 1985-September, 1987)
(Possible Photon Production During A Remote Viewing Task:
A Replication Experiment)
(verbatim transcription -- not edited)
NAME:
Michael Wartell
Comments:
1.
Design and analysis of experiments are acceptable. Experiment appears to have
been carefully accomplished.
Recommendation:
Yes
11-11-87
NAME:
Nicholas Yaru
Comments:
1.
I agree with the conclusion that the Chinese experimenters observed the results of
transients. This is conclusive data to terminate this experiment. The shielding
and signal processing techniques you have evolved should be useful in future
work.
Recommendation:
Yes
12-21-87
.
1
A-12
Approved For Release 2009/13CLASSIFIED90787R000300100001-9

1
.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
SOC Reviewers' Comments on Objective E, Task 2
(Experimental Protocol For Hemolysis: Confirmation Experiment)
(verbatim transcription -- not edited)
NAME:
Herbert L. Ley
Comments:
1.
The report is excellent. There is one variable that I did not pick up in earlier
reviews (probably because with 32 subjects the drawing of blood must be
staggered) that could be looked at in greater depth. Blood samples were drawn
for 14 to 42 hours before the experimental period. This is a three-fold variation
in time. Although ACD is a good blood preservative, it does not totally eliminate
the effects of aging on red cell fragitity. Therefore, the time between drawing
blood and the testing could be influencing results. Critics of this experiment may
be expected to focus on this uncontrolled variable.
2.
I would recommend, if possible, that the variable of time between drawing blood
and testing be tested for significance in the recent test using ANOVA. If no
significance can be associated with this variable, fine. Any future testing should
include controlling the time variable. Obviously, if the samples are refrigerated
sufficient time must pass before the testing to permit thermal equilibrium of all
samples. Whether the delay is 24 + 1 hours or 4 + 0.2 hours or some other figure
makes no difference. The important thing is that the time variable be controlled
in any future tests of hemolysis.
Recommendation:
12-18-87
1
Yes
Response To Herbert L. Ley's Comments:
NAME:
Jessica Utts
1.
Response to comment 1: The differences in time should be at least partially
controlled for by interspersing the control and protest periods. A statement to
this effect has been added to the report. Also, the suggestion will be passed on to
the Mind Science Foundation for their use in designing future experiments.
NAME
Brian Skyrms
Comments:
1.
This appears to be a well-designed experiment, which produced results which are
equivocal for the theories under consideration.
A-13
Approved For Release 20UNCLASSIFIED0787R000300100001-9

.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
Recommendation:
Yes
11-18-87
NAME:
Michael Wartell
Comments:
1.
I am concerned that the protocol became unduly complicated when the "own"
blood versus "others" blood was added as a variable. The results remain
ambiguous, but the extra variable clouds the issue. I suggest simply allowing this
experiment to disappear with the other RA experiments.
Recommendation:
Yes
11-12-87
1
A-14
Approved For Release 2000 0& CLASSE DE b.787R000300100001-9

,
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
SOC Reviewer's Comments on Objective F, Tasks la and 1b
(Feedback and Precognition Dependent Remote Viewing Experiments)
(verbatim transcription -- not edited)
NAME:
Brian Skyrms
Comments:
1.
Additional analysis should be done. What is the analysis of the sum total of
evidence presented here. We have 3 experiments; two negative and the third
with 4 viewings of which 2 are negative and 2 significant. Are the aggregate
results significant or not?
Recommendation:
Yes
11-13-87
NAME:
Michael Wartell
Comments:
1.
Protocol (page 5) would have been stronger had all time parameters been
matched (assistant leaving site and arrival back at SRI).
2.
Cluster analysis approach is excellent - othogonal target selection important to
efficiency and final analytical approaches.
3.
Last statement about Viewer 372 (top of page 11) is inappropriate and should be
deleted.
4.
Page 11-explanation is marked paragraph (2) inappropriate. Excuses regarding
"rushing trials" inappropriate.
5.
This experiment provided inconclusive results. I do not believe that inferences
should be drawn from the results.
Recommendation:
Yes
11-13-87
NAME:
Philip Zimbardo
1
Comments:
1.
The inconclusive pattern of hits and misses across these 3 experiments is
disappointing.
2.
I find as an over generalization and not acceptable the statement on page 3, "yet,
the evidence strangely suggests the (sic) [that] precognition is a fact of nature."
A-15
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
-
.

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
.
UNCLASSIFIED
3.
The out-bound protocol is wasteful of time and resources. I recommend
substituting "cinema verite" video displays of dynamic "real world" scenes as the
target material.
4.
Page 8 -- line 6 from bottom (typo - three experienced viewers).
5.
Page 10 -- line 2 from bottom in other words.
6.
Page 11 -- last line confused/also page 12, line 7, from bottom.
7.
The procedure and design and analysis are exemplary, but I am not convinced
that they allow for RV versus pre-cognition evaluation as competing theories.
8.
Given the predictions, the results for Ss with significant RV, 177 and 009 would
have to support real-time RV and fail to support pre-cognition hypotheses.
9.
The authors must avoid the stylistic bias of too quickly dismissing null findings
with a rush toward explaining them away.
10.
Conclusion: more research is definitely needed using the T-Scope procedure.
Why not have a set of practice trials to do an initial screening of S's RV
functioning on that day. If + then test, if not + then suspend -- give a rain check.
Recommendation:
Yes
11-13-87
Response To Philip Zimbardo's Comments:
NAME:
Edwin C. May
1.
Response to comments: Conclusions were modified appropriately.
1
-
A-16
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED

1
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
4
UNCLASSIFIED
SOC Reviewers' Comments on Objective H, Task 2
(A Quantum Measurement Experiment With A Single Photon Interferometer)
(verbatim transcription -- not edited)
NAME:
Brian Skyrms
Comments:
1.
This experiment aims to test the role of consciousness in the collapse of the wave
packet. A resolution of this question would be of considerable philosophical
importance and would bear on what sort of physical theories are relevant to other
phenomenas investigated by this project.
However, I believe that the interpretation of the experiment is controversial and
suggest that the results and discussion be submitted to appropriate physics journal
for peer review.
2.
Subcontractor's report (page 23) says the results contradict quantum theory.
The discussion under III: summary at results and discussion is better "one of the
possible outcomes allowed by q.m."
Recommendation:
Yes
11-18-87
NAME:
Michael Wartell
Comments:
$
1.
Elegant experiment! Design and analysis acceptable.
Interesting result and
discussion. (Fix report typos.)
Recommendation:
Yes
11-11-87
-
A-17
Approved For Release 2000/08/10 : CJA-RDP96-00787R000300100001-9
UNCLASSIFIED

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
SOC Reviewers' Comments on Objective H, Tasks 3 and 3a
(A Remote Action Experiment With A Piezoelectric Transducer)
(verbatim transcription -- not edited)
NAME:
Brain Skyrms
Comments:
1.
The experiment was carefully done, with good experimental design. I visited and
inspected the experiment.
Sources of noise and artifacts were carefully
controlled (much better than in the pilot study!). The conclusion of our evidence
for RA should be accepted.
2.
This counts negatively on subcontractor JFKU's track record.
Recommendation:
Yes
N
11-13-87
NAME:
Michael Wartel
€
Comments:
1.
Design and analysis acceptable.
Recommendation:
Yes
11-11-87
NAME:
Philip Zimbardo
Comments:
1.
This research represents "state of the art" protocols for the control or suppression
of RA sources of artifact.
2.
The apparatus and procedural features are impressive.
3.
Subjects were rated as very high in psi ability (8/10 were "practitioner-level" psi
people).
4.
Q: What about using the PIF scores for screening of your RV population?
5.
Caveats on pages 16, 17, and ff, are important to note: Anti-scientific personal
biases of JFKU research team led to poor research decisions!
6.
Post hoc motivatorial explanations are not acceptable (as on page 23) for failure
to find effects predicted. Researchers cannot resort to explanations that are
based on too much and/or too low levels of motivation, inhibitory and excitatory
motivation.
A-18
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
7.
In general, this null effect of a well-designed study and protocol seriously
questions any continued research on RA with changes in physical/biological
systems as the DV.
Recommendation:
Yes
11-13-87
.
1
5
-
A-19
Approved For Release 29999 LA's STPRef 00787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
1
SOC Reviewers' Comments on Objective I, Task 2
(Intuitive Data Sorting)
(verbatim transcription -- not edited)
NAME:
S. James Press
Comments:
1.
There were no really significant results of this effort although the care taken to
obtain scientific results is impressive.
2.
I would want to know more about the degree of randomness in these RNG before
suggesting the research be continued.
3 ..
Page 2, paragraph 2, first sentence. I don't agree that we would need to "have
access to future events." If we knew the seed and we knew the number
generating mechanism, we could state precisely what the entire sequence of digits
would be, and it would be the same sequence on every run.
4.
Page 2, paragraph 3. All information about the future is probabilistic. We can
assign (subjective) probabilities for all future events. Sometimes the events are
repeatable and the probabilities are objective. For example, the event that a
"six" will appear on a fair die when I cast it, 10 days from now. Are we being
given the probabilities of certain events in the future? If so, whose probability is
it?
€
5.
Page 6. Is there any meaning to the numerical ordering of seed numbers? How
are the "random" numbers generated by two successive seed numbers related?
Recommendation:
Yes
11-13-87
Response To S. James Press' Comments:
NAME:
Jessica Utts
1.
Response to comment 2: Details about the pseudo-random number generator
used in our experiments are being sent to Dr. Press.
2.
Response to comment 3: The experiments discussed in this section of the report
are based on true random sources, so the sequences are not determined by a
seed, but rather are generated after the buttons are pressed. Thus, one would
need access to future events.
The experiments we conducted, using a
A-20
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED

.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
pseudo-random number generator, would require access to future events only to
account for human reaction time to press the button.
We have not claimed
otherwise in this report.
3
Response to comment 4: The statement in question is a philosophical one. If the
future is already determined, and if precognition exists, then one might expect
that perfectly reliable information about the future would sometimes be available.
However, it appears that precognition does not operate that way. Instead, from
the potential futures, those predicted in precognition trials appear to be actualized
at a higher rate than expected by chance. This indicates that certain future
events may be more likely to occur than others, and that this information is
available. An extra sentence has been added to the report for clarification.
Response to comment 5: They aren't related at all.
NAME:
Michael Wartell
Comments:
1.
I do not find any problem with the experimental approach, but I am concerned
about the lack of definitive result, one way or the other. I do not favor dropping
the experiment -- it seems cheap and useful in developing some necessary
theoretical framework.
2.
Unless some better experimental verification (or anti-verification) of IDS can be
developed, I would favor continuing this set of experiments.
Recommendation:
Yes
12-03-87
NAME:
Nicholas Yaru
Comments:
1.
The research design and statistical protocols are valid.
2.
I agree with the recommendation that the experiment be continued at such a date
as the availability of significant performers can be fulfilled.
Recommendation:
Yes
12-21-87
A-21
Approved For Release 2099CLASSIFIED0787R000300100001-9

.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
SOC Reviewers' Comments on Final Technical Report
(verbatim transcription -- not edited)
NAME:
S. James Press
Comments:
1.
I find the overall approach used scientifically "open." The research staff is open
to new ideas and criticism, and anxious to attend to any problems found in the
research. Designs are well thought out, and analyses are careful. My only
concern is that too many diverse experiments are being pursued. A few narrow
areas should be defined and experiments pursued in these areas.
2.
Page 5. Are all consultants listed: I'm not for example.
3.
Page 10. Under "features" in Figure 3, items are not exhaustive of the category.
Also, under "contours," "flat" and "not flat" would improve upon "flat" and
"hilly." Factors seem to be equally weighted for DACOS.
4.
Page 11 (4). There should be more data given. For example, the water attribute
was correctly identified 17 times, but how many cases were there with water
present?
Perhaps instead of asking whether water is present or not, the question should be,
was there 50% or more water in the picture? Also, were there clouds and sky in
50% or more in the picture?
5.
Page 13, paragraph 2. Distributions of these variables are skewed, so the
correlations computed have questionable merit. I recommend first logging the
variable and then doing the correlations.
6.
Page 13, paragraph 3. Eye movements -- we can study change in pupil size by a
remote camera focused on the subject's eye.
7.
Page 14, last paragraph. Research has shown that while some people make good
subjects for hypnosis (and there are specific personality types for this), they are
very often inaccurate in their recollections, assessments, and evaluations of
situations -- their reports under hypnosis are composites of relaxed experience,
reality, and fantasies.
-
8.
Page 20, section 17. Again, what is recalled under hypnosis is not reliable. This
is all hokey stuff. "Eye witness" accounts are usually wrong!
9.
Page 32, paragraph 3. This description of an experiment is complicated
unnecessarily by the mention of an earlier experiment which was successful. The
earlier one has nothing to do with this one.
10.
Page 40, last paragraph. If only one individual was successful, it's not clear from
the 1:100 rate that this is more than what would be expected.
A-22
Approved For Release 200br:02 ASSIPPED00787R000300100001-9

1
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
.
UNCLASSIFIED
Recommendation:
Yes
11-13-87
NAME:
Brian Skyrms
Comments:
1.
If the viewer classifies his or her own drawing using the descriptor list developed
using the fuzzy logic techniques, the rest of the statistical analysis can be
automated. This has several advantages: (1) it makes the experiment more
objective and (2) it eliminates enormous amounts of labor on the part of the
analyst. Page 6, this would make the whole expert system project pages 6-12
superfluous. Frankly, the expert system project does not seem worthwhile to me.
It's preliminary success is slight, and it will impact all the problems of
¥
logic" into the analysis of experimental design.
2.
Screenings. There is no sense trying to find indications of psi ability until a large
pool of subjects who reliably demonstrate psi ability is identified. - Effort should
be to test large numbers of subjects in simple psi tasks, rather than complicated
tasks.
3.
External contractors.
There seems to be bad luck with a lot of external
experimentation. It would be better to concentrate on in-house experimentation
which can be tightly controlled.
Recommendation:
Yes
11-13-87
NAME:
Michael Wartell
Comments:
-
1.
Hypnosis experiment shows promise. Some stylistic changes needed in report as
noted.
2.
Page 22, paragraph 2. Discussion not productive. Simply state that further study
is necessary but don't blame lack of significance on small sample size.
3.
Screening. For future screening, identify high possible payoff target populations
based on intuitively derived characteristics, e.g., successfully dealing with
ambiguity. Thus, I would suggest making educated guesses concerning target
populations without doing strictly random sampling. Additionally, I am unsure
that the psychological profile information provides enough information for aid
with the screening process. Automated procedures are not only a good idea, but
a necessary one.
4.
I am also highly supportive of the dynamic target set option. Add to the
parameters for developing a screening mechanism the idea that extremely low
labor intensiveness is critical. Methodology: I'm not sure standardization is as
important as speed and ease of application -- after all, this is screening, not
research. I can't get over the feeling that the screening protocol discussions
describe potential methodologies that border on the baroque.
A-23
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
5.
Neuropsychological assessment.
Interesting possible correlates, but at which
level of screening would they be useful, initial screening, follow-on, etc .?
.
6.
Investigate RV of Analytical Information. I am not convinced about "less than
chance expectation" explanation for non-contact with target. Once again, I'm
not sure that "one-too-many" variables wasn't added to the variable set (in
contact or not in contact) to allow meaningful interpretation of results.
7.
RA Effects on Single «-Particles Agree strongly with discontinuation of this
experiment.
Recommendation:
Yes
11-13-87
NAME:
Philip G. Zimbardo
Comments:
1.
This a valuable summary overview of the project's successes and misses for FY
1987.
2.
It is clear, concise and informative.
3.
Good distinction between what did not work, what worked and what is mixed.
4.
Worth pursuing further - NExpert system development with:
a. Specifications for ways to improve its utility
for RV analysis
b. Find sources of support for resource library
c. Video disk technology
d. . RV training models
e. Hypnosis as tool.
5.
Not worth pursuing:
a. Gross physical correlates to RV
b. Skeptical about the value of PAS and a continuing commitment to this task
c. Objective E, Task 1, RA marine algae
d. Objective E, Task 2, Hemolysis
e. All PK research
6.
Unclear as to how to interpret the mixed results of Objective F, Task 1 -
which is very important.
7.
Unclear how much reliance can be placed on the hunches, intuition and
introspection of RV subjects. Staff seems to be overly impressed/sensitized to
them. Problem is the low correlation between S's experience and the quality of
A-24
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
their RV. (See Objective F, Task 10 - for S's failure to relate +/-/0 days to his
accuracy.)
8.
Need to clarify mechanism(s) by which hypnosis debriefing enhances RV
performance.
a. Relaxation
b. Right hemisphere activation
c. Rid mind of irrelevant noise
d. Emotional "flushing" other?
9.
Need to work further on Objective F, Task 7, mass screening, but do not support
PAS use or neuropsych testing. Recommend we look for Ss higher: REMOTE
ASSOCIATION (RAT TEST) imaginative ability, openness to new experience,
field dependence, cognitive flexibility. More work suggest on forced choice
format, computer "search" and no more work on Objective H, RA effect.
Recommendation:
Yes
11-12-87
1
1
A-25
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
APPENDIX B
PHYSIOLOGY CONFERENCE LETTERS
(This Appendix is Unclassified)
1
B-1
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED

.
1
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
Los Alamos
Los Alamos National Laboratory
Los Alamos, New Mexico 87545
August 4, 1987
Dr. Peter McNelis
SRI International, G-206
333 Ravenswood Avenue
Menlo Park, CA 94025
Dear Pete:
Thank you for inviting me to the Physiology Seminar. I really enjoyed
myself and I found the project to be quite interesting. I'm impressed by the
general quality and thoroughness of the research.
My suggestions for studying physiological correlates and indicators are as
follows:
1. Conduct a remote classical conditioning paradigm (a la the Hungarian
study that was mentioned, I don't know the reference). Set up a remote
light or tone conditioned stimulus (CS) that is predictive of a subsequent
shock unconditioned stimulus to the subject. Monitor EEG from scalp
electrodes and record the evoked response to the remote CS. You may also
want to record the GSR as well. If you find an effect, you have provided
strong support for the "informational" hypothesis. Additionally, it would
be quite interesting to subsequently monitor the effect using MEG. In this
way, one could distinguish the physiological source of a remote
conditioned response, as compared to the traditionally evoked conditioned
response. Positive results from such a study would be of remarkable
scientific interest.
-
2. It would be fairly straight forward to record motor evoked responses in
the computer search paradigm. The digitization of brain activity could be
triggered (in a pre-trigger mode) by the subject's press of the mouse (ie.,
record EMG and use it as a trigger). Single trial data could subsequently be
sorted and averaged according to correct and incorrect responses. One
would look for differences in the "readiness potential" (the broad
negativity that precedes the motor response) that distinguish between the
accuracy of the responses. I would predict an enhancement of the
readiness potential in trials preceding a correct response.
Approved For Release 2000/08/10 : CIA2RDP96-00787R000300100001-9
UNCLASSIFIED

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
3. In a similar vein, one could record the premotor activity in the computer
based remote action paradigm. However, a difficulty I perceive here is that
I'm not so sure a correct versus incorrect response is as well defined as in
the computer search game. It is my impression that a positive outcome is
determined on a statistical basis. However, there are responses that fall
outside of the normal curve, and perhaps the brain activity preceding them
is distinct from activity preceding the selection of a normal sequence of
random bits.
4. Determining physiological correlates to RV phenomena will be more
difficult because of the difficulty in timing the process. While one could
attempt to record brain activity evoked by the cue "target", I'm less
optimistic about this approach. There are simply too many unknown
variables in the situation. As was stated in the meeting, it may be that a
less time locked, more global physiological approach than evoked
responses would be better. How would the cerebral blood flow (or PET)
vary between a remote viewing , an actual viewing of a comparable visual
stimuli, and a session where one simply imagines a comparable visual
scene?
I would also push the "proof of principle" strategy and continue to
forcefully demonstrate the existence of the phenomenon in the absence of
known physical cues. I think a demonstration of the effect in a
magnetically shielded room would be of value in that sense.
I hope this is of help. I would be more than willing to go into greater
experimental detail if any of these ideas are of interest for your program.
Once again, many thanks for the stimulating conference.
1
Sincerely,
Dab
Deborah Arthur
B-3
Approved For Release 200009:10CAS3PPER387R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
SPECTRA Research Institute
3700 Osuna Rd. NE, #503
Albuquerque, New Mexico 87109
(505) 344-1040
.
September 2, 1987
Dr. Peter J. McNelis
c/o Ms. Cathy Flowers
SRI International, G-206
333 Ravenswood Avenue
Menlo Park, California 94025
Dear Peter,
I was glad to have the opportunity to speak briefly with you yesterday.
As I mentioned, I am sorry for the delay in getting my comments to you. Things
have been rather hectic since our meeting and time for quiet reflection hard
to come by.
Starting premises
Making my premises explicit seems an appropriate way to begin, starting
with an acceptance, at least implicitly implied as a basis for, our symposium,
of the proposition that there are real behavioral phenomena that correlate
with observable events shielded from human senses, and the possibility that a
second class of events are produced by humans without mechanical interaction
(the notion of time, or the time correlation of these events, will be
introduced later). Given this generalized beginning, the premises I wish to
introduce are that neuronal activity in the brain determines behavior, that
with physiological methods it is possible to look at macroscopic brain state
phenomena, and that EEG/MEG techniques provide a window into objective
information on brain states. (Other physiological correlates such as GSR and
EMG may also yield useful quantitative information, but for now, let the
burden of the discussion fall on the EEG). Additionally, it is assumed that
pervasive parallel processing obscures (even at one locus) single pieces of
information unless special techniques such as time-averaging are used.
If EEG techniques are used for establishing the existence of
physiological correlates and ERP phenomena, close attention should be paid to
the experimental design regarding placement of the electrodes, recording, and
analysis techniques, presuming more than a crude indicator of an unusual
event is desired. I am not so pessimistic as some researchers regarding the
existence of some clues as to what the brain is doing. Accordingly, for
example, I would not record the usual differential signal between active test
electrodes, but rather take the differential between each test electrode and
the reference electrode on that hemisphere. Note that by that last statement I
am encouraging the researcher not to tie together the reference electrodes
from opposite hemispheres. By using the two suggestions above, the chances are
increased of observing any unique information that exists from within each
hemisphere, and any information that may be in phase between two active test
sites is not thrown away. At any rate, these and other questions may be useful
B-4
Approved For Release 2000/DJ/GLAS SbFbE -. 787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
to examine as regards experimental design, including choice and application of
data analysis techniques.
Training and screening
As regards experimental RV subjects, I feel a need to be convinced that
there is no training effect possible with known adepts. At least with respect
to quality of RV responses, I have been exposed to information claiming that
enhanced quality of RV responses due to specific physiological training
techniques was possible; don't know if the claim is true or not, but I find it
conceivable that an induced change in brain state such that a higher
percentage of the subject's time can be spent at his/her cortical tone rest
frequency (read alpha frequency) under alert, rest conditions would contribute
to improved performance. The brief question I raised at the symposium
regarding this issue was not really answered or discussed due to the press of
time.
On another tack, I believe I remember commentary at the symposium to the
effect that to date no personality screening techniques were able to separate
RV adepts from non-adepts. My follow-on question to this, given Ed's proposal
of time-independence for RV phenomena, is whether PAS or some other test(s)
might separate those adepts who are good at precognitive RV from those who are
not? If there is a separation, this raises other questions that may impact the
experimental design, how the ERP data is taken, and analysis.
ELF and the brain state panorama
Perhaps a physiological screening technique for adepts is possible based
on their normal EEG "alpha" rest frequency. The suggestion is that there may
be a useful correlation between rest frequency and RV ability (as there may be
between EEG rest frequency and EM sensitivity), based in part on some of my
other endeavors with which you are familiar. As I mentioned to you at one our
breaks, some attention still needs to be paid to the tenability of an ELF
hypothesis despite negative glib arguments to the contrary. The Schumann ELF
resonances yield a continuous, uniform field operating in a transverse
magnetic mode that does not attenuate as the inverse square of distance. It is
at least coincidence that the brain is upset sensitive to H-component
frequencies outside the first Schumann resonance frequency regime, and will
try to decode ELF frequencies or modulations that look like the brain's own
signaling system. Conventional shielded rooms should improve performance for
remote viewers inside them, as the rooms will diminish EM noise with which the
brain may have to contend, but the rooms do nothing to block ELF. RV performed
in a specially shielded MEG room may be a tougher test but an H- component
attenuation of 2000 at 10 Hz (such as the room at LANL) still may not block RV
even on an ELF hypothesis. Further, an ELF hypothesis for RV can be envisioned
that does not entail sending/receiving information with a human as its point
source and thereby the usual inverse square problem (not to mention power
generation and antenna length), but rather coding/decoding by the brain
on/from an already continuously existing, uniform field ELF carrier containing
information. In addition, known brain architecture reveals millions or tens of
millions of potential simultaneous information processing channels rather than
a single channel operating under Shannon bit information constraints
appropriate to a single ELF input channel. With regard to timing, since we're
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
dealing with Maxwell's equations, I have no a priori difficulty contemplating
advanced potential solutions as part of this information transmission
possibility, thereby leading into the precognitive realm or what may be called
advanced time (AT) solutions. In fact, the paper by Nelson, Jahn, and Dunne
given to us by Ed at the end of the symposium has in it a time event density
plot that looks suspiciously as though it might be gaussian if the number of
events on an ordinate scale were plotted along the same time axis. The number
of events appear to fall off sharply in either direction at about three days
but there would still be a small finite probability at longer times in either
direction. I suspect such a plot could afford some interesting discussion with
respect to Maxwell's theory.
Neuro team member (s)
Some commentary was made regarding having a neurologist aboard as a team
member. Most neurologists are trained intensively to look for signs of human
pathology and may not be expert in the neurology of the normal, so one would
want to be sure of their motivation and expertise as a team member oriented to
look at human potential, and perhaps new neurological phenomena. In any case,
it would be most desirable if a new person is added that besides the requisite
background he/she be thoroughly fascinated with brain states and coding
phenomena in the brain, for I think that it is from those issues that the
break will come.
Miscellaneous comments on RV and PK
For RV at t = +0, 0, or -0 a known event either existed, exists now or
will exist as arranged by some action. The issue is therefore access to an
information "channel". If there is a possibility in PK that there is no human
instituted action but rather an adept is proficient at picking out anomalies
in real, past, or advanced time then we are also left with the issue of
information "channel" access, as in RV. In both cases, the EEG experimental
design could encompass multiple instances of time-locked events to produce ERP
experiments, and time-averaging procedures to extract the signal from the
background activities. From just an on-going EEG record of subjects in an
experimental situation nothing of any critical significance should be expected
to be extracted since the signal being looked for is totally masked by myriad
other processing activities. Given a time-locked approach, analysis is
suggested for times prior to time zero as well as subsequent to time zero, as
will be remembered from my comments at the blackboard. To produce multiple
instances of time-locked events of PK, I suggest something akin to the
ping-pong ball normal distribution experiment. In this case, time-average
those events to the right and left as independent groups, and in the ERP
analyses look for significant differences etc.
As a point of curiosity, left-right reversal reported in RV may not be so
peculiar after all, since no one is postulating the use of visual channel, and
therefore there is no initial reversal as in the case of information from left
and right visual fields to opposite cortices.
B-6
Approved For Release 200g/PICL: AIS SPFRE DP787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
Potential Spectra Research Institute participation
SPECTRA is willing and interested in participating in
electrophysiological correlates research with SRI International, should that
be appropriate. Two areas are suggested for consideration:
(1) Computer-controlled ERP remote RV experiments between different rooms
with both subject and "outbound experimenter" instrumented. The experimental
design is envisioned to be essentially the same approach as I outlined at the
blackboard during our symposium.
(2) Remote hypnosis experiments between different rooms based on an
information access model much as in RV. Instrument both participants and use
protocol involving time-locked wake and sleep commands. Monitor wake and sleep
EEG brain state onset and time course in subject. (At last, here is some
macroscopic and quantitative EEG difference to which most electrophysiologists
can agree).
Finis
Hope these few cryptic thoughts provide input for useful reflection. The
comments and arguments are not polished nor fleshed out, as is evident. Even
if they were, I suspect they are ripe for engendering some controversy and
debate. Some or all of these avenues of thought might be useful to explore in
greater detail with at least you and Ed.
.
Thanks again for inviting my participation in a most stimulating
symposium. I thoroughly enjoyed all the conversation and points of view.
Best regards,
Richard H. Dickhaut
B-7
Approved For Release 2008/59/4. AS SUFLE 50787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
Los Alamos
Los Alamos National Laboratory
Los Alamos, New Mexico 87545
August 5, 1987
Dr. Peter J. McNelis
Cognitive Sciences Program
Geoscience and Engineering Center
SRI International
333 Ravenswood Ave.
Menlo Park, CA 94025
Dear Peter:
I am responding to your request for possible ways
we
can help
your program.
I
can see
three ways immediately that may be
useful to you.
As we discussed in Menlo Park, the magnetic shielded room could
be of use
for measuring the influence of very low frequency
magnetic fields on subjects. We can do this in a variety of ways
which could involve repeating of some of your standard RV
experiments inside the room, repeating the EEG experiment you did
inside the shielded room or looking at the influence of the room
on some evoked response studies such as I
will suggest below.
Obviously such experiments will need to be done at Los Alamos or
some other shielded room. (Actually, there is some sort of room
at Stanford used for magnetotelluric studies but it is primarily
a DC shielded room using iron instead of mu-metal) .
A second category of experiments involves the use of evoked
responses (ER) for measuring time locked responses. Again, many
of these were discussed at the Menlo Park meeting. However,
to
summarize the discussions as I see them, I would like to make the
following comments.
The RV experiments could be timed locked to
the prompt word
or some
other type of stimulus.
These
experiments should be first done with EEG to look for effects and
then could be done with MEG to localize physiological features if
the EEg results are positive. A more definitive experiment along
these lines would be the light or auditory experiment using
stimuli from other rooms and time locking the recording of ER's
to stimuli presentation. The original stimuli should be local so
that the ER's could be examined
for their patterns for each
subject and the subject would know what the stimuli were. Again
these could be done first with EEG although the MEG
setup
is
directly applicable. The EEG could be done at SRI, Los Alamos or
another contractor. There are many
variations
of
these
experiments which could be examined if they are of interest.
Again the effect of the shielded room would be interesting.
B-8
Approved For Release 2099'NCAIS SPPPE BP787R000300100001-9

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
The third category of experiments would be
the presentation
of
multiple stimuli
with the subject told to pick one out.
As we
discussed, the subject could press a button in anticipation
of
the correct stimuli or other procedures could be used. In any
case, the experiment would be conducted in a pretrigger mode with
data collected before and after each stimulus presentation to
examine the ER's for early components. This and the previous
experiments have the advantage of time locked signal averaging to
improve signal-to-noise.
They also have
the advantage
of
allowing many control checks on the data as well as a wealth of
existing data both from EEG and MEG. In addition, power spectrum
can also be obtained simply be doing FFT's on the ER data if
these are desired. Clearly these experiments would want to
utilize some of your special subjects. This latter group of
experiments would best be done at Los Alamos because we have all
of the equipment and paradigms basically available now. This
would also have the advantage that we would have the MEG
available for localizing the sources of any effects we saw and
the subjects would be used to the shielded room environment.
It also occurs to me that if any of the experiments involving the
alpha or piezo gadgets work out, it would be desirable there also
to monitor the brain for distinctive changes and
perhaps
establish physiological correlates.
I hope this is of some use to you. I really enjoyed the meeting
in Menlo Park and it would be nice to have future correspondence
with you. In any case I will be more than happy to assist you in
any way possible. Should I have more thoughts on the subject I
will communicate them with you. I hope to hear from you in the
near future with your response to this letter.
-
Sincerely,
3
Edward R. Flynn
Los Alamos National Laboratory
Box 1663, MS D434
Los Alamos, NM 87545
Tel: (505) 667-4746
B-9
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
1

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
MICHIGAN STATE UNIVERSITY
DEPARTMENT OF MEDICINE . ROOM B220 LIFE SCIENCES
DEPARTMENT OFFICE (517) 353-6625
EAST LANSING . MICHIGAN . 48824-1317
30 July 1987
Dr. Edwin C. May
Cognitive Sciences Program
Geoscience and Engineering Center SRI International
333 Ravenswood Avenue
Menlo Park, CA
94025
Dear Ed:
I'm writing to thank you and Peter and your colleagues for introducing me
to a most interesting field and including me among a fascinating group of
participants -- and for your warm hospitality. I'm writing also to summarize
my thoughts regarding the problems of the present state of development of
your program and a few recommendations for future research.
First, some general comments:
0
The "science" of the present efforts at SRI International
is a welcome finding in a field overloaded with interesting
anecdotes and poorly-designed experiments -- and apparently
some fraud. Your background as a well-trained broadly-experienced
and accomplished physical scientist lends great credibility to the
work. I liked your open, honest presentation style.
0
It's an enormously difficulty field to study well and your interest
in moving more deeply into the "basic science" aspects of the
phenomena and observations is welcome. My intuition tells me that
there won't be much found by physiological research studies the
first time around -- but the success record of my intuition isn't
very good in fields I know little about. Yet, even if that turns
out to be true, it should not disuade you from carrying out a
'series of carefully-designed and conducted studies using the
best methodologies and technologies available. You may get
interesting and perhaps useful results. You may get some leads
that will direct future research efforts that will be successful.
And/or you will bring into the field a corps of first-class scientists
and graduate students who, over time, may or will come up with answer
This last isn't a trivial outcome.
O
The scientific quality of the investigators who collaborate with
you in the basic science studies is very important, if not crucial.
Especially in a field like yours, you need the kind of face-value
creditability they will bring, so that their results will be unchal-
lengeable. Getting them to become involved will
probably be difficult
and expensive and inconvenient for you, but will be worth the effort
entailed.
B-10
Approved For Release 200PreCLASS1PHE90787R000300100001-9

-
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
" I would like to suggest for your consideration two groups of
extramural scientists, consultants.
- Your present group, if properly constituted could
serve as close advisors/collaborators with frequent
visits and even participation in design, data assessment,
etc.
-
Another subgroup of scientists might be impaneled
to serve as arm-length, objective evaluators, critics
and challengers. I've seen such a division of responsi-
bility and expectations work well elsewhere.
Now, for some more specific suggestions:
o
I liked Steve's general approach to the neurophysiological
evaluation of RV (and perhaps RA's).
0
I doubt very much that there will be any value in studying
biochemical or nonneurological physiological changes in humans,
but it might be useful to ask the question of a group of sophisticated
neurophysiologists working closely in the field of neurochemical -- be-
havioral interactions.
0
It would seem to be very important to develop a major effort
to create animal models -- small animals to start, but later a
larger animal, such as monkeys. If successful, this would
permit all sorts of useful short-term and long-term investigations.
0
Despite the apparent (to me) failures of the past, I feel that
a careful assessment of psychoactive drugs of various sorts
would be a useful initiative. If one or more performance enhancers
or blockers could be found, this could be extremely valuable.
0
Although you've excluded children and bizarre-types (for
want of a better term) from your study groups, I wonder if
it might be useful to seek out from these groups and others
all sorts of people who can document strong and consistent
performance.
· I realize there have been some studies of primitive organizisms
(algae) and even cells, but further studies at these levels of
complexity might be productive. For example, a colleague of
mine at Michigan State University in the Department of Pediatrics
(Dr. James Trosko) is very successful in studying cell -- cell
communication.
It would be useful to know if such systems could
be perturbed.
0
I agree with you that a high priority ought to be assigned to
the discovery of means to screen larger groups to uncover people
with high performance capabilities.
Finally, it might be time or organize and carry out a two-four day "idea
session", bringing together very creative scientists from a variety of
fields who can be stimulated to elicit new and promising approaches to
Approved For Release 2009/08/19 :
: CA RDR96 00787R000300100001-9
UNCLASSIFIED

.
ʻ
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
address your problems. These sessions are difficult to pull off. Such
groups must be brought together by one or two very well respected scientists
who can recruit other top flight scientists from many fields. They must
be introduced to the fields sufficiently, comprehensively, to give them a
pretty good understanding of the issues. And, the discussions must be
structured to prevent them from being unfocused and nonproductive. Kindling
creativity and focusing it into certain directions using brilliant, independent
scientists is an awesome challenge -- but it is doable sometimes.
I enjoyed my trip. Good luck!
Regards,)
Ray
Raymond H. Murray, M. D.
Professor and Chairman
Department of Medicine
RHM/1p
B-12
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED