ENHANCED HUMAN PERFORMANCE INVESTIGATION

• Document Type: CREST
• Collection: STARGATE
• Document Number (FOIA) /ESDN (CREST): CIA-RDP96-00787R000300100001-9
• Release Decision: RIPPUB
• Original Classification: U
• Document Page Count: 78
• Document Creation Date: November 4, 2016
• Document Release Date: June 11, 1998
• Sequence Number: 1
• Content Type: REPORT

ved For Release 2000/08/10: CIA1-RDP96-00 7l 000300100001-9
Final Technical Report
Covering the Period 1 October 1986 to 30 September 1987
ENHANCED HUMAN PERFORMANCE
INVESTIGATION (U)
SRI Project 1291
PY ??-?-???? of ... Copies.
is document consists of S3 pages.
oved For ft1 os,Z80OQ&L10? :IUQ IA-RDPWLfOOrP8 603UbfQ0001-9
14151 ^26-;;200 . ('able: SRI INTL MPK ? TWX: 910-373-204
 Approved For Release 200 6flet Itgli f tEB787ROO0300100001-9
TABLE OF CONTENTS (U)
LIST OF TABLES .........................................................
LIST OF FIGURES ........................................................
iv
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
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
I
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
 Approved For Release 200p abkS?r?tb787R000300100001-9
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
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
A
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
 Approved For Release 200fflet jItkpf ffp787R000300100001-9
(U)
University of Delaware, plus the consultants having expertise in specific areas of interest to the
program.
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
 Approved For Release 200M1e 1C -5fflFjt jJ787R000300100001-9
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.
(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:
SGFOIA2
? Consultant Agreement - Used for a single individual within a large
organization.
? Services Contract - Used for contracts having total funding of~less than
$100,000.
? Full Subcontract - Used for contracts let by prime contractor having total
funding of greater than $100,000.
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
 Approved For Release 200fMrLAtSJftElU787R000300100001-9
SGFOIA2 vv''`v~
C. (U) Progress to Date for Each Objective/Task
(U) The progress to date for each Objective and Task in the Statement of Work is
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.
UNCLASSIFIED
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
 Approved For Release 200p jtL iIt VffffY787R000300100001-9
ff_,%ftff?
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.
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 numbeiv 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-NExperto*
development system, an expert system shell which exploits non-monotonic reasoning (i.e.,
simultaneous forward and backward chaining oit, 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 Release(Wt//A6-007878000300100001-9
 Approved For Release 20U00/08/10 :CL
(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 use the resulting
target types as the basis for his work.
(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).
I
Water Structures Interfaces Contours Land Features
Present Absent Waterfall Hilly Fertile Rocky
Absent Non-Urban Island Flat Arid Volcanic
Urban Peninsula Mountainous Dense City Snowy (Glacial)
Coastline Rugged Open City Green (Verdant)
Lake Plowed
RIver Monuments
Ruins
T Dunes
Applicable Applicable
if water if water
is present is absent
FIGURE 1 (U) THE ATTRIBUTE SET FOR THE DACOS SYSTEM
~IN #-%~S IFIED
Approved For Release 200 / 8 1 . CI DP96-00787R000300100001-9
 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
Island Volcanic
Non-Volcanic
Waterfall
F Structures
Absent Coastline r Glacial (Snowy)
I- Lake ` Non-Glacial (Snowy)
v VIl:Cl1
Hilly 11V
Water .
Present
Structures
Non-Urban
L Island
L Coastline
L Lake
River --E Flat
Hilly
Island
Peninsula
L_ Structures Coastline
Urban
Non-Volcaniu*
FIGURE 2 (U) DACOS HIERARCHY,FOR WATER-PRESENT TARGETS
IM 11
Approved For Release 2Y0D1Oo"76 !lE!00787R0003001 00001-9
 Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
Water Structures Contours Land Features
L Volcanic
Mountainous
Non-Volcanic
r Structures Rocky
Absent
Rugged Dunes
Snowy
Plowed
Flat C Fertile --E Verdant (Green)
Arid
Plowed
Water Structures Verdant (Green)
Absent Non-Urban
Structures
Urban
Fertile
Arid
I
Monuments
Dense Cityy
Non-Monuments
Ruins
Open City
Non-Ruins
Monuments
Dense City
Non-Monuments
Open City
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 tub-tree asking further questions. The final output is a list of pairs, the category type
with its corresponding confidence factor. Theseiconfidence 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.
Approved For Release 200 10'8l'I~~Cri~R~~9TU0787R000300100001-9
 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 question, 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
Approved For Release 200 LA6 IHE 787R000300100001-9
 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.
(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.
Approved For Release 2000f re} 6?-M%Flg P787R000300100001-9
 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 PA, 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.
Approved For Release 200087'I'9lL`1k!5jWVb787R000300100001-9
 Approved For Release 2000/08/10 ASS I ~rb00787R000300100001-9
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.
t
9. (U) Objective D, Task 1 (see Objective G, Task 1)
10. (U) Objective D, Task 2 (see Objective H, Task 3a)
11. (U) Objective E, Task 1--RA Effects on Marine Algae
s (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.
Approved For Release 20007c lcLCfi-RE! - 0787R000300100001-9
 Approved For Release 20U/f $/a,,Q A IMP .Q0787R000300100001-9
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 tlf 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 rata 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.
t (U) The study was designed to try to determine whether causal forces or intuitive
data sorting were responsible for any observed phi 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.
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 programmed 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
w
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 reinfording 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
Approved For Release 200"W-WAR787R000300100001-9
 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 e
time without feedback after each session.
The p riod of
examine what effect lack of feedback would have uonosubse uentproposed experiment was to
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 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 the
photograph). The specific opinion is that the RV impressions are richer, more varied, the
limited to visual for actual scenes than for the photographs. It follows that vie' and not as
ponses
might also be more detailed and not as limited to visual impressions. This hYpothe sisr
co resuld be
tested by conducting a series of viewings where targets are randomly chosen from
outdoor sites and photographs. a pool of
(U) A third proposal reached the stage of a formal written protocol. It was to test
idea that experienced viewers could perform just as well without a monitor as with the
present in the RV session. This experiment was rated as a high priority i a mmbn
f it were
successful it would (1) reduce the resources necessary to conduct an RV experiment, were
eliminate potential monitor cues, and (3) allow multiple viewers to work on the same (2)
once. target at
(U) Formal twork on experiments previously proposed and approved reem
further efforts on these proposals. Advanced viewers spent the second hal p pear
participating in several experiments with large numbers of remote viewin s, f of the year
the group 'ceased for the remainder of the year. g Weekly meetings of
16. (U) Objective F, Task 4--
Develop RV Training Hypothesis
(U) This task, originally intended to
Consultants International (CI), was converted to a consulting through a subcontract with
Mr. Gary Langford. During FY 1987, g tionship with the founder of CI,
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
Approved For Relea-UbL-GS 4MP96-00787R000300100001-9
 Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U)
Precognition experiment and (2) the Feedback experiment (see
Objective F, Tasks la and 1b).
s 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.
k
(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
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.
Approved For Release 200&"1&WFJ9_P787R000300100001-9
 Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100061-9
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
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
experiericed 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.
Approved For Release 20I& 0 0 f4 I-A Jl II 0787R000300100001-9
 Approved For Release 2000 ,( .$CLAgI -l 960787R000300100001-9
Target RV 1
randomly
selected 15 minutes
= Hypnosis
(30 minutes)
Proofread
I Feedback
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.'* 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.
v
(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.
(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
(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.
Approved For Release 20VOY0'8l' CTR-PXENP00787R000300100001-9
 Approved For Release 2yfflef AC rgf 0787R000300100001-9 *0 #10
(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
Y
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.
(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,
Approved For Release 200Y INI 0~-C1X UP'!0787R000300100001-9
 Approved For Release 20URCI 4tff11!6D787R000300100001-9
(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 ofsthe 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,
Approved For Release 200V/OSIc C1?ZQU0787R000300100001-9
 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.
(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 ?A of signal. If all 48 channels are high until the threshold voltage is
increased to 10 volts, then the noise floor will be 20 ?A. As was determined later, the alpha
particles produce 40 ?A pulses, not visible in the original system which exhibited 48 ?A of noise.
The discriminator was on the outside of the vacuum chamber and was connected to the strip
detector by twisteds-pair ribbon cables and two vacuum feedthroughs. The contacts on the
feedthrough connectors contributed 10 ?A of noise each. Because there was a connector on
each side of the feedthroughs, the noise contribution was 20 ?A. 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 ?A). This noise was reduced by relocating
the discriminator cards inside the vacuum chamber and eliminating the feedthroughs in this part
of the cirduit. Now, only logic signals pass through the feedthroughs in the base plate, not the
(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
Approved For Release 2 d C1LA&$ FIfO0787R000300100001-9
 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:
e The 2735B cards may be damaged in a high vacuum.
e 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 ?C. 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
Approved For Release V -9
 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 ?A.
(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 ?A 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-?F ao 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.
(U) After these modifications were made to the system, the noise floor was
brought down to 24 ?A and discriminated away. Only signals larger than 24 ?A 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 bee found and need to be eliminated for an infinite signal-to-noise ratio. These types are
as follows: {
? Spurious, intermittent, random noise.
? Parallel noise pulses.
(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
Approved For Release 20GW/N)U.A$,$E P95Q787R000300100001-9
 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
(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
k
used. If the telephone is lifted off the hook, it induces several parallel noise pulses 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.
(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
Approved For Release 200 MAi55tfED787R000300100001-9
 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.
(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 ?A.
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.
Approved For Release 200U '(GLA&SlF4ED787R000300100001-9
 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
17-32
0
0
0
0
0
0
33-48
0
0
0
0
0
0
Parallel noise pulses
1-16
0
1
1
1
1
1
17-32
1
1
1
0
1
2
33-48
1
1
1
1
1
1
0 0 0 0 0 0 0 0 0 0
0 30 36 833 3 4 78 0 14 0
0 0 0 0 0 0 1 0 1 0
1 1 1 1 0 1 1 1 1
0 40 22 848 0 8 71 1 13
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
1
0
1
0
0
0
1
0
0
1 2
17-32
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0 0
33-48
0
0
0
1
1
0
0
0
0
0
0
0
1
0
6 0
h. (U) Proposed System Testing And Modifications
(U) We are now able to demonstrate the presence of the alpha particle beam
well aboveithe 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:
Approved For Release 2( NCL1b &EID0787R000300100001-9
 Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(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.
(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.
(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
Approved For Release 200WNfo9 A i1FJg 0787R000300100001-9
 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',neets 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. WI 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.
Approved For Release 20U Lik9mmOW0787R000300100001-9
 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 Maya 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.
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 4nd host a conference of SRI Cognitive Science Program staff, subcontractors, and
designated consultants. The conference was held4at 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
Approved For Release 20U1C(LASOO787R000300100001-9
 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
(U) There are no deliverables required for this task.
Approved For Release 20`JN LA& 96QO787R000300100001-9
 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
- 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)
A Alpha Particle RA Equipment Development
- Labor charges for two Senior Electronics Technicians (hardware)
- Labor charges for Real-Time Computer Specialist (software).
Approved For Release 200A~~iS5JM1Ef)787R000300100001-9
 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.
Approved For Release 200t/PW 1.0 I HE&UR000300100001-9
 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:
C Postpone 3
Major formal report
Pilot, exploratory, approximately
3 to 5 pages
Wild guess, few paragraphw
Postpone.
(U) Table 3 summarizes the assignments on a task-by-task basis.
Table 3
(U) PRIORITY/DELIVERABLE ASSIGNMENTS FOR FY 1987
RV
Rating Task
B3 A4
B3 QI
Al Fa
Al Flb
A2 F2
A2 F3
B2 F4
B3 F5
B2 F6
A3 F7
B3 F9
B2 F10
Al G1
Name
Physical Correlation
Dowsing
RV/Precognition
Feedback + SL Tachistoscope
Video Disk
Advanced Training
Training Concepts
Stimulus/ Response Correlation
Hypnosis
Special Targets
Neuropsychology
Analytics
Computer Search
TECH SUPPORT t
Rating Task
Al Al
Al A2
Al A3
B2 B1
A2 C1
Al C2
C3 C3
A2 F1l
Al 11
A3 J3
A3 J4
Name
SOC Design
SOC Assessment
RV Analysis
Library
Med/Psych Baselines
MARS
PAS Review
MDS
PRL
Computer Maintenance
Computer Hardware
RA
Rating Task Name
Al El Delaware
B2 E2 Mind Science
A2 H1 Alpha-Particle
Al H2 QM Photon
Al H3 JFK
Al I i IDS-Dynamic
A2 J2 RA-Hardware
Al H3a SRI Part of JFK
GENERAL
Rating Task Name
A3 13 Theory Conference
Al 14 Subs Conference
Al Jl Administration
A3 J2 Publications
A3 J6 Travel
A3 J7 Additional SRI Staff
Al F8 Physiology Conference
Approved For Release 20c N?tA1S-ctptED787R000300100001-9
 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.
(U) AUTHORIZED TASK CHANGES TO FY 1987 STATEMENT OF WORK
TASK CHANGE
O
RIGINAL SOW
CHANGED SOW
Old
New
Task
$ K
Activity
Task
Activity
Justification
D-2
35.0
E&M correlates to
H-3a
SRI portion of
Provide recognition
dowsing
JFK
of significant SRI
participation in the
JFK project
F-7
30.0
"Applications" targets
F-7
Develop video
Augment talented
for novice viewers
"mass" screening
viewer pool
F-8
16.2
MEG with
F-8
Physiology
More effective search
Los Alamos
Conference
for physiology
correlates
(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
Balanced protocol
RV
F-lb
87.5
Subliminal Perception
Continue FY 1986
Experiment
feedback experiment
contains SL
$
G-1
100.0
Abstract to real-world
Continue FY 1986
Necessary pilot
target link
computer search
phase for link
activity
investigation
Approved For Release 200 ( 9L lifFIEE1787R000300100001-9
 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
Objective A--Protocols:
Design
Task 1
Task 2
Task 3
Task 4
Objective B--Library Task 1
Objective C--Psychophysical Task 1
Profiling Task 2
Task 3
Objective D--Field "Dowsing" Task 1 0-
Objective E--Continue RA Task 1
Experiments Task 2
Objective F--RV Parameters Task 1
Task 2
Task 3
Task 4
Task 5
Task 6
Task 7
Task 8
Task 9
Task 10
Task 11
010, Begin Key*
d End With Deliverable
? Deliverable
o End w/o Deliverable
* Tasks with no beginning indicator (11o.) are continuations of an FY 1986 effort.
Approved For Release QASIJlIFEI 00787R000300100001-9
 Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
(U) ENHANCED HUMAN PERFORMANCE INVESTIGATION--FY 1987
Objective G--Computer
"Search"
Objective H--RA Parameters Task 1
Task 2
Task 3a
Task 3
Objective 1--IDS Model / Task 1
Conferences Task 2
Task 3
Task 4
Objective J--Administrative / Task 1
Hardware Task 2
Task 3
Task 4
Task 5
Task 6
Task 7
0.
Begin Key*
End With Deliverable
? Deliverable
UNCLASSIFIED
* Tasks with no beginning indicator (/-) are continuations of an FY 1986 effort.
Approved For Release 2040,IQ$l &.On787R000300100001-9
 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.
Approved For Release 2081 f4CL.AIS.SWMEP787R000300100001-9
 Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
APPENDIX A
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.
Approved For Release 20i0787R000300100001-9
 Approved For Release 2000/08/10 : CIA-RDP96-00787R000300100001-9
UNCLASSIFIED
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:
Approved For Release 20U4 l (kJkWH f 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