Remote Perception at Stanford Research Institute
We present results of experiments suggesting the existence
of one or more perceptual modalities through which individuals
obtain information about their environment, although this information
is not presented to any known sense. The literature 1-3 and our
observations lead us to conclude that such abilities can be studied
under laboratory conditions.
We have investigated the ability of certain people to describe
graphical material or remote scenes shielded against ordinary
perception. In addition, we performed pilot studies to determine
if electroencephalographic (EEG) recordings might indicate perception
of remote happenings even in the absence of correct overt responses.
We concentrated on what we consider to be our primary responsibility
-- to resolve under conditions as unambiguous as possible the
basic issue of whether a certain class of paranormal perception
phenomena exists. So we conducted our experiments with sufficient
control, utilising visual, acoustic and electrical shielding,
to ensure that all conventional paths of sensory input were blocked.
At all times we took measures to prevent sensory leakage and
to prevent deception, whether intentional or unintentional.
Our goal is not just to catalogue interesting events, but
to uncover patterns of cause-effect relationships that lend themselves
to analysis and hypothesis in the forms with which we are familiar
in scientific study. The results presented here constitute a
first step towards that goal; we have established under known
conditions a data base from which departures as a function of
physical and psychological variables can be studied in future
work.
First, we conducted experiments with Mr. Uri Geller in which
we examined his ability, while located in an electrically shielded
room, to reproduce target pictures drawn by experimenters located
at remote locations. Second, we conducted double-blind experiments
with Mr. Pat Price, in which we measured his ability to describe
remote outdoor scenes many miles from his physical location.
Finally, we conducted preliminary tests using EEGs, in which
subjects were asked to perceive whether a remote light was flashing,
and to determine whether a subject could perceive the presence
of the light, even if only at a noncognitive level of awareness.
In preliminary testing Geller apparently demonstrated an
ability to reproduce simple pictures (line drawings) which had
been drawn and placed in opaque sealed envelopes which he was
not permitted to handle. But since each of the targets was known
to at least one experimenter in the room with Geller, it was not
possible on the basis of the preliminary testing to discriminate
between Geller's direct perception of envelope contents and perception
through some mechanism involving the experimenters, whether paranormal
or subliminal.
So we examined the phenomenon under conditions designed to
eliminate all conventional information channels, overt or subliminal.
Geller was separated from both the target material and anyone
knowledgeable of the material, as in the experiments of ref. 4.
In the first part of the study a series of 13 separate drawing
experiments were carried out over 7 days. No experiments were
deleted from the results presented here.
At the beginning of the experiment either Geller or the experimenters
entered a shielded room so that from that time forward Geller
was at all times visually, acoustically, and electrically shielded
from personnel and material at the target location. Only following
Geller's isolation from the experimenters was a target chosen
and drawn, a procedure designed to eliminate pre-experiment cueing.
Furthermore, to eliminate the possibility of pre-experiment target
forcing, Geller was kept ignorant as to the identity of the person
selecting the target and as to the method of target selection.
This was accomplished by the use of three different techniques:
(1) pseudo-random technique of opening a dictionary arbitrarily
and choosing the first word that could be drawn (Experiments 1-4);
(2) targets, blind to experimenters and subject, prepared independently
by SRI scientists outside the experimental group (following Geller's
isolation) and provided to the experimenters during the course
of the experiment (Experiments 5-7, 11-13); and (3) arbitrary
selection from a target pool decided upon in advance of daily
experimentation and designed to provide data concerning information
content for use in testing specific hypotheses (Experiments 8-10).
Geller's task was to reproduce with pen on paper the line drawing
generated at the target location. Following a period of effort
ranging from a few minutes to half an hour, Geller either passed
(when he did not feel confident) or indicated he was ready to
submit a drawing to the experimenters, in which case the drawing
was collected before Geller was permitted to see the target.
To prevent sensory cueing of the target information, Experiments
1 through 10 were carried out using a shielded room in SRI's facility
for EEG research. The acoustic and visual isolation is provided
by a double-walled steel room, locked by means of an inner and
outer door, each of which is secured with a refrigerator-type
locking mechanism. Following target selection when Geller was
inside the room, a one-way audio monitor, operating only from
the inside to the outside, was activated to monitor Geller during
his efforts. The target picture was never discussed by the experimenters
after the picture was drawn and brought near the shielded room.
In our detailed examination of the shielded room and the protocol
used in these experiments, no sensory leakage has been found.
The conditions and results for the 10 experiments carried
out in the shielded room are displayed in Table 1 and Fig. 1.
All experiments except 4 and 5 were conducted with Geller inside
the shielded room. In Experiments 4 and 5, the procedure was
reversed. For those experiments in which Geller was inside the
shielded room, the target location was in an adjacent room at
a distance of about 4 m, except for Experiments 3 and 8, in which
the target locations were, respectively, an office at a distance
of 475 m and a room at a distance of about 7 M,
A response was obtained in all experiments except Numbers
5-7. In Experiment 5, the person-to-person link was eliminated
by arranging for a scientist outside the usual experimental group
to draw a picture, lock it in the shielded room before Geller's
arrival at SRI, and leave the area. Geller was then led by the
experimenters to the shielded room and asked to draw the picture
located inside the room. He said that he got no clear impression
and therefore did not submit a drawing. The elimination of the
person-to-person link was examined further in the second series
of experiments with this subject.
Experiments 6 and 7 were carried out while we attempted to
record Geller's EEG during his efforts to perceive the target
pictures. The target pictures were, respectively, a tree and
an envelope. He found it difficult to hold adequately still for
good EEG records, said that he experienced difficulty in getting
impressions of the targets, and again submitted no drawings.
Experiments 11 through 13 were carried out in SRI's Engineering
Building, to make use of the computer facilities available there.
For these experiments, Geller was secured in a double-walled,
copper-screen Faraday cage 54 m down the hall and around the corner
from the computer room. The Faraday cage provides 120 dB attenuation
for plane wave radio frequency radiation over a range of 15 kHz
to 1 GHz. For magnetic fields the attenuation is 68 dB at 15
kHz and decreases to 3 dB at 60 Hz. Following Geller's isolation,
the targets for these experiments were chosen by computer laboratory
personnel not otherwise associated with either the experiment
or Geller, and the experimenters and subject were kept blind as
to the contents of the target pool.
For Experiment 11, a picture of a kite was drawn on the face
of a cathode ray tube display screen, driven by the computer's
graphics programme. For Experiment 12, a picture of a church
was drawn and stored in the memory of the computer. In Experiment
13, the target drawing, an arrow through a heart (Fig. 2c), was
drawn on the face of the cathode ray tube and then the display
intensity was turned off so that no picture was visible.
To obtain an independent evaluation of the correlation between
target and response data, the experimenters submitted the data
for judging on a 'blind' basis by two SRI scientists who were
not otherwise associated with the research. For the 10 cases
in which Geller provided a response, the judges were asked to
match the response data with the corresponding target data (without
replacement). In those cases in which Geller made more than one
drawing as his response to the target, all the drawings were combined
as a set for judging. The two judges each matched the target
data to the response data with no error. For either judge such
a correspondence has an a priori probability, under
the null hypothesis of no information channel, of p 1/(10!)
= 3 x 10 to the power -7.
A second series of experiments was carried out to determine
whether direct perception of envelope contents was possible without
some person knowing of the target picture.
One hundred target pictures of everyday objects were drawn
by an SRI artist and sealed by other SRI personnel in double envelopes
containing black cardboard. The hundred targets were divided
randomly into groups of 20 for use in each of the three days'
experiments.
On each of the three days of these experiments, Geller passed.
That is, he declined to associate any envelope with a drawing
that he made, expressing dissatisfaction with the existence of
such a large target pool. On each day he made approximately 12
recognisable drawings, which he felt were associated with the
entire target pool of 100. On each of the three days, two of
his drawings could reasonably be associated with two of the 20
daily targets. On the third day, two of his drawings were very
close replications of two of that day's target pictures. The
drawings resulting from this experiment do not depart significantly
from what would be expected by chance.
In a simpler experiment Geller was successful in obtaining
information under conditions in which no persons were knowledgeable
of the target. A double-blind experiment was performed in which
a single 3/4 inch die was placed in a 3 x 4 x 5 inch steel box.
The box was then vigorously shaken by one of the experimenters
and placed on the table, a technique found in control runs to
produce a distribution of die faces differing nonsignificantly
from chance. The orientation of the die within the box was unknown
to the experimenters at that time. Geller would then write down
which die face was uppermost. The target pool was known, but the
targets were individually prepared in a manner blind to all persons
involved in the experiment. This experiment was performed ten
times, with Geller passing twice and giving a response eight times.
In the eight times in which he gave a response, he was correct
each time. The distribution of responses consisted of three 2s,
one 4, two 5s, and two 6s. The probability of this occurring
by chance is approximately one in 10 to the power 6.
In certain situations significant information transmission
can take place under shielded conditions. Factors which appear
to be important and therefore candidates for future investigation
include whether the subject knows the set of targets in the target
pool, the actual number of targets in the target pool at any given
time, and whether the target is known by any of the experimenters.
It has been widely reported that Geller has demonstrated
the ability to bend metal by paranormal means. Although metal
bending by Geller has been observed in our laboratory, we have
not been able to combine such observations with adequately controlled
experiments to obtain data sufficient to support the paranormal
hypothesis.
A study by Osis 5 led us to determine whether a subject could
describe randomly chosen geographical sites located several miles
from the subject's position and demarcated by some appropriate
means (remote viewing). This experiment carried out with Price,
a former California police commissioner and city councilman, consisted
of a series of double-blind, demonstration-of-ability tests involving
local targets in the San Francisco Bay area which could be documented
by several independent judges. We planned the experiment considering
that natural geographical places or man-made sites that have existed
for a long time are more potent targets for paranormal perception
experiments than are artificial targets prepared in the laboratory.
This is based on subject opinions that the use of artificial
targets involves a 'trivialisation of the ability' as compared
with natural preexisting targets.
In each of nine experiments involving Price as subject and
SRI experimenters as a target demarcation team, a remote location
was chosen in a double-blind protocol. Price, who remained at
SRI, was asked to describe this remote location, as well as whatever
activities might be going on there.
Several descriptions yielded significantly correct data pertaining
to and descriptive of the target location.
In the experiments a set of twelve target locations clearly
differentiated from each other and within 30 minutes driving time
from SRI had been chosen from a target-rich environment (more
than 100 targets of the type used in the experimental series)
prior to the experimental series by an individual in SRI management,
the director of the Information Science and Engineering Division,
not otherwise associated with the experiment. Both the experimenters
and the subject were kept blind as to the contents of the target
pool, which were used without replacement.
An experimenter was closeted with Price at SRI to wait 30
minutes to begin the narrative description of the remote location.
The SRI locations from which the subject viewed the remote locations
consisted of an outdoor park (Experiments 1, 2),- the double-walled
copper-screen Faraday cage discussed earlier (Experiments 3, 4,
and 6-9), and an office (Experiment 5). A second experimenter
would then obtain a target location from the Division Director
from a set of traveling orders previously prepared and randomised
by the Director and kept under his control. The target demarcation
team (two to four SRI experimenters) then proceeded directly to
the target by automobile without communicating with the subject
or experimenter remaining behind. Since the experimenter remaining
with the subject at SRI was in ignorance both as to the particular
target and as to the target pool, he was free to question Price
to clarify his descriptions. The demarcation team then remained
at the target site for 30 minutes after the 30 minutes allotted
for travel. During the observation period, the remote-viewing
subject would describe his impressions of the target site into
a tape recorder. A comparison was then made when the demarcation
team returned.
Price's ability to describe correctly buildings, docks, roads,
gardens, and so on, including structural materials, color, ambiance,
and activity, sometimes in great detail, indicated the functioning
of a remote perceptual ability. But the descriptions contained
inaccuracies as well as correct statements. To obtain a numerical
evaluation of the accuracy of the remote viewing experiment, the
experimental results were subjected to independent judging on
a blind basis by five SRI scientists who were not otherwise associated
with the research. The judges were asked to match the nine locations,
which they independently visited, against the typed manuscripts
of the tape-recorded narratives of the remote viewer. The transcripts
were unlabelled and presented in random order. The judges were
asked to find a narrative which they would consider the best match
for each of the places they visited. A given narrative could
be assigned to more than one target location. A correct match
requires that the transcript of a given date be associated with
the target, of that date. Table 2 shows the distribution of the
judges' choices.
Among all possible analyses, the most conservative is a permutation
analysis of the plurality vote of the judges' selections assuming
assignment without replacement, an approach independent of the
number of judges. By plurality vote, six of the nine descriptions
and locations were correctly matched. Under the null hypothesis
(no remote viewing and a random selection of descriptions without
replacement), this outcome has an a priori probability
of p = 5.6 x 10 to the power -4 since, among all possible
permutations of the integers one through nine, the probability
of six or more being in their natural position in the list has
that value. Therefore, although Price's descriptions contain
inaccuracies, the descriptions are sufficiently accurate to permit
the judges to differentiate among the various targets to the degree
indicated.
An experiment was undertaken to determine whether a physiological
measure such as EEG activity could be used as an indicator of
information transmission between an isolated subject and a remote
stimulus. We hypothesised that perception could be indicated by
such a measure even in the absence of verbal or other overt indicators.
6, 7
It was assumed that the application of remote stimuli would
result in responses similar to those obtained under conditions
of direct stimulation. For example, when normal subjects are
stimulated with a flashing light, their EEG typically shows a
decrease in the amplitude of the resting rhythm and a driving
of the brain waves at the frequency of the flashes 8. We hypothesised
that if we stimulated one subject in this manner ( a sender),
the EEG of another subject in a remote room with no flash present
(a receiver), might show changes in alpha (9-11 Hz) activity,
and possibly EEG driving similar to that of the sender.
We informed our subject that at certain times a light was
to be flashed in a sender's eyes in a distant room, and if the
subject perceived that event, consciously or unconsciously, it
might be evident from changes in his EEG output. The receiver
was seated in the visually opaque, acoustically and electrically
shielded double-walled steel room previously described. The sender
was seated in a room about 7 m from the receiver.
To find subjects who were responsive to such a remote stimulus,
we initially worked with four female and two male volunteer subjects,
all of whom believed that success in the experimental situation
might be possible. These were designated 'receivers'. The senders
were either other subjects or the experimenters. We decided beforehand
to run one or two sessions of 36 trials each with each subject
in this selection procedure, and to do a more extensive study
with any subject whose results were positive.
A Grass PS-2 photostimulator placed about 1 m in front of
the sender was used to present flash trains of 10 s duration.
The receiver's EEG activity from the occipital region (0z), referenced
to linked mastoids, was amplified with a Grass SP-1 preamplifier
and associated driver amplifier with a bandpass of 1-120 Hz. The
EEG data were recorded on magnetic tape with an Ampex SP 300 recorder.
On each trial, a tone burst of fixed frequency was presented
to both sender and receiver, and was followed in one second by
either a 10 s train off lashes or a null flash interval presented
to the sender. Thirty-six such trials were given in an experimental
session, consisting of 12 null trials - no flashes following the
tone - 12 trials of flashes at 6 f.p.s. and 12 trials of flashes
at 16 f.p.s., all randomly intermixed, determined by entries from
a table of random numbers. Each of the trials generated an 11-s
EEG epoch. The last 4 s of the epoch was selected for analysis
to minimise the desynchronising action of the warning cue. This
4-s segment was subjected to Fourier analysis on a LINC 8 computer.
Spectrum analyses gave no evidence of EEG driving in any
receiver, although in control runs the receivers did exhibit driving
when physically stimulated with the flashes. But of the six subjects
studied initially, one subject (H. H.) showed a consistent alpha
blocking effect. We therefore undertook further study with this
subject.
Data from seven sets of 36 trials each were collected from
this subject on three separate days. This comprises all the data
collected to date with this subject under the test conditions
described above. The alpha band was identified from average spectra,
then scores of average power and peak power were obtained from
individual trials and subjected to statistical analysis,
Of our six subjects, H. H. had by far the most monochromatic
EEG spectrum. Figure 3 shows an overlay of the three averaged
spectra from one of this subject's 36-trial runs, displaying changes
in her alpha activity for the three stimulus conditions.
Mean values for the average power and peak power for each
of the seven experimental sets are given in Table 3. The power
measures were less in the 16 f.p.s. case than in the 0 f.p.s.
in all seven peak power measures and in six out of seven average
power measures. Note also the reduced effect in the case in which
the subject was informed that no sender was present (Run 3).
It seems that overall alpha production was reduced for this run
in conjunction with the subject's expressed apprehension about
conducting the experiment without a sender. This is in contrast
to the case (Run 7) in which the subject was not informed.
Siegell's two-tailed t approximation to the nonparametric
randomisation test 9 was applied to the data from all sets, which
included two sessions in which the sender was removed. Average
power on trials associated with the occurrence of 16 f.p.s. was
significantly less than when there were no flashes t = 2.09,
d.f. = 118, P < 0.04). The second measure, peak power, was
also significantly less in the 16 f. p. s. conditions than in
the null condition ( t = 2.16, d.f. = 118, P < 0.03). The
average response in the 6 f.p.s. condition was in the same direction
as that associated with 16 f.p.s., but the effect was not statistically
significant.
Spectrum analyses of control recordings made from saline
with a 12 kOhm resistance in place of the subject with and without
the addition of a 10 Hz, 50 microvolt test signal applied to the
saline solution, revealed no indications of flash frequencies,
nor perturbations of the 10 Hz signal. These controls suggest
that the results were not due to system artifacts. Further tests
also gave no evidence of radio frequency energy associated with
the stimulus.
Subjects were asked to indicate their conscious assessment
for each trial as to which stimulus was generated. They made
their guesses known to the experimenter via one-way telegraphic
communication. An analysis of these guesses has shown them to
be at chance, indicating the absence of any supraliminal cueing.
So, arousal as evidenced by significant alpha blocking occurred
only at the noncognitive level of awareness.
We hypothesised that the protocol described here may prove
to be useful as a screening procedure for latent remote perceptual
ability in the general population.
From these experiments we conclude that:
- A channel exists whereby information about a remote location
can be obtained by means of an as yet unidentified perceptual
modality.
- As with all biological systems, the information channel
appears to be imperfect, containing noise along with the signal.
- While a quantitative signal-to-noise ratio in the information-theoretical
sense cannot as yet be determined, the results of our experiments
indicate that the functioning is at the level of useful information
transfer.
It may be that remote perceptual ability is widely distributed
in the general population, but because the perception is generally
below an individuals level of awareness, it is repressed or not
noticed. For example, two of our subjects (H.H. and P.P.) had
not considered themselves to have unusual perceptual ability before
their participation in these experiments.
Our observation of the phenomena leads us to conclude that
experiments in the area of so-called paranormal phenomena can
be scientifically conducted, and it is our hope that other laboratories
will initiate additional research to attempt to replicate these
findings.
This research was sponsored by The Foundation for Parasensory
Investigation, New York City. We thank Mrs. Judith Skutch, Dr.
Edgar D. Mitchell of the Institute of Noetic Sciences - as well
as our SRI associates, Mr. Bonnar Cox, Mr. Earle Jones and Dr.
Dean Brown - for support and encouragement. Constructive suggestions
by Mrs. Jean Mayo, Dr. Charles Tart, University of California,
and Dr. Robert Ornstein and Dr. David Galin of the Langley Porter
Neuropsychiatric Institute are acknowledged.
Russell Targ
Harold Puthoff
Electronics and Bioengineering Laboratory,
Stanford Research Institute,
Menlo Park, California 94025
Fig 1. Target pictures and responses drawn by
Uri Geller under shielded conditions.
Fig 2. Computer drawings and responses drawn by
Uri Geller. a, Computer drawing stored on video display; b, computer
drawing stored in computer memory only; c, computer drawing stored on
video display with zero intensity.
Fig 3. Occipital EEG spectra, 0-20 Hz, for one
subject (H. H.) acting as receiver, showing amplitude changes in the 9-11
Hz band as a function of strobe frequency. Three cases: 0,6, and 16 f.p.s.
(12 trial averages).
Tables
1 Pratt, J., Rhine, J.B., Stuart, C., and Greenwood, J.,
Extra Sensory Perception after Sixty Years (Henry Holt,
New York, 1940).
2 Soal, S., and Bateman, F., Modern Experiments in Telepathy
(Faber and Faber, London, 1954).
3 Vasilliev, L.L., Experiments in Mental Suggestion
(ISMI Publications, Hampshire, England, 1963).
4 Musso, J.R., and Granero, M., J. Parapsychology,
37, 13-37 (1973).
5 Osis, K., ASPR Newsletter, No. 14 (1972).
6 Tart, C.T., Physiological Correlates of Psi Cognition,
Int. J. Parapsychology, V, No. 4 (1963).
7 Dean, E.D., Int. J. Neuropsychiatry, 2 (1966).
8 Hill, D., and Parr, G., Electroencephalography: A Symposium
on its Various Aspects (Macmillan, New York, 1963).
9 Siegel, S., Nonparametric Statistics for the Behavioural
Sciences, pp. 152-156 (McGraw-Hill, New York, 1956).
Back to main index.
Back to Uri Geller's home page.