VISUAL EVOKED RESPONSES:
THEIR RELATIONSHIP TO INTELLIGENCE
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
TIE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The author wishes to express his gratitude to
Dr. Nathan Perry, the chairman of his committee, for
his encouragement during the planning stages of this
study, and for his continuing interest and assistance
during its completion. Appreciation is also extended
to the other members of the committee: Dr. Louis
Cohen, Dr. Herschel Elliott, Dr. Paul Satz, and Dr.
Vernon Van De Riet.
TABLE OF CONTENTS
ACKNOWLEDGMENTS . . . .
LIST OF TABLES . . . .
LIST OF FIGURES.......
INTRODUCTION . . . . . .
Background. . .. .
The Present Study . . . .
Apparatus . * * * * **
Procedure . . . * *
Data Processing and Analysis . .
RESULTS . . . . . . ..
Preliminary Analyses: Parametric
Data . .. . . . .
Main Analyses: VERs and Intelligence
IV DISCUSSION . . . . . .
V SUMARY.. . . . .
1 Summary of Statistical Analyses .
2 Rules for Discrimination of Peaks:
Criteria I and II . . .
3 Intercorrelation Matrix of
Test Scores * * *o .
LIST OF REFERENCES . . . . . .
BIOGRAPHICAL SKETCH . . . . . .
. . . . . . ii
* . . . .. iv
0 6 0 0 0 0 0 0 0 0 V
* 0 0 0
* 0 *
. . .
LIST OF TABLES
1 Parametric Studies of Scalp VER
in Man . . . . . . . . 10,11
2 Means, Ranges, and Standard Deviations
of Major Intellectual and Perceptual-
Motor Tests for 37 Subjects Used in
Most Analyses. . . . . . . 48
3 Mean Latencies of the Left-Motor
VER Peaks of the Trial 1 Presenta-
tion which were Significantly
Different Among IQ Groups. . . . 57
4 Correlations Between the Latency of
all Trial 1 Peaks (Criteria I and
II) and Full Scale WAIS . . . . 61
5 Correlations of Intellectual and
Perceptual-Motor Tests with Latency
of Peak 2 (Criterion I, Trial 1) . . 63
LIST OF FIGURES
1 Two consecutive visual evoked responses,
recorded at the occipital, left-motor,
and right-motor locations, from a high,
medium, and low IQ subject *. * 53
2 Mean latency of each left-motor VER
peak for the high, medium, and low IQ
groups . . * * * * 55
3 Left-motor responses of the four most
intelligent and the four least intel-
ligent subjects . . * * * * 59
4 Complex left-motor responses of three
high or medium IQ subjects contrasted
with their simpler right-motor responses 60
5 Scatter plots showing the relationship
between Full Scale scores of the
Wechsler Adult Intelligence Scale,
and the latencies of Deaks 2 and 3
(Trial 1, Criterion I) of VERs recorded
from the left-motor area. *. . * * 62
6 Comparison of Chalke and Ertl's (1965)
left-motor VER peaks 3, 4-, and. 5 with
left-motor VER peaks 2, 3, and 4 of the
present study * * * 68
The relationship between the neural processes of
the human brain and intellectual activity is one of the
most challenging of all problems facing the sciences of
human behavior. Yet the relevant literature is either
non-existent (Wechsler, 1958) or highly fragmented, and
a recent review has concluded that there are "no general
points of view about the neurophysiological correlates of
intelligence..." (Ferguson, 1965, p. 55). It is there-
fore quite understandable that a recent study (Chalke and
Ertl, 1965) which found that certain electrical proper-
ties of the cortex varied with psychometric intelligence
has been received with both excitement and scepticism.
The variable which these authors related to
intelligence was the visual evoked response (VER), the
reaction of the brain to visual stimuli. It was obtained
by recording, through scalp electrodes placed over the
motor cortex, the neural responses of subjects to a
flashing light. It had been theorized that the late com-
ponents of evoked potentials are the electrical signs of
information processing or associative activity in the
brain .(John, Ruchkin, and Villega 1964; Uttal and
Cook, 1964). From this, Chalke and Erti postulated that
"a biologically efficient organism should process infor-
mation more rapidly than a less efficient organism and
that the delay of components of the evoked potentials
is a measure of the efficiency of this process" (p.
1319). A pilot study (Barry and Ertl, 1963) had provided
support for this hypothesis.
On the basis of these assumptions, Chalke and
Ertl recorded visual evoked responses from 33 postgradu-
ate students of superior intelligence, 11 Army Cadets of
low average intelligence, and 4 individuals who were
mentally retarded.1 They compared each of these groups
with the other on the latencies of five peaks which were
claimed to have been identified in most of the subjects.
Depending on the groups compared, they found that the
latency of the later peaks was shortest for the superior
subjects and longest for the retardates, with the low
normal falling between them. Mann-Whitney U Tests were
used to test the differences between means, which were
found significant at the .05 and .01 levels.
Since the publication of Chalke and Ertl's (1965)
study, several other authors have reported related
Intelligence was measured by the Otis Scale.
findings. Probably the most important comparison to
their study is that of Rhodes, Dustman, and Beck (1967).
They recorded VERs from parietal and occipital scalps of
20 bright children (Full Scale WISO scores ranging from
120-140, mean of 130), and 20 dull children (Full Scale
WISC scores ranging from 70-90, mean of 79). All sub-
jects were 10 and 11 years old and each group was com-
posed of 10 boys and 10 girls (analyzed separately).
Aside from the latency of one late occipital component,
which was significantly shorter for bright children,
Rhodes and his colleagues found no IQ-latency relation-
ships. However, they did find smaller VERs, lack of
hemispheric differences in amplitude, and.less waveform
stability in their low IQ subjects. These amplitude
differences did not appear to be simply a function of
maturation, and Rhodes et al. suggested that dull chil-
dren may be less "alert" than bright children. This non-
cognitive explanation is in sharp contrast to the cogni-
tive theorizing of Chalke and Ertl (1965).
Several other studies have related prolonged
latency to pathological states which involve decreased
intellectual ability. Straumanis, Shagass, and Schwartz
(1965) compared the occipital VERs of 20 elderly pa-
tients with severe chronic brain syndrome due to cere-
bral arteriosclerosis with those of 18 healthy subjects
of the same age and sex. The patients, whose IQs were
much lower, had prolonged latencies of peaks after 100
msec. It was speculated that these VERs were physio-
logical conebcitants of clouded sensorium. However, as
Shagass (1967a) commented in a separate paper, all of
these patients were disoriented, the majority had
abnormal EEGS, and their evoked responses had many
features similar to those found in drowsiness. When
the data for healthy controls alonewere examined, how-
ever, the significant correlations between abbreviated
Weohsler IQ (based.on the block design and picture
completion subtests of the WAIS) and the latencies of
sequential evoked response peaks were in exactly the
opposite direction reported by Chalke and Erti (1965).
Instead of faster latencies being associated with higher
IQ, longer latencies were found to be associated with
the higher IQ.2 In a similar vein, Bigun and Whitmer
(1967) compared the visual and somatosensory evoked
responses of 24 Mongoloid children, ranging in age from
6 to 17, with those of 24 children of average intelli-
gence and matched for chronological age. Among the
results was the finding that the late components of the
Mongoloid VER when recorded from the occiput were of
longer latencies than normal. This study must also be
2Rank-order correlations ranged from .19 to
*75, with a mean of .46.
viewed with considerable reservation, however, since the
Mongoloid group, like the chronic brain syndrome patients
of Straumanis, et al. (1965), has a number of well-known
Finally, Callaway and Jones (1968), using audi-
tory evoked responses, found in a pilot study involving
21 children, aged 3 to 16, that latency was negatively
related to performance on a group of simple cognitive
measures. These included such tasks as finding or match-
ing odd forms and colors, and the 10 test-latency corre-
lations reported ranged from -.12 to -.48, with seven of
the r's falling between -.19 and -.31.
It is apparent from these findings that, as
Shagass (1967a) has noted, the question of intelligence
vad evoked response latency is still very much open.
Indeed, further more rigorously controlled research,
especially replications of Chalke and Ertl's (1965)
original positive findings, has been called for by a
number of authors (e.g. Callaway, 1966; Liberson, 1967).
The present study was designed, therefore, to replicate
and extend Chalke and Ertl's (1965) original pilot study
in hopes of remedying its major shortcomings.
Before proceeding to describe this study, how-
ever, it will be profitable to examine the recent re-
search which has pointed towards the possible interrela-
tionships between the visual evoked response, the
electroencephalographic activity from which it is drawn,
and a number of perceptual, cognitive, and emotional
states. This review will (1) examine the nature of the
visual evoked response; (2) investigate the possibility
that the VER reflects other complex mental processes
besides intelligence; and (3) explore the possibility
that the, VER may be related to the EEG, and the EEG to
Origin and Parameters of the VER
The visual evoked response (VER) can be recorded
in man from large numbers of specific and non-specific
subcortical and cortical areas, such as the hippocampus
(Brazier, 1964), the medial thalamus (Ervin and Mark,
1964), the optic tract, mesencephalic reticular forma-
tion, and visual cortex (Guerrero-Figueroa and Heath,
1964); and the frontal and temporal lobes (Walter,
1964).3 The extensive interaction of visual responses
with those of other sensory modalities has also been
documented (Walter, 1964).
Until the development of modern computer tech-
niques, however, it was nearly impossible to isolate
3The ubiquitous appearance of responses to
visual stimuli has also been emphasized by animal
studies (Fischer-Williams, 1966; Torres and Perpich,
1966; Vaughan and Gross, 1966).
the small evoked responses from the scalps of waking sub-
jects since they were embedded in the much larger "noise"
of the EEG. Computers can now sum the evoked responses
which correspond in time to the sensory stimuli (e.g.
flashes striking the retina) to form a single multi-
waved representation of the individual potentials. This
becomes large enough to be separable from the EEG, which
does not sum and averages to a relatively straight line.
Visual evoked potentials can be recorded from
all over the scalp (Remond, 1964; Kooi and Bagchi, 1964a;
Clones, Kohn, and Lifshitz, 1964; Puchinskoya, 1966),
but have been most often recorded from areas overlying
the occipital areas 17, 18, and 19 of the visual cortex.
While there is little doubt that placements other than
those over the occiput reflect non-specific responses to
visual stimuli, there has been considerable debate as to
the origin of the occipital scalp response. Several
studies have suggested, on the basis of the early la-
tency of the first VsE components and their disappearance
in patients with visual pathway lesions, that these
waves are derived from primary visual cortex (Ciganek,
1961; Vaughan and Katzman, 1964; Vaughan, 1966). Other
studies, however, have provided evidence that the entire
response arises from non-specific or secondary visual
areas. Thus, Hill and Parr (1963) note that the sec-
ondary areas 18 and 19 probably contribute much more
than primary area 17 to the scalp response because of
the secondary area's much closer proximity to the skull.
Ciganek (1966, 1967), furthermore, has reversed his
former position (Ciganek, 1961) upon finding that loud
clicks evoked the same waveform over the occiput as did
flashes (although stimuli from different sensory modali-
ties'can be differentiated on the basis of latency and
amplitude), suggesting that both responses reflect
independent non-specific sensory pathways. Corletto,
Gentilomo, Rosadini, Rossi,' and Zatoni (1967), who
recorded VERs from an epileptic patient before and af-
ter ablation of the occipital pole, implicated both
primary and secondary areas, as well as those of subcor-
tical origin, in the response. Finally, Walter (1964)
has pointed out that visual responses recorded in pri-
mary cortex are stereotyped and do not particularly
habituate, while the scalp response is variable and
extremely subject to habituation. In addition, visual
responses can be recorded in frontal cortex whose earli-
est latency (about 20 msec) approximates that of the
occipital cortex. He concluded from these data that the
scalp response is probably an inseparable combination of
primary and secondary reactions.* There is more
4The issue is complicated by the fact that af-;
ferent impulses from the eye, the labyrinth, and the
non-specific thalamus converge on neurons in the visual
cortex, suggesting that the visual areas not only are
specialized receiving areas, but are also involved in
integration of various sensory inputs (Jung, 1961).
agreement, however, that the later occipital waves
reflect non-specific processes and can also be found all
over the scalp (Remond, 1964; Wicke, Donchin, and
Lindsley, 1964; Uttal and Cook, 1964).
In any case, considerable data have been accumu-
lated concerning the form, distribution, and parameters
of this scalp response. Some of these studies are
listed in Table 1. A semblance of order and agreement
is now beginning to emerge, and several reviews (Berga-
mini and Bergamasco, 1967; Jonkman, 1967) have begun
to identify common and stable features of the VER, its
relationship to other sensory evoked potentials, and
important stimulus and subject parameters. In addi-
tion, a number of symposia have continued to stimulate
interest (Cobb and Morooutti, 1967; Katzman, 1964).
Perceptual, Cognitive, and Emotional Correlates of VERs
Concurrent with the primarily parametric studies
which have been done with the VER, there have been a
large number of experiments based on the premise that
the VER, as a response of the brain, should correlate
with significant features of human experience or be-
havior.5 Thus, the VER has been investigated as
uor a fuller discussion and review of the
degree to which this assumption has been borne out, see
Uttal (1965) and Callaway (1966).
PARAMETRIC STUDIES OF SCALP VER IN MAN
VARIABLE STUDIED AUTHORS
Childhood through Old Age
Drugs and Anesthesia
Site of Retinal
Dustman and Beck (1966);
Copenhaver and Perry (1964);
Ebe, et .l* (1962)
Engel (1964); Hrbek, et a.
1966); Ferriss, ,t .
1966); Ellingson 966b,
1968)' Rose and Ellingson
Dustman and Beck (1965b)
Kooi and Bagchi (1964a);
Vanzulli,, al. (1960);
Shagass (1967b)i Bergamini
and Bergamasco (1967)'
Domino and Corssen (1964)
.Vaughan and Katzman (1964);
Copenhaver and Beinhocker
Liberson (1966); Eilingson
and McBeath (1967):
Eason, rt al. (1967);
Dawson, et al. (1968)
TABLE 1 Continued
VARIABLE STUDIED AUTHORS
Latency and Amplitude Ciganek (1961, 1966, 1967);
Cobb and Dawson (1960);
Rietveld (1963) Gastaut
and Regis (1964)
Configuration over Werre and Smith (1964);
Location Crighel and Ciurea (1966);
Blatt and Offner (1966);
Remond (1964); Matsumiya,
et a. (1968)
Stability over Time Dustman and Beck (1963)
Habituation Perry and Copenhaver
Descriptive Laws and Dubouloz et al. (1966);
Statistical Analysis Donchin (19665
Area and Intensity Tepas and Armington (1962);
Shipley, et al. (1966);
Clynes, et al (1964);
Kitajiimat 17); Vaughan,
et al. (1966)
Pattern Spehlmann (1965); Gross,
et al. (1967); Rietveld
17967); John, et al. (1967)
Color Cavonius (1965); Clynes
and Kohn (1964); Shipley,
et al. (1965)
aMany of these studies are concerned with more
than one variable. For simplicity's sake, however, each
study is listed only one time.
a correlate of such perceptual experiences as stabilized
retinal images (Lehrmann, Becker, and Fender, 1965,
1967), subjective brightness (Wicke, Donchin, and Lindsley,
1964; Bartlett and White, 1965), paired flashes (Donchin,
Wicke, and Lindsley, 1963), perceptual masking and
enhancement (Donchin and Lindsley, 1965), metacontrast
(Schiller and Chorover, 1966; Vaughan and Silverstein,
1968), and perceived number (Harter and White, 1967).6
Buchsbaum and Silverman (1968) and Callaway and Jones
(1968) have found that individuals who showed a reduction
of the experienced intensity of stimulation, as inferred
from their performances on a figural after-effects pro-
cedure, also showed a comparable tendency in the ampli-
tude of their evoked responses. An interesting exten-
sion of this work has consisted of studies investigating
the effect of hypnotic and waking suggestions on the
visual response (Beck, Dustman, and Beier, 1966; Emrich,
1966; Guerrero-Figueroa and Heath, 1964; Hernandez-
Peon and Donoso, 1959; Clynes, Kohn, and Lifshitz, 1964;
Beck and Barolin, 1965; and Plum, 1965).
Possibly more relevant to the study of VER cor-
relates of intelligence, however, are studies which have
been concerned with attentional and expectancy states,
6Excellent reviews of this and other work are
provided by Goff (1967), Vaughan (1966) and White and
with the cognitive and emotional significance of the
stimuli used to elicit the response, and with evoked
response correlates of psychiatric conditions.
Thus, the amplitudes of visual evoked responses
have been found to increase with attentional states pro-
duced, for example, by counting the flashes, and to de-
crease with distraction to another stimulus (Garcia-Austt,
Bogacz, and Vanzulli, 1964). Similarly, visual responses
to flashes which s failed to detect during a vigilance
task were typically reduced in amplitude when compared
with those of an equal number of signals which were
correctly detected (Haider, Spong, and Lindsley, 1964).
This finding was later extended to both visual and
auditory stimuli recorded from the occipital and temporal
areas under three conditions of selective attentiveness
(Spong, Haider, and Lindsley, 1965). Another approach to
attentional correlates of the VER was made by Chapman
and Bragdon (1964) and Chapman (1965) who studied the ef-
fects "meaningfulness" or "task relevance" on VERs by
having subjects solve simple problems that required the
perception of visual stimuli presented in a sequence
which also included stimuli which were irrelevant to the
task. They found that relevant stimuli consistently
evoked larger responses than did the irrelevant stimuli.
More recently, these findings have been extended by
more sophisticated studies which have
controlled variables such as alertness and excitation
level of the Ss (Donchin and Cohen, 1967; Spong and
Lindsley, 1968).7 In another vein, Sutton, Braren,
Zubin, and John (1965) found that evoked responses to
sound and light stimuli showed differences as a function
of the S's degree of uncertainty with respect to the
sensory modality of the stimulus to be presented.
Differences were also found in the evoked potentials as
a function of whether or not the sensory modality of the
stimulus was anticipated correctly. In a later study
(Sutton, Tueting, Zubin, and John, 1967), a positive
component of the VER was demonstrated whose latency was
determined by the point in time at which ambiguity is
reduced. In addition, the shape and amplitude of this
component was influenced by whether the S was in the
presence or absence of an external event which delivered
A further significant advance in the study of
evoked response correlates of cognitive activity was
made by Walter and his associates (Walter, Cooper, Al-
dridge, McCallum, and Winter, 1964; Cohen and Walter,
1966) in their discovery of the Contingent Negative Varia-
tion (CNV) or Expectancy Wave. This phenomenon was
7Results have been by no means straightforward
in this complex area. For further reviews, see Tecce
1968), Bergaminin and Bergamasco (1967), and Callaway
1966 Also, see Eason, Aiken, White, and Lichtenstein
1964), on VER correlates of activation.
found in the course of a study of responses evoked in
non-specific areas of the brain when two stimuli were
regularly presented in association. It was found that a
prolonged, slow, surface negative wave appeared in the
evoked potential of the first stimulus if the second was
made a signal for some response by the S. Some flavor
of this work is provided by a recent abstract by Walter
The E-wave can be recorded consistently from all
normal adult subjects. The mental state in which
the E-wave develops is compounded of readiness,
motivation, attention, and expectancy and the
potential rise is terminated by recognition, deci-
sion, action and consummation. The signals needed
to initiate and terminate the E-wave can be purely
semantic, in the form of words or pictures, and
the engagement of the subject need involve only a
Averaging with suitable time-delays . Adi-
cate*g. . that a similar Readiness or Intention
wave appears a second or so before a spontaneous
voluntary decision or action, but this also need
not involve a physical movement. When the volun-
tary act is arranged to provide an experience (such
as the appearance of an interesting picture)
the Intention Wave persists through the action
until the picture disappears. The Intention Wave
. can then be made to trigger the projector
and computer directly so that the subject has the
desired.experience before any action has been
taken. Similarly, an Expectancy Wave can be made
to initiate or arrest an imperative stimulus
directly, thus by-passing the operant effector
system (p. 616).
The sophistication of Walter and his associates'
technique in the direct operation of machines through
expectancy states and their resultant COY is paral-
leled by their application of multi-channel
radioteleaetry to the recording and analysis of evoked
responses while subjects have been free to move around
and engage in activities such as reading, talking, walk-
ing, riding in a car and on a bicycle, and playing
games (Walter, Cooper, Crow, McCallum, Warren, and
Aldridge, 1966). This instrumentation has allowed ob-
servations such as the following to take place:
Thus, when a ball was tossed to a subject the
first trials showed a ONV peaking just before
the ball was caught But with practice the COV
appeared as soon as the first movement was made.
This response disappeared when the experimenter
was told to feint occasionally-so that the sub-
ject had to time his catch from the flight of
the ball (p. 617).
Recent reports have demonstrated that the CNV
can be brought under voluntary control (McAdam, Irwin,
Robert and Knott, 1966), and have linked it with other
experimental manipulations such as conditioning (Low,
Frost, Borda, and Kellaway, 1966), anxiety and stress
(Knott and Irwin, 1967) and motivational paradigms
(Irwin, Knott, McAdam, and Robert, 1966). Indeed, the
COV has even been found in rhesus monkeys (Low, Borda,
and Kellaway, 1966).8
Emotional states, furthermore, have been
reported to correlate with VER amplitude and latency
For related work, see Cohen, Offner, and Pal-
mer (1967), Hillyard and Galambos (1967), Walter (1967),
Guibal and Lairy (1967), Cant and Bickford (1967).
changes. Thus, Begleiter, Gross and Kissin (1967)
added pleasant, neutral, and unpleasant meanings to
visual stimuli through conditioning procedures of which
the subjects were unaware, and then found different VERs
to these stimuli. Lifahitz (1966), using a different
procedure. to obtain the same affective stimuli, showed
young men slides of scenic, medical, and nude figures
which were assumed to evoke the three emotional reac-
tions, and again found VER changes.
Finally, reports that the VERs of psychiatric-
ally ill individuals can be differentiated from those
of normal subjects is of especial-interest in view of
the more long-standing nature of both psychiatric syn-
dromes and intellectual characteristics, compared with
the large number of VER studies which have concerned them-
selves with transitory states of the organism.
Straumanis, Shagass, and Schwartz (1965), e.g., demon-
strated prolonged latencies in later waves of the VER
in elderly patients. Speck, Dim, and Mercer (1966)
showed that depressive patients also tend to have longer
latencies, and that the ratio of amplitude of paired
flashes to the amplitudes of single flashes, considered
to be an index of cortical recovery function, was sig-
nificantly lower in schizophrenics as compared to nor-
mals. Following improvement, furthermore, this amplitude
ratio shifted towards the non-patient levels. These
patients also showed several other VER correlates of
improvement, including a significant latency decrease
with a gain in insight (Heninger and Speck, 1966). Other
results, however, have provided conflicting evidence.
Shagass and Schwartz (1965) and Shagass, Schwartz, and
Krishnanoorti (1965) reported, among other findings (in-
cluding a longer latency of the early VER in schizo-
phrenics compared to normals), that psychiatric patients
have a significantly greater amplitude of response of
one component and a steeper intensity response curve
than non-patients, but no difference in cortical recov-
ery function between schizophrenics and normals. Rodin,
Zacharopoulos, Becket, and Frohman, (1964), on the
other hand, found that schizophrenics have decreased
amplitude of responses compared to normals. Thus, as
Begleiter, Porjesz and Gross (1967) noted in their
review, "no clear conclusions can be drawn about the
utilization of evoked potentials as a criterion for dif-
ferentiating psychological disturbance from 'normal'
functioning" (p. 758).9 This is due, perhaps, to problems
of diagnostic criteria which plague most studies involv-
ing psychiatric populations.
9More consistently encouraging have been a series
of reports of studies using auditory evoked responses
(Callaway, Jones, and Layne, 1965; Jones, Blacker, and
The VER and the EEG
This synopsis of experimental results has so far
emphasized a number of VER correlates of complex psycho-
logical states in human beings. At this point, it may
be useful to examine the relationship of the VER to the
EEG. Since the VER is drawn from these ongoing rhythms
of the brain, it is reasonable to expect that the
character of the VER may be partially determined by the
nature of the EEG. If the VER is found to be signifi-
cantly related to the EEG and the EEG is related to
intelligence, then there would be further reason to
expect a correlation between the VER and intelligence.
While some authors have failed to find such a relation-
ship (Ebe, Mikami, Aki, and Iiyazadki 1962; Ciganek,
1961), the more recent evidence has tended to support
more positive conclusions.10
Four studies provide evidence for a relationship
between EEG frequency and VER latency or number of
peaks. Kooi and Bagchi (1964), in'a study which con-
sidered the latency and amplitude of VER peaks in the
first 250 msec- found that the latency of a late nega-
tive wave correlated -.27 with alpha frequency,
10While a number of studies have documented that
the phase of alpha at which the stimulus is presented
influences the VER (Lesevre and Remond, 1967; Callaway
and Layne, 1964), this section confines itself to EEG
correlates of VER latency or number of peaks.
significant at the .02 level. Rodin, Grisell, Gudohba,
and Zachary (1965), who presented separate analyses for
males and females (on the basis of reliably greater
amplitude for females), found some statistically sig-
nificant :crrelations between the latencies of various
components and characteristics of their frequency
analysis, but since these occurred unevenly in the two
groups, they did not include the specific results. These
authors concluded that the complexity of the evoked
response curve-as reflected in the total number of
positive peaks--was significantly correlated (g's
averaging about .6 between total number of peaks and
energy =aount in the high frequency 22-33 cps band.of
the EEG) with the amount of fast activity in the basic
EEG. Finally, Barlow (1960) reported on the parallel-
ism between the frequency of resting alpha activity
and the frequency of the late rhythmic components of
the VER, a relationship which was later quantified by
Dustman and Beck (1963) by a correlation of .58 between
average alpha and late component frequency. It seems
likely% therefore, that a relationship exists between
latency and EEG frequency, one which may be negligent
in the early components of the VER and becomes increas-
ingly significant in the later components. This find-
ing is especially interesting in light of Chalke and
Ertl*s (1965) finding that these same later components
were the ones which were found to vary with intelligence.
It provides some justification for the assumption that
variables which have been found related to EEG fre-
quency might bear a similar relationship to VER latency.
Prime among these relationships are reported correla-
tions between intelligence and EEG frequency.
EEG Frequency and Intelligence
The extensive literature concerning the rela-
tionship between EEG variables and intelligence is beset
by conflicting and unreplicated findings.12 Most
11This assumption has already been borne out in
several cases. For example, EEG frequency and reaction
time have been found to be inversely related (Surwillo,
1961, 1963, 1964a,b; Williams, Granda, Jones, Lubin,
and Armington, 1962), as have reaction time and the
later components of the VER (Dustman and Beck, 1965a;
Donchin and Lindsley, 1966; Nawratzki, Auerbach, and
Rowe, 1966). Similarly, the alpha frequency of the EEG
slows in old age (Otomo, 1966), as do thei.Lencies of
most peaks of the VER (Ebe, Mikami, Aki, and Hiyazaki,
1962; Straumanis, Shagass, ahd Schwartz, 1965).
12Vogel and Broverman's (1964) recent review con-
cluded that the evidence primarily supported significant
positive relationships between EEG frequency and intelli-
gence (usually conceived as mental age), mostly in
feeble-minded subjects, children, institutionalized geri-
atric subjects, and brain-injured adults, and much less
so in normal adults. A similar view was advanced by
Liberson (1967), whose review concluded that most inves-
tigators had found a positive relationship between alpha
frequency and mental age, provided that chronological
age of the sample was held within narrow limits. This
view has been challenged by Ellingson (1966) who criti-
cized Vogel and Broverman's (1964) conclusion and con-
curred with former reviewers in the belief that signif-
icant relationships had yet to be demonstrated. Vogel
and Broverman (1966) had the final word, reiterating
and defending their original conclusions, but these
authors later provided additionally confusing findings
(Vogel, Broverman, and Klaiber, 1968; see text).
pertinent to the present study are the six experiments
which have examined the relationship between EEG fre-
quency and intelligence in normal adults.
Biesheuvel and Pitt (1955) found no relation-
ship between the Raven's Progressive Matrices and alpha
frequency. Gastaut (1960), using French translations
of several of Thurstone's Primary Mental Ability Tests,
reported negative results with French army recruits.
Shagass (1946) found no significant relationships using
the Royal Canadian Air Force Classification Tost.
Sugarman (1961), employing the South African Group
Intelligence Test, obtained a significant negative cor-
relation between alpha frequency and the total test
score in 35 university students, but not in a group of
15 staffrmembers. On the other hand, Mundy-Castle (1958)
and Mundy-Castle and Nelson (1960) found significant
positive relationships (y .51, r .34, .01 level of
significance) between WAIS Full Scale IQ and a number of
subtest scales (South African version) and alpha fre-
quency.13 An explanation for these widely disparate
results was proposed by aundy-Castle, who noted that he
alone used the WAIS to measure intelligence. He sug-
gested that the correlations between the WAIS and other
measures of normal adult intelligence indicated
11n addition, Saunders (1961) reported a sig-'
nificant positive relationship between memory' for digits
and EEG frequency.
considerable variance unique to each type of intelligence
test, and,therefore, that alpha frequency may be corre-
lated with the Wechsler, and the Wechsler significantly
correlated with the other intelligence tests without
alpha frequency being correlated with tests of intelli-
gence other than the Wechler. In contrast, he sug-
-gested that the tests used by experimenters who found
negative results were not comprehensive tests of intel-
ligence or were tests whose factorial structure or
relationship to other tests had not been investigated.
Finally, Vogel, Broverman, and Klaiber (1968) recorded
EEGs during rest and during periods of mental effort
in two samples of normal young adult males and inves-
tigated the relationships between a number of EEG
measures, including frequency, and three separate in-
dices of mental ability. They found that certain EEG
measures were related to "Automatization Cognitive
Style," defined as greater ability (strong automatiza-
tion) to perform simple repetitive tasks than expected
from the individual's general level of mental ability.
It may be more significant, however, that they found
general intelligence and EEG frequency completely unre-
lated. Moreover, slow waves and slow alpha frequencies
were positively associated both with automatization
ability and with efficient cognitive performance under
conditions of mental effort. Vogel,et al. (1968)
discussed their results, including the discrepancy
between their own study and the conclusions they drew
from their review of the literature, in terms of several
kinds of inhibition which they suggested are involved
in intellectual activity. Unfortunately, both the
"cognitive style" upon which they focused, and the con-
structs which they used to explain their results, are
esoteric and complex. Their findings, as well as the
others which have been previously noted, warrant atten-
tion. However, lack of independent replication and
clear theoretical constructs leave the relationship
between EEG frequency and intelligence suggestive but
If it is appropriate to partially equate
latency of the VER and frequency of the EEG, then it
may also be profitable to briefly compare theoretical
speculations about these variables.
Surwillo (1964a,b), citing a number of studies
which have shown that higher EEG frequencies are a con-
comitant of the more complex mental processes, and his
own studies relating reaction time to EEG frequency,
suggested that, in terms of information theory, the time
required to process one bit (the informational capacity
of the central nervous system) is a function of EEG
frequency. Based on this point of view, he presented
data which predicted information capacity on the basis
of EEG frequency which corresponded with behavioral
studies of information handling capacity. This theo-
retical viewpoint relates EEG frequency to the temporal
aspects of behavior, suggesting that the brain wave
cycle is the unit of time in terms of which simple
behavior is organized by the CNS.14
Another approach to information processing was
taken by Mundy-Castle (1958) and Mundy-Castle and
Nelson (1960), who accounted for their findings of a
relationship between EEG. frequency and intelligence in
terms of a cognitive style which they termed "primary-
secondary function." Mundy-Castle (1958) has suggested
that primary function
refers to the initial activity of a sensori-
motor process during stimulation, while
secondary function characterizes the tendency
of the process to continue for varying periods
of time after cessation of the stimulus.
Secondary function is believed to influence all
subsequent associations, tending to limit them
to the themm' of the primary function. The
general effect of SF is to give continuity and
integration to mental events, since it favors
persistence of attention and rate of work,
relative stability of moods and interests, and
action in the light of past experience (p. 185).
The authors pointed to a number of studies
which have indicated that SF varies over a continuous
See Harter (1967) for a critical review of
scale and is relatively enduring and constant for any
individual. At one end of the scale (the so-called
"primary" pole), where SF is relatively lacking, the
personality is characterized by an extensive but shallow
conscious field, dominated by primary experience. At
the other extreme, where primary conscious content is
dominated largely by SF, the conscious field is said to
be given greater depth due to the resultant reduced
stimulability and more frequent evocation of past
It was noted previously that the late waves of
the VER are; the ones which tend to correlate with EEG
frequency, and were also found to vary with intelligence.
It is striking, therefore, that Chalke and Ertl (1965),
in speculating on these findings, took an approach which
was quite similar to the theorists who considered EEG
frequency, and suggested that VER latency might reflect
the speed of information processing in the nervous sys-
tem. Speck, Dim, and Mercer (1966) reasoned similarly
in discussing the prolonged latencies of depressed
patients, suggesting that in terms of a cybernetic model,
the information processing channels may be slowed. A
number of other authors, taking note of additional evi-
dence on the late VER waves, such as their presence all
over the scalp, their sensitivity to conditioning, bar-
bituates, sleep, and symbolic value, have also suggested
that they are more reflective of cognitive processes
when compared to the early more purely specific sensory
waves (cf. Callaway, 1966). Kooi and Bagchi (1964b)
have noted the independence of one of the late waves of
the VER from the actions of the earlier waves and sug-
gested that it has a functionally different nature.
Others (Dustman and Beck, 1965a; Donchin and Lindsley,
1966) have suggested that the later waves (appearing
after latencies of 50-80 msec.) may be related to dis-
charges from the reticular system and the diffuse
thalamic projections, suggesting that this added neural
activity is necessary for conscious awareness of the
direct input reflected by the earlier waves of more
Summary and Critique
Recently developed computer techniques have al-
lowed the recording of VERa from the human scalp, result-
ing in the description of its basic parameters. The VER
seems reliably to reflect certain.other complex mental
processes besides intelligence. By analogy, therefore,
it would not be surprising to find that intellectual
processes as well were reflected by VER measures. In
addition, the latency of the VER has been related to the
frequency of the EEG. EEG frequency has, in turn, some-
times been related to intelligence. To the degree that
such relationships held up, they would also lead to the
expectation that VER latency would reflect intelligence.
The force of these arguments, and the empirical
data which support them, is lessened, unfortunately,
by the shortcomings of a research area whose methodology
is still exploratory, and which lacks a rigorous theo-
Difficulties arise specifically in VER research,
for example, from the necessity of recording many
responses from the scalp out of EEG activity in order to
arrive at a coherent representation of the cortical
response. First, the conventional electrode arrangements
placed on the scalp may not provide a "true" picture of
the underlying evoked response activity. The scalp May
sometimes act as an average of electrical activity from
underlying cortical areas, averaging out the local
random rhythms and transmitting the rhythms which are
common to and synchronous over relatively large areas
(De Lucchi, Garoutte, and Aird, 1962). For example,
the electrical field structure of the scalp is such that
large evoked potentials are recorded from the vertex.
However, this placement doesn't lie above any particu-
larly active cortical zone, but rather over the sagit-
tal sinus, a pool of venous blood as remote from
active tissue as any part of the head (Walter, 1964).
Similarly, it is possible to record considerable EEG
activity from the scalp which overlies an absent hemi-
sphere (Cobb and Sears, 1956). On the other hand, the
latency and waveform of responses recorded on the scalp:
have also been shown to be similar to those recorded from
the brains of experimental animals (Katzman, 1964),
and the distribution of responses on the scalp has been
found to correspond largely in several patients to
potentials generated by the underlying visual cortex
(Vaughan, 1966; Corletto, Gentilomo, Rosadini, Rossi,
and Zatoni, 1967; Rayport, Vaughan, and Rosengart,
1964). However, conflicting findings have been pro-
vided by Heath and Galbraith (1966) and Cooper, Winter,
Crow, and Walter (1965), and these findings, like many
others, may vary with stimulus conditions such as light
Second, the computer averaging technique itself,
in which responses which are timelocked to the stimulus
are summed while random activity averages out, may yield
evoked potentials whose components may be somewhat arti-
factual. The amplitude of individual components con-
stituting the average VER may reflect mostly the stabil-
ity of that component. Small components which are
stable in individual potentials may appear grossly
exaggerated in computed averages, and the less stable
larger components may not be so prominent (Abrahams,
Finally, the studies in both the VER and EEG
literature are weakened because many of them are not
comparable to each other in variables, methods, equip-
ment, or analyses and interpretation of data. Lifshitz
(1966) and Begleiter, Gross, and Kissin (1967), for
example, while ostensibly comparing the same emotional
states, analyzed their data so differently from each
other that it is almost impossible to compare their
results. Another difficulty is that most of these find-
ings remain unreplicated, and others are openly contra-
dictory of one another. Thus, the performance of
mental arithmetic, when used in an attention experiment
to operationalize distraction from a visual stimulus,
has been found to reduce the amplitude of the VER
(Garcia-Austt, 1963). The same operation, however,
when used to define activation, seems to increase the
VER amplitude (Eason, Aiken, White, and Lichtenstein,
1964; of. Callaway, 1966). Part of this problem can
probably be explained by the fact that some of the rele-
vant parameters of VERs are insufficiently explored;
while others, such as pupillary diameter (Bergamini,
Bergamasco, Mombelli, and Gandiglio, 1965), which havO
been shown to influence the VER under some conditions,
are often left uncontrolled.
The Present Study
Within the context of these shortcomings of
the VER literature in general, this study was designed
to replicate and extend Chalke and Ertl's (1965) original
experiment, and to deal more rigorously with the follow-
1. Are the latencies of certain peaks of the
VER (and by implication, the general waveform) related
to intellectual behavior?
2. Might such a relationship vary with the
area of the brain from which the VERs are recorded?
3. Might such a relationship also vary with
different kinds of intellectual performance? If so,
is it dependent on the higher order cognitive processes
of abstraction, manipulation.of symbols, and integra-
tion of present and past experience that underlie most
thinking and problem-solving? On the other hand, might
it be due to a lower-order concomitant of most of the
intellectual activities which are measured by standard
tests, such as speed of perceptual information proces-
sing? In short, what underlying factor or factors could
account for any-correlation that might be found between
the latency of the VER and an individual's performance
on an IQ test?
4. Do evoked response correlates of intelli-
gence become more accentuated and apparent while sub-
jects are engaged in a simple cognitive task rather than
while mentally "at rest?"
Forty-six right-handed males served as Ss in
this study. Of these 46 Ss, at least 35 were used in
all of the analyses; 37 _s were used in most of them;
and 39 were used in several more. These 7-11 Ss were
deleted, depending on the specific analysis, because of
insufficient signal/noise ratios (4 Ss), data loss
(2 s), equipment malfunctioning (1 S), or incomplete
data collection (4 Ss). The age range of the 37 Ss
used in most of the analyses was from 16-25 years, with
a mean of 18 years, and a standard deviation of 2.4
The Sa were drawn from MENSA-a local society
whose membership is based on an IQ above 150 (4 Ss);
introductory psychology classes (13 Ss), PK Yonge High
School in Gainesville (16 Ss); and local high school
dropouts who answered an ad in the local newspaper (13
gs). The Ss were chosen to emphasize the normal and
higher range of intelligence, thus avoiding the neces-
sity of basing conclusions largely on retarded $s, many
of whom might be brain-damaged, and maximizing the appli-
cability of the findings to the largest number of people.
This sample differed from that of Chalke and Ertl (1965)
mainly in its continuous distribution of intellectual
abilities over a more limited range. as who had any
history of visual or neural trauma were not accepted.
Recording of Evoked Responses
Evoked responses were recorded with three pairs
of cup electrodes held in place by a partially elastic
headband; the earlobe was grounded through a clip-on
electrode; and skin contact was made through EEG elec-
The Ss were seated in a deep chair and faced
a Grass PS-2 photostimulator mounted in a wooden hemi-
sphere and set at level 8 intensity. They sat 24 inches
from the light with their eyes open, and looked down at
a point on the floor in front of them. The only other
light was provided by a 60-watt bulb placed above and
behind the Ss. These stimulus conditions paralleled
those of Chalke and Ertl (1965).
The signals from the three sets of electrodes
were led to three Grass P5 amplifiers with frequency
responses of .15 to 50 cps, amplified 100,000 times,
and then fed to a Mnemotron Computer of Average
Transients (CAT 4000) for summation and display. The
CAT displayed a waveform of the VER (analog data) which
was recorded by a Sanborn ink writer, and also provided
100-number descriptions of each response (digital data)
which was printed out by a TWC digital print-out unit.
This averaging device differed from Chalke and Ertl's
(1965) sero-orossing technique, but it had been claimed
that the two approaches yield almost identical results
(Erti, 1966). Similarly to Chalke and Ertl (1965)
however, the light was flashed randomly at an average
frequency of one flash every two seconds; 120 flashes
constituted one complete evoked potential trial; and
the computer analysed bioelectric activity from the scalp
for 500 msec. after the flashes. The amplifiers for
each channel were calibrated at weekly intervals by-
delivering a 10 uv. signal to each of them and then
summing 100 samples on,the CAT.
Tests of Intellectual Function
Simple and disjunctive reaction time.-These
tests constituted the most basic measure of perceptual-
motor speed, with the smallest cognitive component.
They required the S to depress a key when a light came
on in front of him (simple RT) or to depress the key
only if one of two lights came on (disjunctive RT). RTs
were measured with a Lafayette RT apparatus attached to
a timer set to begin automatically when a light in
front of the S came on. When the S pressed the key to
turn it off, this timer stopped and gave the RT in
msec. A ready signal was given randomly from 1-3
seconds before the stimulus, and each S was given 3
warm-up trials followed by 10 trials of SRT and 10 trials
of DRT. There were two final scores, one consisting of
the mean of the 10 trials of SRT, the other of the mean
of the 10 trials of DRT.
Minnesota Clerical Test.-The MCT was used
to measure the speed of more complex perceptual informa-
tion processing. It consisted of two separately timed
subtests, Number Comparisons and Name Comparisons. The
first required the S to compare 200 pairs of numbers
of 3 to 12 digits each. If the two numbers in the pair
were identical, he was to place a check mark between
them. The task was similar in the second subtest,
with proper names substituted for numbers. A time-
limit of eight minutes was allowed for numbers, and
seven minutes for names. The scores were arrived at by
summing the number of correctly checked pairs of items,
and subtracting from this total the number of errors.
Paired Associates Learning.-The PAL task con-
stituted an additional measure of associative memory.
It consisted of presenting the S with 10 pairs of
simple words, and then requiring him to give the second
word of each pair when presented with the first. This
was repeated three times, and the score was the total
number of correct words memorized.
Otis Group Intelligence Scale.--The Otis added
an additional general intelligence measure which
loaded most highly on verbal comprehension. It was in-
cluded to replicate Chalke and Erts (1965) study, and
the 20-minute version was used.
Wechsler Adult Intellipence Scale, Satz Short-
Form.-The WAIS, a comprehensive scale consisting of sub-
tests covering the factors of verbal comprehension,
numerical reasoning, spatial organization, associative
memory, and perceptual speed, served as the reference
measure of intelligence in this study. It was chosen
because it is the beat known and most frequently used of
the individually administered intelligence tests. Its
division into verbal and performance scores also ap-
peared to have some neural correlate in that verbal
scores have been found to decrease quite consistently
with left-hemisphere lesions, independent of language
disorders. Right-sided lesions, however, have been
found to lead to decreased performance scores only in-
consistently (Satz, 1967; Reitan, 1955; Milner, 1962;
On arrival at the laboratory, s were first
tested for visual auity with a Snellen Eye Chart and
questioned to detect the possibility of neural or visual
defects. They were then given the various intellectual
tests in the following order: Otis, Minnesota Clerical
Test, Reaction Time, Paired Associate Learning, and
Wechsler Adult Intelligence Scale. The Ss' pupils were
then dilated with a myadratic agent (10% Neosenephrine)
in order to control for differences across Ss in pupil
size, since it was know that pupillary diameter helps
to modulate light intensity and also varies with cog-
nitive processes, and that the latency of the VER
varies with light intensity. Following dilation, the
Es were fitted with electrodes over the left-motor,
right-motor, and occipital areas for evoked response-
recording. The left- and right-motor electrodes were
placed 6 ca on either side and parallel to the midline,
and 3 cm. on either side of a line joining the two ears.
The occipital electrodes were placed on the midline,
with the lower electrode on the inion and the upper
electrode 6 cm. higher. The left- and right-motor elec-
trodes were also located 6 cm. apart from each other.
These brain locations were chosen for the following
Left-motor area.--Chalke and Ertl (1965) re-
corded VERs whose latency correlated with intelligence
from this region of the brain. They chose this location
to maximize input-output delay and thus to enhance the
possibility of actually measuring central processing
time. It might be additionally expected that verbally
loaded tests might show higher correlations with VERs
recorded from this hemisphere than with those recorded
from the right hemisphere due to the probable mediation
of verbal intelligence by this hemisphere in most Ss.
Right-motor area.-This area was chosen to pro-
vide a symmetrical comparison of hemispheric brain func-
tion. In view of the evidence which has often located
visual-constructual and spatial abilities in this hemi-
sphere, possible correlations with tests measuring
these abilities were expected.
Occipital area.-This location includes both
primary and secondary visual receiving areas, and there
is evidence that these areas are not essentially in-
volved in most cognitive tasks (Piercy, 1964). Cor-
relationsbetween VERs recorded from this area and intel-
ligence, therefore, were not anticipated.
Following the placement of electrodes and the
positioning of the s, three VER trials were adminis-
tered as follows:
(1) A non-counting evoked response trial
consisting of 120 flashes, presented in two 60-flash
sequences to allow the % to rest briefly.
(2) A control trial of 120 flashes in which the
light was occluded (similarly broken into two 60-flash
sequences, but with a shorter interval between them).
(3) A counting evoked response of 120 flashes,
also broken into two 60-flash sequences. During this
trial, the gs were instructed to press a button on the
first flash, then press the button after two more flashes,
then after three more flashes, etc. The Ss' responses on
this task were monitored by the E through use of a light
which flashed each time the button was pressed.
Trials 1 and 3, the non-counting and counting
conditions, were counterbalanced across all Ss, with mean
intelligence held equal in the two groups, to control
for order effects.
Data Processing and Analysis
Upon completion of the experiment, each S's
scores on the various intellectual measures, and his
digital and analog VER data, were available for analysis.
These data were processed in two stages: a preliminary
one in which general information about this sample's
intellectual and VER characteristics were assessed; and
the main analysis which was devoted to exploration of the
relationship between latency (and other VER measures)
and intelligence.1 Since this study was designed to
both replicate and extend Chalke and Ertl's (1965)
original study, several different statistical approaches
were used to answer various questions. These approaches
included the use of the Pearson Correlation Coefficient
(r) for analysis of intercorrelations of intelligence
test scores and VER digital waveform data, and non-
parametric tests (Mann-Whitney and Kendall Tau) for
analysis of the VER-latency data, in parallel with
Chalke and Ertl (1965).
Preliminary Analyses: Psranetric Data
Each S's intellectual test scores were pro-
cessed through an IBM 360 computer for intercorrela-
tional analysis. The resulting correlation matrix pro-
vided a measure of linear relationship between the vari-
ous intellectual tests, and gave some idea of how many
relatively discrete or independent intellectual abili-
ties were tapped by these tests. Means, ranges, and
standard deviations were also computed for each test.
VER digital data
Digital data, comprised of two VERs per S (the
counting and non-counting trial),wre: analyzed by the
The various statistical comparisons which were
made in this study are summarized.in Appendix 1.
IBM 360 in the following three ways:
Reliability within Ss.--Correlations between
the first and second evoked response trials for each
P's three locations yielded rough measures of reliability.
These measures provided evidence for the stability or
lack of stability of the VERs over time, and informa-
tion about the importance of the two conditions.
Brain conmunality within Sa.-Correlations among
VERs collected from each of the three locations from
each S yielded a rough measure of the relationships
between the different parts of the brain.
VER Piiilarity across Ss for location. -Corre-
lations between each S's VER and every other S's VER
resulted in a measure of similarity between locations
across So. This measure helped to determine whether
the VERs recorded from each location of the brain were
similar from S to S.
Peak-to-peak amplitudes.--Peak-to-peak ampli-
tudes were also computed for each S's VERa, and con-
parisons were made between conditions for each location,
and among locations with data pooled for conditions.
Main Analysis: VERs and Intelligence
The relationships between latency of the VERs
and intelligence were determined in the following ways:
Assessment of peaks.--In order to assess the
relationship between intelligence and the latency of the
various peaks of the VERs for each condition, objective
criteria were developed to determine whether deflections
in the VER should be considered peaks. Criterion I was
designed to take into account only the obviously larger
peaks in the VERs. Five major peaks were fairly readily
identifiable in most So, and this criterion (the "High
Five") was designed to include them and exclude the
smaller deflections, or those which came quite close to-
gether. Criterion II (the "First Five") was designed to
take into account the smaller deflections which occurred
in the VER, and also to treat those peaks which fell
close together.2 The latencies of the peaks which
emerged from use of these criteria were then determined
for the VERs from each location and condition. This was
done through reference to the digital data, which pro-
vided a voltage reading eery 5 aseo.
IQ group analysis.-After the determination of
latencies, the Ss were divided into three groups based
on high, medium, and low FS WAIS intelligence scores.
2The specific rules which comprised the criteria
and which governed the selection of peaks are found in
Appendix 2. In cases in which VERs did not contain five
peaks which met the criteria, those peaks which were
valid were included and blanks left for the remaining
The mean latencies for each peak of the VERs of the $s
in each of the three groups were then compared through
use of Mann-Whitney U Tests (Hays, 1963) for each loca-
tion. This analysis paralleled that of Chalke and Ertl
(1965), with the exception that their low IQ group was
close to matching this study's medium IQ group, and
this study's medium IQ group fell into the range between
their high and medium group.
IQ correlational analysis.-In order to obtain
further information as to whether latency was an accu-
rate predictor of intelligence in individual cases,
Kendall Taus (Hays, 1963) were computed for a number of
the peaks derived from the left-motor VERs for both the
counting and non-counting conditions.
Differential abilities analysis.-In order to
explore the aspects of intellectual function that ap-
peared to be most responsible for correlations between
latency and IQ, the latency of one of the LM peaks was
correlated with a number of the remaining intellectual
measures. These results were checked through use of
random correlations between other peaks and tests.
Additional VER measures
The relationship of peak-to-peak amplitude,
reliability, and communality among brain locations, to
intelligence was examined through comparison of the
means of the high and low IQ groups (Mann-Whitney U
Tests) on these various VER measures.
Correlational waveform analyses
The purpose of this third analysis was to test
the hypothesis that Ss with similar intelligence would
have similar VER waveshapes. A measure of VER similar-
ity was provided by correlating each S's VER with those
of every other S, location by location. A measure of
intellectual similarity was generated by subtracting each
S's WAIS full scale score from that of every other S. In
each case, these operations formed matrices, in which
each cell in the VER matrix corresponded to a cell in
the IQ matrix, comparing the degree of intellectual and
electrophysiological similarity between pairs of Ss.
To gain a preliminary view of this data, the
VER correlations and IQ difference scores were plotted
against each other for each location. The data were
further examined by selecting 18 Ss with the most stable
VERs from across the IQ range (6 Ss from each of three
groups based on IQ scores of 70-95, 96-120, and 121-
145, with no ties). Each of these VER correlations of
all pairs of Ss from these groups were then entered
into a matrix which was re-ordered on the basis of the
difference scores of these pairs of Ss. This resulted
in a matrix in which the Ss who were intellectually
similar were at one end and those who were dissimilar
were at the other end. If the hypothesis that waveform
similarity would correspond to intellectual similarity
was correct, then each column in this matrix would be
in ascending order of VER correlation magnitude. Ken-
dall Taus (Hays, 1963) were computed for each column of
the matrices which were generated for each location.
This procedure was repeated for the 10 Ssacross the
IQ range whose VERs were least stable.
Preliminary Analyses: Parametric Data
The results of five of the major intellectual
and perceptual-motor tests used in this study are shown
in Table 2. This table illustrates the means, ranges,
and standard deviations of the Full Scale of the
Wechnler Adult Intelligence Scale (WAIS FS), Otis Scale,
Paired-Associates Learning (PAL), Simple Reaction Time
(SRT), and the Numbers section of the Minnesota Cleri-
cal Test (MOT Nu). Inspection of this table indi-
cates that the Ss used in this study had a wide range
of abilities, and that the sample centered about six
to eight points above average in intelligence.
Intcrcorrelations among the test scores sug-
gested several conclusions about the nature of the
abilities which were measured.1 First, there were strong
relationships among all of the intelligence measures,
1This table of intercorrelations is found in
MEANS, RANGES, AND STANDARD DEVIATIONS OF MAJOR INTEL-
LECTUAL AND PERCEPTUAL-MOTOR TESTS FOR 37 SUBJECTS
USED IN MOST ANALYSES
WAIS FS Otis PAL SRT HCT-Nu
Mean 108.6 105.8 13.7 .196 91.0
Range 70-137 60-128 3-23 .26-.14 26-147
S.D. 17.5 19.2 8.9 .032 28.6
suggesting that intelligence was validly measured.
Second, inspection of the intercorrelations of the intel-
lectual measures revealed that tests which measured the
same abilities (e.g. object assembly and block design
of the WAIS as measures of spatial analytic ability,
digit span and paired-associates learning as measures of
recent memory) correlated only slightly higher with
each other than with the remaining intellectual measures.
These intercorrelations, moreover, paralleled those of
Wechsler (1958). This suggested that most of the vari-
ance due to intellectual factors could be accounted for
by an overriding factor of general intelligence, with
relatively small additional contributions of specific
group factors. Third, the correlations of all remaining
test scores with the full scale scores of the WAIS, the
reference measure of intelligence, provided a measure of
the degree to which each of these remaining tests was
related to general intelligence. Inspection of these
correlations revealed that the WAIS verbal tests and the
Otis Scale correlated most highly (mean of .83),
followed by the VAIS performance tests and the combined
HOT scores (mean r of .68). The paired-associate
learning task followed this group (mean r of .47).
Simple and disjunctive RTs,however, were almost com-
pletely unrelated to the intellectual measures (mean r
of -.14), and correlated only about -.30 with
perceptual speed tasks such as the NOT and the digit
symbol test of the WAIS. This indicated that simple
motor speed was independent of cognitive abilities, and
emphasized that even moderately complex perceptual
speed performance is highly related to intelligence.
VER Digital Data
Reliability within Ss.-Mean reliabilities for
the three locations (correlations between the first and
second evoked response trials) were .83 (occipital),
.73 (right-motor), and .69 (left-motor). These strong
relationships demonstrated that the VERs were stable
and reproducible for most sa. In addition, doubt was
cast on any effect of the counting vs. non-counting
conditions on the VERs.
Brain communality within Ss.-The mean corre-
lations among VERs collected from each of the three
locations for each S (each of two VER trials combined
for each S) were .02 (occipital-left motor), .06
(occipital-right motor), and .77 (left-motor right-
motor). These figures indicate that the occipital
response was independent of the two motor responses,
which were, in turn, highly similar to each other.
Part of this similarity may be accounted for by the fact
that the motor electrodes, although placed on different
sides of the brain, were located much more closely to
each other than they were to the occipital electrodes.
VER similarity across Ss for location.-The
left- and right-motor areas yielded symmetrical dis-
tributions with mean r's of .00 (LM area, combined
counting and non-counting trials) and .01 (right-motor
area, combined counting and non-counting trials). The
occipital VERs, on the other hand, centered around a
mean a of .52, and were skewed to the left. These
figures suggest that insofar as waveshape may be said
to characterize the VER, there is a general occipital
VER whose waveform is roughly the same from subject to
subject for any given condition. This is not the case,
however, for the left- and right-motor VERs. One fac-
tor which may help to explain this discrepancy between
the motor and occipital VERs is the possibility that
the occipital electrodes were placed on more similar
areas from subject to subject, since the inion pro-
vided an easily identifiable point from which to start.
The occipital electrodes were always placed, therefore,
in relationship to a specific point on the scalp. The
motor electrodes, in contrast, were always placed in
reference to each other, the aidline, and the ears, and
thus may have varied more with respect to the cortical
area over which they were placed due to difference in
size and shape of the head. In addition, most Ss had
a good deal more hair on the scalp covering the motor
areas than over the occipital area. This necessitated
the placement of a greater quantity of electrode paste
in order to make proper electrical contact with the
scalp, and thus may have led to the recording of
cortical activity from a larger area of the brain.
Peak-to-peak amplitudes.-Computation of peak-
to-peak amplitude for all waveforms yielded the follow-
ing general results. The mean amplitude of the occipi-
tal response was 13 uv., the mean amplitude of the
left-motor response was 7 uv., and the right-motor
response had a mean amplitude of 8 uv. Habituation
across trials 1 and 2, spanning an average time period
of about 10 minutes, was negligible for all locations.
Comparison of the counting and non-counting conditions
revealed that the counting condition resulted in evoked
responses with larger amplitudes for all locations.
About two out of every three Ss showed these larger
counting responses, a significant difference when
only this fact was taken into account (occipital and
right-motor locations, p < .05 level; left-motor
location, p < .09 level; one-tailed Sign Test for
Matched Pairs, Hays, 1963). However, when the magni-
tude of the differences was also considered, then only
the occipital area even approached significance (p < .1
level, one-tailed, Wilcoxen Test for Two Matched Samples,
Hays, 1963). Such an amplitude difference in favor
of the counting VERs was expected due to either an
increase in attention or arousal during this condi-
Finally, the right-motor responses proved
to be consistently larger in amplitude than the left-
motor response (pooled conditions). Twenty-six
of the 37 Ss analysed showed this right-motor superior-
ity, and the magnitude of the differences was also
sufficient to make this difference significant at
the .02 level, two-tailed (Wilcoxen Test for Two
matched Samples, Hays,. 1963).
A number of these amplitude and waveform char-
acteristics are found in Figure 1, a presentation of
MEDIUM IQ S
500 msec. 500 msec. 500 msec.
Fig. 1. Two consecutive visual evoked responses, re-
corded at the occipital, left-motor, and right-motor lo-
cations, from a high, medium, and low IQ subject.
HIGH IQ S
LOW IQ S
two VER trials for each location for each of three Ss,
chosen from the high, medium, and low IQ groups.2
Study of this figure reveals the stability of the wave-
forms within each S across time, the stability of the
occipital waveform across all Ss as well (in contrast
to the left- and right-motor waves), and the increasing
amplitude size as one moves from left-motor to right-
motor to occipital locations.
Main Analyses: VERs and Intelligence
IQ group analysis.--The means of the three
groups divided on the basis of Full Scale score of the
Wechsler Adult Intelligence Scale were 127 (high,
S=- 13), 110 (medium, N 12), and 90 (low, N 13).
When the mean latencies for the three groups were com-
pared using Mann-Whitney U Tests (after Chalke and
Ertl, 1965), only the left-motor area VER latencies
proved to bear a significantly inverse relationship to
intelligence. The mean latencies of each left-motor
area peak for the three groups of Ss (Criteria I and
II) are found in Figure 2. This figure presents three
sets of data for each peak: the counting and non-
counting conditions, and the first trial (Tl) given
, 2In order to conserve space, the waves shown in
Figure 1 are the results of 60 light flashes rather
than the 120 used in the statistical analyses.
200 3 00
Fig. 2 Mean latency of each left-motor VER peak
for the high, medium, and low IQ groups. Two cri-
teria and three sets of data (counting and non-
counting conditions, and Trial 1 regardless of con-
dition) are presented, o- o = Trial 1,
S.............a Counting, A----A = Non-Counting.
to each S with the conditions randomly distributed
over this trial. Trial 1 was included in the analysis
on the chance that the conditions might prove to be
irrelevant to the relationship between intelligence
and latency, and might even obscure a relationship
best seen by taking the first evoked potential sample
Inspection of this figure indicates that there
was a generally negative relationship between intelli-
gence and latency as hypothesized. This relationship
seem to hold roughly over both conditions, although
it was less apparent for the counting condition. As
suggested, Tl seemed to result in the most linear rela-
tionship across the two criteria. As might have been
expected, when the data from the counting and non-
counting conditionswere pooled, the resulting latencies
paralleled those of the first VER trial. Similarly,
the most significant differences between the high end
low IQ groups appeared in the Tl presentation, followed
by the non-counting, and then by the counting condition.
The medium group, however, generally fell into the same
range as the high or low groups and did not line up
linearly for any of the sets of data.
Consequently, further analyses were confined to
the T1 presentation. Furthermore, this treatment
paralleled Chalke and Ertl's (1965) data analyses which
also used a single trial, and so facilitated comparison
with their work.
The mean latencies of the peaks of the Tl pre-
sentation which were significantly different among IQ
groups are illustrated in Table 3. Inspection of this
HEAN LATENCIES OF THE LEFT-MOTOR VER PEAKS OF THE TRIAL
1 PRESENTATION WHICH WERE SIGNIFICANTLY
DIFFERENT AMONG IQ GROUPS
IQ Group Peak Latency Probability
Hi vs. Lo 2 131-170 .01
Hi vs. Lo 3 228-278 .05
Hi vs. Lo 5 416-448 .025
Med vs. Lo 3 224-278 .025
eMd vs. Lo 5 422-448 .036
Hi vs. Lo 1 40-60 .05
Hi vs. Lo 3 176-232 .025
table shows that Criterion I yielded more significant
results, but both criteria led to the same basic find-
ing. This convergence in findings adds further validity
to this IQ-latency relationship, and suggests that it
can be demonstrated regardless of the criteria which are
set up to determine the peaks. Study of the actual
waveshapes of the Ss revealed that many of the low IQ
Os had noticeably simplified configurations and fewer
peaks, especially in the first 200 sec. of the response,
although this finding was not invariant. These wave-
shapes and the peaks which were identified by the two
criteria are illustrated in Figure 3, the left-motor
responses of the four most intelligent (X 154) and the
four least intelligent (X 77) So.
As was noted previously, only VERs recorded from
the left-motor area were foiun to be related to intelli-
gence. Study of the waveshapes of the right-motor area of
a number of high and medium IQ Ss revealed that some ef
these Ss had complex left-otor responses, but also had
right-motor responses which were somewhat simplified and
similar in seertai respects to the left-motor responses
of the lower IQ sa. Figure 4 illustrates the left- and
right-motor VERs of three ss who showed this feature.
This simplified wave, while it does eluidate the dif-
ference between the two locations, was noted only in some
cases. The degree of relationship between left-motor
and right-motor responses did not itself seem to be a
function of intelligence.
IQ correlational analysis.--In order to obtain
further information as to whether latency was an
accurate predictor of intelligence in individual cases,
Kendall Taus (Hays, 1963) were computed for the
HIGH IQ Ss
LOW IQ Ss
A Criterion I
e Criterion II
Fig. 3. Left-motor responses of the four most
intelligent and the four least intelligent sub-
jects. The peaks chosen through use of Criteria
I and II are presented.
Fig. 4. Complex left-motor responses of three high
or medium IQ subjects contrasted with their simpler
I = 2,mv
latencies of all T1 left-motor peaks (Criteria I and II)
against Full Scale WAIS IQ. These results are found in
Table 4. They indicate a relatively constant and
CORRELATIONS BETWEEN THE LATENCY OF ALL TRIAL 1
PEAKS (CRITERIA I AND II) AND FULL SCALE WAIS
1 2 3 4 5
Tau -.21* -.29** -.28** -.28** -.28*
N 37 37 37 35 26
Tau -.19 -19.9* -.26** -.23* -.21*
N 37 37 37 37 35
moderately negative correlation between the latency of
the peaks and intelligence. While this correlation was
significant in many cases, study of the scatter plot of
the latencies of peaks 2 and 3 (Trial 1, Criterion I)
against IQ, as seen in Figure 5, emphasizes the great
variability of the score and the resulting limits of
the predictive value of latency for individual cases.
95 0 0
S95 0 * *
70 125 180 235 290
135 0 *
140 215 290 365 440
Fig. 5. Scatter plots showing the relationship between
Full Scale scores of the Wechsler Adult Intelligence
Scale, and the latencies of peaks 2 and 5 (Trial 1,
Criterion I) of VERs recorded from the left-motor area.
Differential abilities analysis.-In order to
explore the aspects of intellectual function that ap-
peared to be most responsible for correlations between
latency and IQ, the latency of peak 2 (Trial 1,
Criterion I), taken to be representative of other
peaks, was correlated with a number of the remaining in-
tellectual and perceptual-motor measures.3 Inspection
of Table 5, which contains these results, reveals that
all the tests which included a cognitive element in
their performance, ranging from verbal abilities to
CORRELATIONS OF INTELLECTUAL AND PERCEPTUAL-MOTOR
TESTS WITH LATENCY OF PEAK 2 (CRITERION 1, TRIAL 1)
V P 8 DSp BD DSy
TAU -.25** -.27** -.26** -.23* -.17 -.32**
Otis PAL MCT-NU DRT SRT
TAU -.26** -.18 -.18 -.10 .03
p < .01
3Random correlations using other peaks of
Criteria I and II yielded essentially the same results.
perceptual speed, and which correlated highly with
Full Scale WAIS IQ, correlated with latency in the
same basic range as did Full Scale WAIS IQ. There was
also a trend for the tests to correlate well with
latency to the degree that they correlated with Full
Scale IQ. Thus, the paired-associate learning task
and the Minnesota Clerical Test tended to correlate
less highly with FS WAIS.and with latency than did
the verbal measures. Simple and disjunctive reaction
time, however, correlated poorly with latency, sug-
gesting that the correlation between latency and in-
telligence seems to be a function of a broad cognitive
factor which is tapped by most tests with intellec-
tual content, but is not a function of simple per-
ceptual motor speed alone.
Additional VER Measures
Comparison of the means of the high and low
IQ groups with Mana-Whitney U Tests on peak-to-peak
amplitude, reliability coefficients, and communality
among brain location correlations showed no clear dif-
ferences between groups based on intelligence.
Correlation Waveform Analyses
When the VER correlations ( 's resulting from
the correlation of each S's VER with that of every
other S) were plotted against the IQ difference scores
(the number resulting from subtracting each S's WAIS
FS score from that of every other S), the scatter plot
which was generated contained about 500 points and was
completely random in appearance for each location. Ken-
dall Taus computed between VER similarity correlations
and IQ difference scores for the 18 most stable Ss and
the 10 least stable Ss across the IQ range yielded no
significant relationships for the left-motor and
occipital areas, and three significant (.05 level) Taus
for the right-motor area. However, these correlations
differed in direction and provided no consistent evi-
dence for any relationship. It was therefore concluded
that Ss with like intellectual abilities do not share
equal electrophysiological likeness.
This conclusion is based, of course, on the
assumption that VER waveform correlations over the en-
tire 500 msec. range of the response provided a valid
measure of similarity between pairs of VERs. It is pos-
sible that correlations between more limited sections of
the waveform, e.g. 100-200 msec., might provide more
positive findings. Inspection of the waveforms sug-
gests that the complexity of the left-motor VERs of
the higher IQ Ss was largely due to smaller peaks found
in the first part of the wave, and it would be logical
to examine this section of the wave first. However, it
seems likely that the same large variance which kept
latency from becoming a good predictor of intelli-
gence in individual cases would also be likely to work
in the case of a grosser measure such as a correla-
The major hypothesis of this study, that intel-
ligence is inversely related to the latency of the
visual evoked response, was supported for VERs recorded
from the left-motor area. Chalke and Ertl's (1965)
study, therefore, was largely replicated. The con-
vergence between these two studies is beet seen in the
independent identification of peaks with similar laten-
cies across the IQ range. This agreement, shown by
Chalke and Ertl's (1965) peaks 3, 4, and 5, and peaks
2, 3, and 4 of the current study (Criterion I), is
illustrated in Figure 6.1
On the other hand, the relationships found in
the present study fell short of those found by Chalke
and Ertl (1965). They were able significantly to dif-
ferentiate their medium IQ group from both the high and
low IQ groups, while the medium grc"p in the present
study fell into the same range as either the high or
1Additional support for the existence of peaks
in this same latency range, without IQ-latency correla-
tions, was provided by Rhodes,et al. (1967).
o---o C & E
n- --a PLUM
Fig. 6. Comparison of Chalke and Ertl's (1965) left-motor VER peaks 3,
4, and 5 with left-motor VER peaks 2, 3, and 4 of the present study.
low groups. Several factors may account for the dis-
crepancies that exist between these two studies. First,
Chalke and Ertl's (1965) low IQ group, which had only
four Ss, ranged in IQ from 50-65. The mean IQ of the
low IQ group in this study was about 90, a mean score
which is much closer to Chalke and Ertl's (1965)
medium IQ group (characterized as having IQs in the
low average range). The Ss in the current study, there-
fore, spanned a smaller range of intelligence and formed
a continuous rather than discrete distribution. Second,
Chalke and Ertl (1965) used a different technique, zero-
crossing analysis, to record their evoked potentials.
While Erti (1966) noted in a separate paper that record-
ings with this equipment produced VERa which were almost
identical to those recorded with an averaging device,
visual comparisons of their VERs and those collected in
the present study show a generally larger number of
identifiable peaks in their Ss. This may be due par-
tially to these differences in VER computer recording
methods, and partially to incidental differences in
variables such as electrodes.
The discrepancy in results may be further ex-
plained by a number of weaknesses in Chalke and Ertl's
(1965) study, including a small number of as (4) in
their low IQ group, uncontrolled age and pupillary
dilation, and,most significant, their failure to
specify objective criteria through which they determined
what deflections of the VER were considered peaks. The
present study rectified these shortcomings and found
the same relationship but in a more attenuated form.
The range of correlations in the present study, more-
over, closely matched those which Callaway and Jones
(1965) found between the latency of auditory evoked
responses and performance on simple cognitive tasks in
children. In addition, the present study focused on
Ss whose IQs lay mainly in the average and above aver-
age range of intelligence, and who were purposely chosen
to form a continuous rather then discrete distribution
to test the value of latency for individual prediction.
This study raises strong doubts, therefore, about the
practical value of substituting VER measures for be-
havioral ones in the measurement of intelligence. The
moderate relationships found in this study, moreover,
possess a certain face validity in view of the large
variety of elements which contribute to the final score
on an IQ test. These include the various factors of
intelligence, non-intellective factors such as set and
attitude, cultural and social factors such as the in-
dividual's learning history, and situational factors
such as time and place of testing.
Correlations between the various intellectual
and perceptual-motor tests and latency indicated that
the relationship between latency and intelligence could
not be explained by reference to any one single intel-
lectual ability, such as recent memory or spatial analy-
sis, but was a function of general intelligence of a
type related to the I of the factor analysts. Latency
was not related, moreover, to simple perceptual-motor
speed alone, as might be expected since this variable
was not related to most of the intellectual measures.
An additional indication that latency may reflect broad
cognitive processes is provided by the results of
studies which have related latency to other matura-
tional or cognitive variables. For example, Ellingson
(1968) has demonstrated that latency has a very strong
relationship to conceptional age and body weight (r -
-.74 to -.87) during development over the first 50
weeks or up to 6 kg. body weight, after which the cor-
relation approaches zero. Engel (1967) found the same
relationship (r -.61) and suggested that photic
latency adds an independent estimate of gestation to
other criteria of maturation in new-borns. In addition,
prolonged latencies have been demonstrated in dementia
(Bankier, 1967; Straumanis, Shagass, and Schwartz,
1965), coma (Lillie, Borlone, Levique, Scherrer, and
Thieffry, 1967) and in old age (Kooi and Bagchi, 1964,
and reduced latencies with an increase in insight in
psychiatric patients (Henninger and Speck, 1966).
The finding that only the latencies of SVER
recorded from the left-motor area were related to in-
telligence provides evidence for the hypothesis that
the left hemisphere is involved in cognitive function-
ing in a way in which the occipital area and right
hemisphere are not. rzaaination of other findings of-
fers firm support of this proposition for the occipital
area (Piercy, 1964). The evidence for the relatively
minor intellectual role of the right hemisphere is much
less clear, but still remains impressive. For example,
many studies of brain lesions have found intellectual
abilities to be more dependent on the left than on the
right hemisphere, while perceptual ability has more
frequently been associated with the right hemisphere
(Teuber, 1962; Weinstein, 1962; McFie and Piercy,
1952; Anderson, 1950, 1951; Reitan and Tarshes, 1959).
This perceptual-intellectual dichotomy continues when
studies are confined to the left- and right-temporal
lobes (Milner, 1962). Other studies have found that
there is a close relationship between aphasia, which is
most frequently due to an impairment of the left hemi-
sphere, and more generalized intellectual impairment
beyond that due to language dysfunction (De Rensi,
Faglioni, Savoiardo, and Vignolo, 1966; Colonna and Fag-
lieni, 1965; Boller and De Resi, 1967). Studies of
two Ss whose cerebral hemispheres were disconnected
have revealed, moreover, that the right hemisphere seemed
able to perform only very simple cognitive tasks, and
could not, for example, calculate, even to the extent of
doubling the numbers 1 to 4 (Sperry, 1966). Finally,
most reviewers have concluded with Piercy (1964) that
"the left hemisphere has prime responsibility for a
wider range of Intellectual2 function" (p. 333).2
Another important outcome of this study was the
description of certain parametric characteristics of
the VER. It was found that the VERs from each location
were stable across time; that the left-and right-motor
VERs resembled each other but not the occipital VERa;
that the occipital VER, but not the left-or right-
motor VER, appeared to have a relatively standard wave-
shape across Ss; and that the right-motor VER was
significantly greater than the left-motor response.
These findings are in close agreement with those of
Rhodes, et al. (1967). For example,thir computation of
five-minute test-retest reliability correlations led
to *'s of .89 for the left-parietal, .93 for the right-
parietal, and .96 for the occipital areas This com-
pared to .83 for occipital, .73 for right-motor, and
2See McFie (1961) and Hecaen (1962) for fur-
ther reviews of this work.
.69 for left-motor in this study. While the latter
figures are noticeably lower, they also do agree in
placing the occipital area as the most stable over
time. The discrepancies may be explained by a longer
test-retest period in the present study (an average of
10 minutes) and the fact that the period studied after
the flash was 500 msec. rather than the 300: msec.
period that was used by Rhodes, et al. (1967). In addi-
tion, each of Rhodes,et all's (1967) electrode-place-
ments varied slightly from those of the present study.
More striking similarity is indicated by comparisons
of the relationships between different areas of the
brain. Rhodes,et, al. (1967) found r's of .88 for
right-parietal left-parietal VERs, .12 for right-
occipital right-parietal, and .05 for left-occipital -
left-parietal areas. This compared with .77 for the
left-motor right-motor areas, .06 for the occipital-
right-motor areas, and .02 for occipital left-motor
areas in the present study. Both studies, therefore,
are in full agreement as to the high relationship
between the hemispheres, and the independence of the
occipital area with respect to each of the other areas.
Finally, measures of similarity between locations
across Ss in each study are in basic agreement that
there is an identifiable occipital VER which is found
in most Ss, while the left and right hemisphere VERa
are much less constant (Rhodes, at al., 1967: left-
parietal .14, right-parietal .23, occipital .64;
the present study: left-motor .00, right-motor .01,
occipital .52) although Rhodes,et al. (1967) again
found higher values for all locations.
It is apparent, therefore, from the results
of this and other studies, that the latency of the VER
bears some significant relationship to intelligence.
Given the nature of VER research, the replication of
this and several parametric findings is striking. On
the'other hand, several studies (Rhodes, et al., 1967;
Shagass, 1967a) have failed to find the IQ-latency
relationship, and it is imperative that future research
be concerned with the reasons for this discrepancy.3
Future work in this arse might take several
forms* First, extensive teplications which pay careful
attention to the details of electrode placement, stimu-
lus and recording apparatus, etc., are of prime impor-
tance. Second, the problem of what aspect of intelli-
gence underlies the correlation might be further ex-
plored. One approach would consist of a refinement of
the technique of correlating a number of intellectual
In addition, the present findings were dis-
crepant with Rhodes,ot al.'s (1967) finding of smaller
VERs, lack of hemispheric differences in amplitude,
and lowered waveform stability in low IQ Ss.
and perceptual-motor measures with latency. A series
of cognitive tasks which havethe lowest possible inter-
correlations, such as those claimed by Guilford (1967),
would lead to a more rigorous test of the hypothesis
that latency is a reflection of a broad overriding
cognitive factor, as suggested by this study, or
whether more careful delineation of abilities might re-
veal patterns which have so far failed to emerge.
Another approach would involve measurement of IQ-latency
relationships in groups of individuals of different
ages. Differential correlations might bear on whether
the neural structure of the brain, as reflected by VER
latency, changes as the intellectual processes based on
the original innate structure are modified through inter-
action with the experience of the organism. In addi-
tion, if the correlation between IQ and latency was
found to be strongest in very young children, and then
declined with increasing age, it could be speculated
that VER latency primarily reflects this basic innate
ability of the individual. If, however, the reverse
were found, then one might speculate that latency is
more reflective of the developing intellectual capacity
that results from the interaction of the individual's
experiential and learning history with his original
This study was designed to replicate and extend
Chalke and Ertl's (1965) original experiment which
found that the latency of the visual evoked response
was inversely correlated with psychometric intelli-
gence, Consequently, the VERs of subjects with a wide
variety of intellectual abilities were recorded from
the left-motor, right-motor, and occipital areas of the
scalp, and compared with these subjects' performances
on a series of intellectual and perceptual-motor tasks.
The VERs were recorded while the subjects' mental
activity remained undirected by the experimenter, and
while they were engaged in simple arithmetic calcula-
tions involving the stimulating flashes.
The major hypothesis of this study, that intel-
ligence is inversely related to the latency of the VER,
was supported for VERs recorded from the left-motor
area, but not for the right-motor or for the occipital
area. This relationship appeared to be a function of
a broad cognitive factor and not a function of simple
perceptual-motor speed. It was most evident in the
first sampling of the subjects' electrophysiological
activity, regardless of whether the Ss' mental activity
was directed by the experimenter.
It was concluded that (1) while VER latency is
significantly related to intelligence, it does not
seem practical at present to substitute VER measures
for behavioral ones in the measurement of intelligence;
(2) that the left hemisphere appears to be involved in
general intellectual activity in a way in which the
right hemisphere and occipital areas are not; and (3)
that future research should focus its attention on
further replications, longitudinal studies, and attempts
to better delineate the aspects of intelligence that
underlie its relationship to VER latency.
Suumary of Statistical Analyses
I. Preliminary Analyses: Parametric Data
A. Test Scores
1. Computation of means, ranges, and
2. Intercorrelation of all scores
B. VER Digital Data
1. Correlations between first and second VER
trial for each location within each
2. Correlations among VER locations for each
trial within each subject,
3. Correlations between each subject's VER and
every other subject's VER for each loca-
tion and each trial
4. Computation of peak-to-peak amplitudes for
VERs from every location and every trial,
a. Comparisons between locations
b. Comparisons between conditions for
II. Main Analyses: VERa and Intelligence
1. Comparison of mean peak latencies of VERs
from subjects in high, medium, and low
IQ groups (separately for each location)
APPENDIX 1 Continued
2. Correlations between the latencies of
all Trial 1 left-motor VER peaks
(Criteria I and II) and Full Scale
3. Correlations between latency of peak 2
(Trial 1, Criterion I) and 11 intellec-
tual and perceptual-motor measures
B. Additional VER measures
1. Comparison of means of high and low IQ
groups on peak-to-peak amplitude,
reliability coefficients, and communal-
ity among brain location correlations
C, Correlation waveform analyses
1. Inspection of scatter plot generated by
correlating VER correlations against IQ
difference scores for each location
2. Correlations between VER correlations
and IQ difference scores in the 18
most stable subjects and 10 least
stable subjects across the IQ range for
Rules for Discrimination of Peaks:
Criteria I and II
Criterion I. High Five
1. No deflection which occurs before 30 msec.
or after 475 Msec. is considered a peak.
2. If two deflections occur at a time interval
equal to or less than 40 msec. apart from
each other, the larger deflection is con-
sidered the peak.
3. A deflection whose trough has an absolute
amplitude in either direction of 12 digital
units (Computer of Average Transients) or
less is not considered a peak, with the
A. A deflection whose trough is equal to
or greater than 5 digital units (CAT)
in both directions is accepted as a
peak if the deflection lasts for 25
msec. or longer within 7 digital units
(CAT) of the high point.
Criterion II. First five
1. No deflection which occurs before 15 msec. or.
greater than 485 msec. is considered a peak.
2. If two deflections occur at a time interval
equal to or less than 25 mseo. apart from
each other, the larger deflection is con-
sidered the peak.
3. To be considered peaks, deflections must be
separated by changes in direction which
last at least 10 msec.
APPEIDIX 2 Continued
4, A deflection whose trough has an
average amplitude in either direction
of less than 7 digital units (CAT)
is not considered a peak.
Intercorrelation Matrix of Test Scores
P I 0 A S DSp
75 68 73
68 55 74 69
72 67 60 66 57
71 75 80 78 68
70 44 62 59 65
64 36 38 46 41
81 47 46 53 52
71 59 47 66 47
78 39 36 48 43
80 69 79 84 81
38 38 38 43 41
-10 -04 -18 -01 -21
-08 -02 -22 -22 -32
69 59 71 73 72
56 43 57 03 61
PC BD PA OA Otis PAL SRT
56 66 58
38 66 58 55
28 34 36 27 42
02 -08 07 02 -10 17
17 -05 -09 -01 -20 02 56
36 54 45 50 77 30 -30
32 36 33 35 58 17 -32
_ __ __ ___
LIST OF REFERENCES
Abrahams, V. C. Some factors affecting the form of
averaged evoked potentials. Proc. physiol. Soc.,
1966, "8 60-61.
Anderson, A. L. The effect of laterality localization of
brain damage on Wechsler-Bellevue indices of
deterioration. J. clin. Psychol., 1950, ,
Anderson, A. L. The effect of laterality localization
of focal brain lesions on the Wechsler-Bellevue
sub-tests. J. clin. Psychol., 1951, 2,
Bankier, R. G. A correlative study of psychological and
EEG findings in normal, physically ill, and
mentally ill seniles. EEG clin. Neurophysiol.,
1967, 22, 189-190. (Abstract)
Barlow, J. S. Rhythmic activity induced by photic
stimulation in relation to intrinsic alpha activ-
ity of the brain in man. EEG olin. Neurophysiol.,
1960, 2, 317-325.
Barry, W., & Ertl, J. Intra-cortical delay and intelli-
gence. Revue de l'Universite d'Ottawa, 1963,
Bartlett, N. R., & White, C. T. Evoked potentials and
correlated judgments of brightness as functions
of interflash intervals. Science, 1965, 146,
Beck, E. C., & Barolin, G. S. Effect of hypnotic sug-
gestions on evoked potentials. J. nerv. ment.
Dis., 1965, 140, 154-161.
Beck, E. C., Dustman, R. E., & Beier, E. G. Hypnotic
suggestions and visually evoked potentials.
EEG clin. Neurophysiol. 1966, 2Q, 397-400.
Begleiter, H., Gross, M., & Kissin, B. Evoked cortical
responses to affective visual stimuli. Psycho-
physiol., 1967, 3, 336-544.
Begleiter, H., Porjesz, B., & Gross, M. Cortical evoked
potentials and psychopathology. Arch. gen.
Psychiat., 1967, 17, 755-758.
Bergamini, L., & Berganasco, D. Cortical evoked poten-
tials in man. Springfield, Illinois: 0.C.
Bergamini, L., Bergamasco, B., Mombelli, A. M. and
Gandiglio, G. Visual evoked potentials in sub-
jects with congenital aniridia. EEG clin.
Neurophysiol., 1965, 12, 594-397.
Bergamini, L., Bergamasco, B., & Mutani, R. Evoked poten-
tials in subjects with intraoranial space-
occupying lesions. EEG clin. Neurophysiol.,
1967, 2_, 186.
Biesheuvel, S., & Pitt, D. Some tests of speed and
tempo of behavior as predictors of the primary-
secondary function temperament variable. J. Nat.
Inst. Personnel Res., 1955, 6, 95-102.
Bigum, H. B., & Whitmer, C. A. Visual and somatonsen-
sory evoked responses of mongoloid and normal
children. Paper read at Western Psychol.
Ass., Cheyenne, Wyoming, May, 1967.
Black, S., & Walter, W. G. Effects on anterior brain
response of variation in probability of associa-
tion between stimuli. J. psychosom. Res.,
1965, 2, 33-45.
Blatt, S., & Offner, F. Distribution of the visual
evoked response of man. EEG clin. Neurophysiol.,
1966, 20, 96. (Abstract)
Boller, F., & De:Ranzi, E. Relationship between visual
memory defects and hemispheric locus of lesion.
Ieurol., 1967, 12, 1052-1058.
Brazier, M. Evoked responses recorded from the deptns
of the human brain. Ann. New York Acad. ScJ.,
1964, 112 33-59.
Buchsbaum, M., & Siivermas, J. Stimulus intensity con-
trol and the cortical evoked response. Psycho-
som. Med., 1968, $M, 12-20.
Callaway, E. Averaged evoked responses in psychiatry.
J. nery. sent. DiPs. 1966, 143, 80-94.
Callaway, E., & Jones, R. T. Evoked responses for the
study of complex cognitive functions. Unpublished
manuscript, Langley Porter Neuropsychiatric
Institute, San Francisco, February, 1968.
Callaway, E., Jones, R., & Layne R. Evoked responses
and segmental set of schizophrenia. Arch. pen.
Psyohiat., 1965, 12, 83-89.
Callaway, Ei, & Layne, R. So Interaction between the
visual evoked response and two spontaneous
biological rhythms: the EEG alpha cycle and the
cardiac arousal cycle. Ann. New York Acad.
Sci., 1964, ;19, 421-431.
Cant, B. R., & Bickford, R. G. The effect of motivation
on the CONV EEG clin. Neurophysiol., 1967, 2,t 595.
Cant, R. B., Pearson, J. E., & Bickford, R. G. The
itchanism of the expectancy wave in man. EE
olin. Neurophysiol., 1966, 21, 622. (Abstract)
Cavonius, C. R. Evoked response of human visual cortex:
spectral sensitivity. Psychonom. Sci., 1965,
Chalke, F. C. R., & Erti, J. Evoked potentials and intel-
ligence. Life Seo., 1965, 4, 1319-1322.
Chapman, R. M. voked responses to relevant and irrelevant
visual stimuli while problem solving. Proc.
73rd Amer. Psychel. Ass., 1965, 177-178.
Chapman, R. H., & Bragdon, H. R. Evoked responses to
numerical and non-numerical visual stimuli while
problem-solving. Nature, 1964, 203, 1155-1157.
Ciganek, L. The EEG response (evoked potential) to light
stimulus in man. EEG clin. Neurophysiol., 1961,
Ciganek, L. Evoked potentials in man: Interaction of
sound and light. BEG olin. Neuroptr siol., 1966,
Ciganek, L. A comparative study of visual, auditory, and
somatosensoryy EEG responses in man. Ex. Brain
11., 1967, 4, 118-125.
M., & Kohn, M. Specific responses of the brain
to color stimuli. Proc. 17th Ann. Conf. Eng.
in Med. Biol., Cleveland, Ohio, 1964, g.
M., Kohn, M., & Lifshits, K. Dynamics and spatial
behavior of light evoked potentials, their modifi-
cation under hypnosis and on-line correlation
in relation to rhythmic components. Ann. New
York Acad. Sci., 1964, 112, 468-509.
Cobb, W. A., & Dawson, G. D. The latency and form in man
of the occipital potentials evoked by bright
flashes. J. Physiol., 1960, 1A, 108-121.
Cobb, W. A., & Morocutti, C. (Eds.), The evoked poten-
tials. EEG Supplement 26, 1967.
Cobb, W. A., Morton, A. B., and Ettinger, G. Cerebral
potentials evoked by pattern reversal and
their suppression in visual rivalry. Nature,
1967, 21, 1123-25.
Cobb, W. A., & Sears, T. A. The superficial spread of
cerebral potential fields. Some evidence provided
in hemispherectomy. EEG clin. Neurophysiol.,
1956, 8, 717-718.
Cohen, J., Offner, F., & Palmer, 0. W. Development of
the CNV in children. EEG clin. Neurophysiol.,
1967, U2, 77* (Abstract)
Cohen, J, & Walter, W. G. The interaction of responses
in the brain to semantic stimuli. PsychoDhysiol.,
1966, 2, 187-196.
Colonna, A. and Faglioni, P. The performance of hemi-
sphere damaged patients on spatial intelligence
tests. Corter, 1965, 2, 293-307.
R., Winter, A. L., Crow, H. J., & Walter, W. G.
Comparison of subcortical, cortical, and scalp
activity using chronically indwelling electrodes
in an. EEG clin, Neurophsiol., 1965, 18,
Copenhaver, R., & Beinhocker, G ...Objective visual field
testing. Occipital potentials evokes from small
visual stimuli. J. Amer. Med. Ass., 1963, 186,
Copenhaver, R., & Perry, N. Factors affecting visually
evoked cortical potentials such as impaired
vision of varying etiology. Invest. Ophth.,
1964, ., 665-673.
Corletto, F., Gentiloeo, A., Rosadini, G., Rossi, G.,
& Zatoni, J. Visual evoked potentials as
recorded from scalp and from visual cortex before
and after surgical removal of the occipital pole
in man. EEG clin. Neurophysiol., 1967, 2g,
Crighel, E., & Ciurea, E. Flash-evoked potentials in
man. Variability of the evoked potentials. EEG
clin. Neurophysiol., 1966, 21, 99. (Abstractl
Dawson, W. D., Perry, N. W., & Childers, D. Flash and
scan stimulation of retinal fields and evoked
.response production. EEG clin. Neurophysiol.,
1968, 24, 467-473.
De Lucchi, N. R., Garoutte, B., & Aird, R. B. The scalp
as an electroencephalographic average. EEG
clin. Neurophysiol.., 1962, LJ, 191-196.
] Renzi, E., Faglioni, P., Savoiardo, M., & Vignolo, L.
The influence of aphasia and of the hemispheric
side of the cerebral lesion on abstract thinking.
Cortex, 1966, 2, 399.
Domino, D., & Corasen, G. Visually evoked response in
anesthetized man with and without induced muscle
paralysis. Ann. New York Acad. Sci., 1964, 112,
Domino, E. F., Corssen, G., & Sweet, R. B. Effects of
various general anesthetics on the visually
evoked response in man. J. Int. Anesth. Res.
Soc., 1963, 42, 735-747.
Donchin, D. Statistical analysis of average evoked
potentials. EEG clin. Neurophsiol., 1966, 21,
Donchin, E., & Cohen, L. Averaged evoked potentials
and intramodality selective attention. EEG clin.
Neurophysiol., 1967, 22, 537-546.
Donchin, E., & Lindsley, D. B. Visual evoked response
correlates of perceptual masking and enhancement.
EEG olin. Neurophysio3., 1965, 19, 325-335.
Donchin, E. & Lindsley, D. B. Average evoked poten-
tials and reaction times to visual stimuli. EEG
clin. Neurophysiol., 1966, 20, 217-223.
Donchin, E., Wicke, J. D., & Lind1ley, D. B. Cortical
evoked potentials and perception of paired flashes.
Science, 1963, lfl, 1285-1286.
Dubouloz, P., Kaphan, G., Corriol, J., & Chau-Huu, D.
New facts on a law which appears to rule the
chronology of evoked potentials. EEG olin. Neuro-
phyaiol., 1966, 21, 514. (Abstract)
Duffy, F. H., Robb, R. M., & Lombroso, C. T. Visual
evoked responses to plain and patterned light
in amblyopia ex-anopsia. EEG c3in. Neuro-
physiol., 1967, 2J, 493. (Abstract)
Dustman, R., & Beck, E. C. Long-term stability of visu-
ally evoked potentials in man. Science, 1963,
Dustman, R. E., & Beck, E. C. Phase of alpha brain
waves, reaction time, and visually evoked poten-
tials. EEG clin. Neurophysiol., 1965a,18, 433-
Dustman, R. E., & Beck, E. C. The visually evoked poten-
tial in twins. EEG. clin. Neurophysiol., 1965b,
Dustman, R. E., & Beck, E. C. Visual evoked potentials:
Amplitude changes with age. Science, 1966, 151,
Eason, R. G., Aiken, L. R., White, C. T., & Lichtenstein,
M. Activation and behavior: II. Visually evoked
cortical potentials in man as indications of
activAtion level. Perc. mot. Skills, 1964, 19,
Eason, R. G., Groves, P., White, C. T., & Oden, D.
Evoked cortical potentials: Relation to visual
field and handedness. Science, 1967, 16*, 1643-
Eason, R., & White, C. T. Averaged occipital responses
to stimulation of sites in the nasal and tem-
poral halves of the retina. Psychonom. Sci.,
1967, 2, 309-310.
Ebe, M., Mikami, T., Aki, M., & Miyazaki, M. Electrical
responses evoked by photic stimulation in human
cerebral cortex. Tohoku J. exp. Med., 1962, 2277
Ellingson R. J. Relationship between EEG and test intel-
ligence: a commentary. Psychol. Bull., 1966a,
Ellingson, R. J. Development of VERs in human infants
recorded by a response average. EG clin.
Neurophysiol., 1966,21, 403. (Abstract)
Ellingson, R. J. Clinical applications of evoked poten-
tial techniques in infants and children. EEG clin,
Neurophysiol., 1968, 24, 293. (Abstract)
Ellingson, R. J., & McBeath, J. Evoked potentials and
serial EEGs following total left hemispherectomy
in a 47 year-old. EEG clin. Neurophysiol., 1967,
22, 190. (Abstract)
Emrich, H. The EEG in hypnotic blindness. EEG clin.
Neurophysiol., 1966, 20, 274. (Abstract)
Engel, R. EEG responses to photic stimulation, and their
correlation with maturation. Ann. New York Acad.
Sci., 1961, 117, 407-412.
Engel, R. I.Maturational differences in EEG responses:
Latencies to photic stimulation of newborn in-
fants in regard to conceptional age, sex, and
race. EEG clin. Neurophysiol., 1966, 21,-413.
Engel, R. Acoustic and phobic response latencies in
Oriental neonates. EEG olin. Neurophysiol., 1967,
22, 293. (Abstract)
Ertl, J. Zero crossings and evoked potentials. Un-
published manuscript, University of Ottawa, 1966.
Ervin, F., & Msrk, V. Studies of the human thalamus: IV.
Evoked responses. Ann. New York Acad. Sci., 1964,
Ferguson, G. A. Human abilities. Annu. Rev. Psychol.,
1965, 16, 39-62.
Ferriss, G. S., Davis, G. D., Hackett, E. R., & Dorsen,
M. Maturation of visual evoked responses in
human infants. EEG clin. Neurophysiol., 1966,
21 404. (Abstract)
Fischer-Williamis, M. The interaction of evoked responses
to flash and to click in the depths and on the
cortex of adult cat and monkey. EEG clin. Neuro-
physiol., 1966, 21, 203. (Abstract)
Garcia-Austt, E. Influence of the states of awareness
upon sensory evoked potentials. EEG clin.
Neurophysiol., 1963, Suppl. 24, 76-89.
Garcia-Austt, E., Bogacz, J., & Vanzulli, A. Signifi-
cance of the photic stimulus on the evoked responses
in man. In J. F. Delafresnaye (Ed.), Brain mechan-
isms and Learning. Oxford: Blackwell, 1961. Pp.
Garcia-Austt, E., Bogacz, J., & Vansulli, A. Effects of
attention and inattention upon visual evoked
responses. EEG olin. Neurophysiol., 1964, 17,
Gastaut, H. Correlations between the electroencephalo-
graphic and the psychometric variables (MMPI,
Rosenzweig, Intelligence tests). EEG clin.
Neurophysiol., 1960, 12, 226-227.
Gastaut, H., & Regis, H. Visual evoked potentials re-
corded transcranially in man. In L. D. Proctor,
& W. R. Adey (Eds.), Symposium on the analysis
of CNS and cardiovascular data using computer
methods. Washington: NASA, SP-72, 1964. Pp.
Goff, W. R. Evoked potential correlates of perceptual
organization in man. Proc. Toddin-ton Conference
on Attention in Neurophysiology, 1967.
Gross, E. G., Vaughan, H. G., & Valenstein, E. Inhibition
of visuals evoked responses to patterned stimuli
during voluntary eye movements. EEG clin
Neurophysiol., 1967, 22, 204. (Abstract)
Guerrero-Figueroa, R. B., & Heath, R. G. Evoked responses
and changes during attentive factors in man.
Arch. Neurol., 1964, 1, 74-84.
Guibal, M., & Lairy, G. C. Preliminary findings concern-
ing the expectancy wave and the visually defi-
cient child. EEG clin. Neurophysiol., 1967, 2t ,
Guilford, J. P. The nature of human intelligence. New
York: McGraw-Hill, 1967.
Haider, M., Spong, P., & Lindsley, D. B. Attention,
vigilance, and cortical evoked potentials.
Science, 1964, 1 180-182.
Halliday, A. .M & Mason, A. A. Cortical evoked poten-
tials during hypnotic anesthesia. EEG clin.
Neurophysiol., l96F, 16, 314.
M. R. Excitability cycles and cortical scanning:
A review of two hypotheses of central intermit-
tency in perception. Psychol. Bull., 1967, 68,
M., & White, C. Perceived number and evoked
cortical potentials. Science, 1967, 156,
Hays, W. Statistics for psychologists. New York: Holt,
Rinehart, and Winston, 1963.
Heath, R., & Galbraith, G. 0. Sensory evoked responses
recorded simultaneously from human cortex and
scalp. Nature, 1966, 212, 1535-1537.
H. Clinical symptomatology in right and left
hemisphere lesions. In V. Mountcastle (Ed.),
Interhemispheric relations and cerebral domi-
nance. Baltimore9 Nd.: Johns Hopkins Press,
i962. Pp. 215-243.
Heilbrun, A. B. Psychological test performance as a
function of laterality localization of cerebral
lesion. J. comp. lphysiol. Psychol., 1956,
Heninger, G., & Speck, L. B. Visual evoked responses and
mental status of schizophrenics during and after
phenothiazine therapy. Arch. gen. Psychiat.,
1966, j1, 419-426.
Hernandez-Peon, E., & Donoso, M. Influence of attention
and suggestion upon subcortical evoked electrical
activity in the human brain. First international
Congress of Feurological Sciences. Brussels,
1959. Vol. III, Electroencephalograhy, clini-
cal neurophysiology, and epilepsy. London:
Pergamon, 1959. Pp. 385-396.
Hill, D., & Parr, G. Electroencephalography, a sym-
posium on its various aspects. New York:
Hillyard, S. A., & Galambos,*R. Effects of stimulus and
response contingencies on a uurface negative
slow potential shift in man. EEG olin. Neuro-
physiol., 1967, 22, 297-304.
Hrbek, A., Vitova, Z., & Mares, P. The development of
cortical evoked responses to visual stimulation
during childhood. Activitas Nervosa Superior,
1966, 8, 39-46.
Irwin, D. A., Knott, J. R. McAdam, 1. W., & Robert,
C. S. Motivational determinants of the "con-
tingent negative variation." =EG clinical Heuro-
physiol., 1966, 21, 538-543.
John, E. R., Herrington, R. N., & Sutton, S. Effects
of visual form on the evoked response. Science,
1967, 155, 1439-1441.
John, E. R., Ruchkin, D. S., and Villegas, J. Experi-
mental background: Signal analysis and be-
havioral correlates of evoked potential configu-
rations in cats. Ann. New York Acad. Sci.,
1964, 112, 362-420.
Jones, R., Blacker, K., & Callaway, E. Perceptual
dysfunction in schizophrenia: clinical and
auditory evoked response findings. Amer. J.
Psychiat., 1966, 123, 639-645.
Jonkman, E. J. The average cortical response to photic
stimulation. Amsterdam: Thesis, 1967,
Jouvet, M., & Courjon, J. Variation of the subcorti-
cal visual responses during attention in man.
EEG olin. Neurophysiol., 1958, 10, 344. (Abstract)
Jung, R. Neuronal integration in the visual cortex
and its significance for visual information. In
W. A. Rosenblith (Ed.), Sensory communication.
Cambridge, Mass.: MIT Press, 1961. Pp. 627-674,