Electrophysiological Properties Of Reaction Time And Aging

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Request accessible transcript

Transcript (if available)

Content ELECTROPHYSIO LOGICAL PROPERTIES
OF REACTION TIME AND
AGING
by
Jytte Busk
A Dissertation Presented to the
Faculty of the Graduate School
University of Southern California
In partial fulfillment of the
Requirements for the Degree
Doctor of Philosophy
(Psychology)
February, 1973
INFORMATION TO USERS
This material was produced from a microfilm copy of the original document. While
the most advanced technological means to photograph and reproduce this document
have been used, the quality is heavily dependent upon the quality of the original
submitted.
The following explanation of techniques is provided to help you understand
markings or patterns which may appear on this reproduction.
1.The sign or “target" for pages apparently lacking from the document
photographed is "Missing Page(s)". If it was possible to obtain the missing
page(s) or section, they are spliced into the film along with adjacent pages.
This may have necessitated cutting thru an image and duplicating adjacent
pages to insure you complete continuity.
2. When an image on the film is obliterated with a large round black mark, it
is an indication that the photographer suspected that the copy may have
moved during exposure and thus cause a blurred image. You will find a
good image of the pege in the adjacent frame.
3. When a map, drawing or chart, etc., was part of the material being
photographed the photographer followed a definite method in
"sectioning" the material. It is customary to begin photoing at the upper
left hand corner of a large sheet and to continue photoing from left to
right in equal sections with a small overlap. If necessary, sectioning is
continued again — beginning below the first row and continuing on until
complete.
4. The majority of users indicate that the textual content is of greatest value,
however, a somewhat higher quality reproduction could be made from
"photographs" if essential to the understanding of the dissertation. Silver
prints of "photographs" may be ordered at additional charge by writing
the Order Department, giving the catalog number, title, author and
specific pages you wish reproduced.
5. PLEASE NOTE: Some pages may have indistinct print. Filmed as
received.
Xerox University Microfilms
300 North Zaab Road
Ann Arbor, Michigan 40100
I
73-18,800
BUSK, Jytte, 1939-
ELECTROPHYSIOLOGICAL PROPERTIES OF REACTION
TIME AND AGING.
University of Southern California, Ph.D., 1973
Psychology, experimental
University Microfilms, A X E R O X Company, Ann Arbor, Michigan
THIS DISSERTATION H A S B E E N M IC R O FILM ED E X A C T L Y A S RECEIVED.
UNIVERSITY O F SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES. CALIFORNIA 9 0 0 0 7
This dissertation, written by
under the direction of hSX .... Dissertation Com
mittee, and approved by all its members, has
been presented to and accepted by The Graduate
School, in partial fulfillment of requirements of
the degree of
D O C T O R OF P H IL O S O P H Y
jJ y .tL tS L .J S u s J s .
_ February 1973
Date.................... .............
'yv^Jrt £
DISSERTATION COMMITTEE
.( .i.
Chmirmau
TABLE OF CONTENTS
Chapter
I.
I
II.
! h i .
i
IV.
INTRODUCTION ...................................................................
Objective
Background
Peripheral Factors
Central processes in slowing of behavior in
senescence
Synapses
The Electroencephalogram
RESEARCH HYPOTHESES..................................................
METHODS...............................................................................
Subjects
Apparatus
Procedure
EEG Analysis
RT Analysis
Statistical Treatment
RESULTS.................................................................................
Reaction time as a function of age and task
Spectral frequency as a_function of age
Prediction of RT from C, auto- and cross-spectra
DISCUSSION...........................................................................
Age effects on RT
Effects of age on EEG auto-spectra
The effects of age on_cross -spectra
The effect of age on C
Prediction of RT from EEG
ii
Chapter
V I. SUMMARY
REFERENCES
LIST OF TABLES
Analysis of variance design for testing hypotheses la and lb ..
Summary of analysis of variance - the effect of age
and task complexity on R T .....................................................
Summary of analysis of variance - the effect of age,
task and electrode derivation on the auto-spectral
power of d e lta ...........................................................................
Summary of analysis of variance - the effect of age,
task and electrode derivation of the auto-spectral
power of th e ta ............................................................................
Summary of analysis of variance - the effect of age,
task and electrode derivation on the auto-spectral
power of alpha I ........................................................................
Summary of analysis of variance - the effect of age,
task and electrode derivation of the auto-spectral
power of alpha II.......................................................................
Summary of analysis of variance - the effect of age,
task and electrode derivation on the auto-spectral
power of b eta............................................................................
Summary of analysis of variance - the effect of age,
task and electrode derivation of the cross-spectral
power of d e lta .........................................................................
Summary of analysis of variance - the effect of age,
task and electrode derivation on the cross-spectral
power of th e ta ............................................................................
Summary of analysis of variance - the effect of age,
task and electrode derivation of the cross-spectral
power of alpha I.........................................................................
Table
11.
12.
13.
Page
Summary of analysis of variance - the effect of age,
task and electrode derivation on the cross-spectral
power of alpha II....................................................................... 44
Summary of analysis of variance - the effect of age,
task and electrode derivation of the cross-spectral
power of beta............................................................................. 45
Multiple intra-S correlations between EEG param eters
and simple and complex RT for young and old Ss............. 60
I
LIST OF FIGURES
Figure Page
1. Block diagram of reaction time apparatus.................................... 26
i
2. The effect of age and complexity of task on R T .......................... 33
]
I 3. Occipital auto-spectra for young and old...................................... 46
4. Parietal auto-spectra for young and old........................................ 47
5. Motor auto-spectra for young and old............................................ 48
6. Frontal auto-spectra for young and o ld ........................................ 49
J 7. Temporal auto-spectra for young and o ld .................................... 50
8. Occipital-parietal cross-spectra for young and old................... 51
9. Occipital-motor cross-spectra for young and o ld ..................... 52
10. Occipital-frontal cross-spectra for young and old..................... 53
| 11. Occipital-temporal cross-spectra for young and old................. 54
{ 12. Parietal-m otor cross-spectra for young and o ld ....................... 55
J 13. Frontal-motor cross-spectra for young and old.......................... 56
14. Frequency of occurrence of a variable entered as one
of the first five predictors in the multiple
regression analysis for simple ta sk ..................................... 65
15. Frequency of occurrence of a variable entered as one
of the first five predictors in the multiple
regression analysis for complex task................................... 66
vi
CHAPTER I
Introduction
Objective
The objective of the study was a description of EEG changes concomi
tant with age related behavioral slowing. Although many studies have de
scribed changes in the EEG with age, very few have obtained simultaneous
behavioral measures indicative of behavioral slowing (Obrist,1963; Surwillo,
1968), or had little success in establishing EEG-RT relationships (Birren,
1965; Boddy,1971; Thompson and Botwinick,1968). The highest correlation
between speed of behavior and EEG has been obtained by Surwillo (1960; 1961;
1963; 1964) who related alpha mean frequency to reaction time. However, his
findings have limited generalizability because his predictor is based on highly
selected EEG periods.
The present project differs from previous studies of EEG and age-
related slowing of behavior in several ways. F irst, the whole frequency
range of EEG is investigated. Secondly, these measures are obtained from
five different cortical areas simultaneously and finally, EEG param eters
reflecting interaction between these brain areas are investigated.
1
Background
Senescence is characterized by a slowing of all behaviors; and due to
the generality of this slowing process, it has been considered as a prim ary
aging factor (Birren,1959). Experimental attempts to determine the neural
mechanisms (e .g ., peripheral or central) most responsible for the slowing
process have included investigations of reaction time (RT), neural transm is
sion rates, and synaptic transm ission (Birren and Botwinick,1955; Birren and
W all,1956; Hugen, N orris, and Shock, 1960; Magladery,1959; M iles,1931;
N orris, Shock, and Wagman,1953; Rabbitt,1965; Singleton, 1955; Suci,
Davidoff, and Surwillo, 1960).
Behavioral measures of RT have in general been classified into two
types, simple and disjunctive. A simple RT paradigm involves the presenta
tion of a single stimulus to the subject (S), who is required to make a response
such as a button press with the shortest possible delay. RT is then defined as
the time elapsing between the stimulus presentation and S's response.
Numerous studies of simple RT have been reported for different sense
modalities. It has been determined that there is an increase in RT as a func
tion of age for all sense modalities and that the increase is independent of the
modality (Birren and Botwinick,1955; Miles, 1931; Singleton, 1954). Some of
the most extensive RT data was collected by Galton's laboratory in the 19th
century from more than 9000 Ss of both sexes and different ages and was later
analyzed by Kogan and Morant (1923). The results showed that simple RT to
both auditory and visual stimulation decreased during childhood and gradually
increased after the age of 40 years. The fastest RT was found between the age
of 18 and 20, who had a mean RT of 182 msec for visual stimuli and 154 msec
for auditory stimuli. In the seventies, the RT had increased to 174 msec and
205 msec respectively, thus demonstrating a sim ilar degree of slowing (20
msec) for the two sense modalities. In addition, Kogan and Morant found that
RT depends little on the acuity of the sense, but rather is more dependent on
central factors.
However, simple RT experiments have been criticized since they do
not typically resemble real life situations. The out-of-laboratory situations
are complex and usually require a series of successive responses to a given
stimulus or a selective set of responses to each of several stimuli. The
results of several studies (e.g. Miles, 1931; Singleton, 1955) have demon
strated that the slowing of RT with age is slightly greater for conditions in
which S has to give a series of multiple responses than for conditions in which
a single discrete response is made. One subset of possible complex tasks is
the disjunctive or multiple choice RT paradigms, where one of several
possible stimuli are presented to S in a given trial and each stimulus requires
a specific response. It has been clearly demonstrated that the more choices
the S has, the more difficult the task and greater the RT (Goldfarb,1941;
Rabbitt,1965; Suci, Davidoff and Surwillo, 1960; Welford,1959).
Goldfarb conducted a series of RT experiments of varying complexity
4
in both young (18-24 years) and old (55-64 years) groups. It was found that the
older group had longer RTs for all tasks. The differences were 11 msec for
the simple RT task, 57 msec for a two choice task and 66 msec for a five
choice situation. Goldfarb concluded that the increase in RT with age
increases with complexity of the task.
Additional evidence for greater slowing with increased complexity of
task has been presented by other investigators. Rabbitt (1965) found that the
response times of old people in a sorting task were more affected by increas
ing the number of stimuli from which they had to discriminate than were the
young. Birren (1955) has reported that line judgment gets progressively
slower with age and the difference between young and old increases as the
discrimination increases in difficulty. These studies have demonstrated that
the increased response times are not due to. slowness of movement, but rather
to an increase of central processing time necessary to perform a specific task
Numerous studies have dealt with the assessm ent of which parts of the input-
output chain are prim arily responsible for the slowing process.
Peripheral Factors
Sensory Factors
One of the first suggestions to account for slower reaction times was
that there were increased sensory deficits due to structural changes of the
sensory organs with age. These changes in the sensory organs then could
diminish the amount of information which would normally be received, and
consequently slow down the decision processes of the organism. Numerous
studies have investigated changes of the visual sensory process as a function
of age and have generally found that there is a decrease in their functioning
(Birren, Bick and Fox, 1948; B irren, Casperson and Botwinick,1950; Birren
and Shock,1950; Goth, Easman and McNelis,1956; G regory,1957; Karpe,
Richenback and Thomasson,1950; Peterson, 1956; W eiss,1959).
The pupillary response has been shown to be important when visual
stimuli are used since the size of the pupil will determine how much light will
reach the retina. For Ss between the ages of 20 to 89 years it has been found
that the pupil size decreases with age both in darkness and in the presence of
1 millilambert illumination (Birren, Casperson and Botwinick, 1950). How
ever, the effectiveness of constriction as a function of intensity is unchanged,
i . e . , the old pupil constricts as much as the young pupil in relation to its
original size at a given illumination. The constriction is slower in old people.
Peterson (1956) found that the latency of constriction was 40 msec longer in a
group of Ss in their sixties than in a group of Ss in their twenties.
It has been repeatedly demonstrated that sensitivity to light decreases
with age. Birren, Bick and Fox (1948) studied the light threshold in the dark-
adapted eye in 130 males (18-83 years) and found that there is a decrease in
light threshold until the end of the twenties, then a gradual accelerating rise
till the sixth decade, whereupon the increase is rapidly accelerating. The
6
increase in threshold from the twenties till the sixties was . 541oguul and to
the eighties 1.23uul. Similar changes have been observed for the lightadapted
eye (McFarland and Fisher, 1955), The rate of adaptation has not been found to
change as a function of age (Birren and Shock, 1950).
In summary, it can be stated that the changes in sensory processes due
to aging causes that the aged receive less information from the environment.
Although the sensory deficits affect RT, studies in which they have been
minimized have shown that the peripheral sensory changes still cannot account
for the longer latencies in RT (Birren, 1964).
Afferent and Efferent Pathways
Several studies have shown that conduction velocity of afferent and
efferent pathways to the brain decrease with age (Birren and Botwinick, 1955;
N orris, Shock and Wagman,1953; Sommer,1941; Wagman and Lesse,1952).
Conduction velocity of the human ulnar nerve has been determined to reach its
peak before the age of ten and begin a pronounced decline around the age of
fifty (Wagman and Lesse,1952). Measured decreases in mean conduction
times from the age 30 to age 80 have varied from 3 msec (Wagman and Lesse,
1952) to 10 msec (Norris, Shock and Wagman,1953). Since the latter only
corresponds to a decrease in RT of 4 msec and general increases in simple
RT exceed 50 msec (Welford,1959), it is evident that decreases in conduction
velocity only accounts for a minor fraction of the slowing of voluntary reactions
with age.
7
This conclusion was confirmed by Birren and Botwinick (1955) who
measured RT of jaw, hand and foot to auditory stimuli. They hypothesized
that if conduction velocity played an important role in the increase of RT with
age, the increase in RT of foot should increase proportionately more than
should RT of hand or jaw, because of the longer pathways to and from the foot.
All RTs increased with age, but since there was no relationship between the
length of the pathway and the increase in RT, it was concluded that conduction
velocity played little role in explaining the increased RT.
Motor Time
Motor time is defined as the interval between the increase in motor
action potential of the muscle eliciting the response and the overt mechanical
response (Surwillo, 1968). This motor component of RT has been investigated
as a possible source of prolonged RT with age. However, no evidence has
been found that motor time contributes to the slowing of RT in senescence
(Botwinick and Thompson, 1966; Surwillo, 1968; W eiss,1956, 1965).
Central Processes in Slowing of Behavior
in Senescence
The previous review of the literature has indicated that peripheral
factors only contribute a sm all percentage to behavioral slowing, while the
integration processes of sensory input and motor output seem to be responsible
for the major proportion of slowing. Since the integration of input and output
8
takes place in the synapses in the central nervous system it is of interest to
determine if there are any age related changes in synaptic delays.
Synapses
Wayner and Emmers (1958) measured the synaptic delays for the mono
synaptic flexor hallucis longus reflex in several groups of rats ranging from
117 days to 822 days. They found that synaptic delay increased progressively
and significantly from a mean value of . 97 msec in the youngest group to
1.36 msec in the oldest group, which represents a 40% increase. This
suggests that a large proportion of slowing of behavior may be accounted for by
summated prolonged synaptic delays of a large number of synapses in the CNS.
Increased synaptic delays with age could account for the differential effect of
age on simple and disjunctive RT. Intuitively, the input-output sequence in
disjunctive RT paradigm passes through more synapses and would therefore be
expected to have greater slowing in RT.
The Electroencephalogram
Since neuronal and synaptic activity is reflected in the electroenceph
alogram (EEG) (Adey,1967; Anderson, 1968; Croutzfeldt,1967; Lindsley,1960)
the EEG should reflect changes in brain activity with age.
Alpha Rhythm, Aging and RT. When S is awake, but in a relaxed state
with closed eyes and/or lacking visual stimulation, his EEG from the occipital
area often contains easily detectable large amplitude waves of 8-12 cps. These
9
are called Alpha waves or Alpha rhythm. As early as 1933 it was noticed that
patients with senile dementia had a slower Alphy rhythm than normal Ss, but
Berger assumed that it was related to their illness, and did not associate the
slowing with aging. Davis (1941), however, demonstrated that the Alpha
rhythm slowed with age, which has been upheld in later research (Busse and
Obrist,1963; Matousek, Volavka, Roubicek and Roth, 1967; Mundy-Castle,
Hurst, Beerstecker and Prinsloo,1954; Obrist,1954; Otomo,1966). Thus, the
mean frequency in young adults is 10.2-10.5 cps (Brazier and Finesinger,
1944), whereas the average frequency for people in their sixties is 9.11 cps,
and for people beyond eighty, it is further decreased to 8.6 cps (Obrist, 1954,
1965). Since these studies were cross-sectional, they did not rule out the
possibility that the decreases were due to selective dropout of Ss. However,
Obrist (1961) found in a longitudinal study that two-thirds of the Ss showed a
progressive decline in frequency over a ten year period. In one S the Alpha
declined from 9.4 cps at the age of 79 to 8.0 cps at the age of 89. In addition,
it was found that dropouts due to deaths occurred more often from those with
slow Alpha than from those with fast rhythms, which indicates that any
observed decreases in Alpha frequency in cross-sectional studies tend to be
underestimates of changes in any individual person.
Alpha recordings from healthy community volunteers between 28 and 99
years of age showed that the Alpha period (reciprocal of Alpha frequency) in
creased at a rate of 4 msec per decade (Surwillo, 1963). Such an increase
results in a 20 msec increase from age 30 to age 80. This finding led to a
series of investigations to determine the possibility that the Alpha rhythm
might function as a basic timing mechanism, and thus be related to slowing of
behavior with age (Surwillo,1960,1961). A significant rank order correlation
of . 81 between RT and Alpha period was obtained in a group of Ss ranging from
18-72 years. The Alpha period measure was computed from the Alpha during
the interval between the stimulus and the response when only clean Alpha was
present.
After the encouraging results Surwillo developed a theory according to
which changes in Alpha rhythm over age could account for changes in RT. His
general hypothesis is that Alpha serves as a m aster timing mechanism in
behavior.
According to his theory, a certain number of waves must occur between
the stimulus and the response in a simple RT experiment. Since Alpha fre
quency is slower in old than in young, a longer time would elapse before an
older person could respond. The theory can be illustrated by considering two
hypothetical Ss, a young S with fast (10 cps) Alpha and an old S with slow
(9 cps) Alpha, performing a simple and a disjunctive RT experiment. Assume
that exactly 2 cycles must elapse between stimulus and the response in a
simple RT experiment. For the young person, two cycles last 200 msec,
while two cycles last 222 msec for the old, a difference of 22 msec. The
theory can thus account for longer simple RT with age.
11
In order for the theory to explain behavioral slowing with age in
general, it must also be capable of accounting for the relatively greater in
crease in RT in older Ss as complexity of the task increases. Assuming that
the complex task requires three cycles to elapse between the stimulus and the
response, the interval will be 300 msec for young and 333 msec for old, a
difference of 33 msec. For young, the increase from simple to complex RT,
thus would be 300 - 200 = 100 msec, while the increase for old would be
333 - 222 =111 msec. Surwillo's theory therefore could account for some of
the differences in simple as well as complex RT for young and old.
His theory has been supported in a series of studies (Surwillo, 1963a,
1963b, 1964a, 1964b) which showed that the slowing of Alpha with age could
account for the slower reactions or longer latencies that are characteristic of
senescence, and the progressively greater time required by elderly individuals
to handle increased amounts of stimulus information (Surwillo, 1967). Never
theless , though the theory and the empirical evidence are impressive it has
certain limitations in term s of generalizability. F irst the decrease in Alpha
frequency can only account for approximately 20 msec increase in RT from
age 30 to 80, while the actual increase most often exceed 50 msec (Welford,
1959). This indicates that there are additional processes in the central
nervous system that are responsible for slowing. Secondly, the results were
based on pure, well developed, synchronized Alpha activity, which is only
present in some people and under special conditions, such as relaxation or
12
lack of visual stimulation. The data, therefore, do not say anything about the
relationship between Alpha cycle and RT when the EEG is activated and
desynchronized. Since the activated state is the more common state during
problem solving, it is evident that investigations of the whole EEG frequency
spectrum during such states are badly needed.
EEG Frequency Spectrum and Age. Most of the investigations of
changes over age have been concerned with Alpha, which is probably because
the 8-12 cps large amplitude regular waves are readily recognized and easily
analyzed by visual inspection of oscillographic recordings. However, Alpha
is only dominant under special conditions, e .g ., when S is relaxed, has closed
eyes, or is lacking visual stimulation. As soon as S is engaged in mental
activity, such as problem solving, the EEG becomes desynchronized into a
composite of superimposed frequencies of lower amplitudes (Lindsley,1952,
1960). To analyze these composite frequencies into component frequencies by
visual inspection is very laborious and unreliable. However, the frequency
analyses can be performed efficiently and reliably by computers, which com
pute auto-spectra expressing the relative power of each frequency. Typically
the data are analyzed into four frequency bands, Delta (1-3. 5 cps), Theta
(3.5-7.5 cps), Alpha (7.5-12.5 cps), and Beta (12.5-25 cps), although greater
frequency resolution can easily be obtained by computer-analysis. At times
it is of interest to know the degree to which two cortical areas share sim ilar
activity in term s of frequency. This can easily be learned by computing
13
cross-spectra that yield a measure of shared activity in each frequency band.
Until now changes in the auto spectrum with age have mainly been
studied by visual inspection of oscillographic recordings or by the use of
electronic filters. It has been found that Delta, Theta, and Alpha emerge
independent of each other between ages one and ten years in children. Typi
cally Delta is prominent until the age of four, while Theta and Alpha dominates
the records of older children (Corbin and Bickford, 1955). Galbraith and
Williams (1972) found that both auto- and cross-spectral mean frequency in
creased linearly with age in a group of Ss ranging from 4 to 14 years. It is
generally agreed that there is little change in the EEG during the adult years
prior to the onset of senescence, but after the onset of senescence several
changes take place.
It has consistently been reported that there is a decrease in the amount
of Alpha activity, a decrease in Alpha amplitude, and a slowing of Alpha fre
quency (Davis,1941; O brist,1954; Surwillo, 1967). Beta activity has been
reported to increase (Matousek, Volavka, Roubicek and Roth, 1967; Mundy
Castle, 1951; Mundy Castle, Hurst, Beerstecker and Prinsloo,1954) though not
always significantly.
The change in Delta and Theta activity seems to depend on the mental
and physical health of the S. Thus in healthy old people, Mundy Castle and
others (Matousek, et al,1967; Mundy Castle, et al. ,1954) have found signifi
cant decreases in Theta activity with age. Matousek, et al. (1967) observed a
14
decrease in Delta activity with age in the tem pero-parietal region when
recording from healthy Ss ranging from 17 to 64 years. On the other hand
increased Delta activity has been reported in senile psychotics (Mundy
Castle, et a l ., 1954; O brist,1964; Obrist and H enry,1958) and in older people
with brain damage (Obrist, 1964; Obrist and Busse,1964). Several studies have
shown that the EEG changes with age are not uniform in all areas of cortex.
The greatest slowing of EEG activity has been observed in the superior
temporal gyrus (Busse, Barnes, Friedman and Kelty,1956; Obrist, 1965) and
this slowing has been associated with intellectual impairment.
EEG Frequency Spectrum, Age and RT. Over several years it was
believed that RT was shorter to a stimulus presented during activated EEG
than during Alpha (Fedio, Mirsky, Smith and P a rry ,1961; Lansing, Schwartz
and Lindsley,1959). In these studies Alpha was blocked by a warning stimulus
to produce the activated condition. In the control conditions no warning
stimulus was given or it was presented at a different time period prior to the
imperative stimulus, such that Alpha blocking either had subsided or had not
occurred at the time of the imperative stimulus. More recent studies
(Thompson and Botwinick, 1966; Leavitt, 1968) have shown that there is no
relationship between the degree of Alpha desynchronization at the time of the
imperative stimulus and RT. The S performs equally well whether Alpha
desynchronization is 0% or 56%. Both RT and Alpha desynchronization a re ,
however, dependent on the length of the period between the warning stimulus
15
and the imperative stimulus with the shortest RT occurring after a 500 msec
preparatory interval (Leavitt, 1968). It is interesting to note that Lansing et al,
(1959) had data to support this conclusion but still emphasized the correlation
between Alpha blocking and RT.
The lack of differential effect on RT of such different EEG patterns
may make one question the validity or usefulness of the EEG as a measure of
central integration of input and output. However, before dismissing it, it is
important to keep in mind that a great deal of information in the EEG is lost
by merely dichotomizing it into Alpha and non-Alpha. For example, if it were
assumed that Theta is associated with long RT, Beta with short RT, and Alpha
with medium RT, grouping together Beta and Theta would eliminate any dif
ferences in RT during Alpha and non-Alpha.
As a matter of fact, specific frequency bands in auto-spectra have been
demonstrated to be sensitive to such variables as level of activation (Walter,
Rhodes and Adey, 1967), anxiety (Adey,1967), levels of stress (Berkhout, Adey
and Walter, 1969), and RT (Galbraith, 1970). In these studies the specific
frequency bands have been used as predictors in multiple regression analyses.
On the basis of the above discussion it therefore seems valuable to investigate
the relationship between the entire EEG frequency spectrum, age and RT.
All the EEG param eters that previously have been used in the investiga
tion of the slowing of brain function with age have been derived from single
areas. None of the indices have reflected the degree to which different brain
16
areas interact with each other. This is surprising, because the very process
that has been investigated is highly dependent on interaction of different brain
areas. For instance, a visual RT task requires transm ission of perceived
sensory input from the occipital-parietal cortex to the motor cortex (probably
via the pre-m otor area) before a response can be executed (Pandya and
Kuypers,1969).
A study by Galbraith and Williams (1972) suggests that cross-spectral
data might provide additional information regarding aging. They observed a
linear increase in both auto- and cross-spectral mean frequency as a function
of chronological age in a group ranging from 4 to 14 years of age. However,
the standard erro r of estimate for the cross-spectral data is less than that of
the auto-spectral data, indicating that the cross-spectrum is a more reliable
index of chronological age in the developmental years. In the present study it
will be investigated if it also relates to aging of the nervous system.
Other statistics for measuring degree of shared activity between brain
areas are available. It has been demonstrated that they are useful in pre
dicting amplitude of evoked potentials (Galbraith, 1967) and in discriminating
between different behavioral states, i . e . , levels of consciousness (Walter,
et al. ,1967), and levels of stress (Berkhout, et al. ,1969), and tasks requiring
different degrees of visual-motor coordination (Busk, 1971). Walter et al. and
Berldiout et al. used linear coherence and cross-spectra. Linear coherence is
an expression of the strength of the linear relationship between the activity of
17
each pair of cortical areas for each frequency band. Galbraith and Busk used a
summary statistic, wrighted-average coherence (C), which expresses the
overall degree of interaction between two brain areas.
Galbraith was able to predict the amplitude of the primary and
secondary components of a visual evoked potential recorded directly from the
occipital cortex (in monkey) on the basis of the obtained C values between a
cortical and subcortical structure prior to the onset of the visual stimulus. As
expected, high C value between structures within the prim ary visual system
yielded high primary and low secondary components in the succeeding visual
evoked potentials, whereas, high C values between the visual system and the
reticular formation resulted in increasing secondary components and lower
prim ary components. This is in agreement with the fact that the earlier com
ponents of the evoked potential are due to visual input via the prim ary pathways
whereas, the secondary components reflect the input via the reticular forma
tion (Rose and Lindsley,1965). The study thus demonstrated that C reflected
known physiological functional relationships.
In Busk's study EEG was recorded via scalp electrodes while S was
engaged in three types of visual-motor coordination tasks requiring different
degrees of visual-motor interaction. C was of greatest magnitude during the
task requiring greatest visual-motor coordination, thus indicating greater
interaction between the relevant areas. After practice of the visual-motor
coordination task C decreased between the occipital and motor areas, which
18
was interpreted to reflect a decrease in interaction or involvement of the two
areas as the task became learned.
C has also been related to individual differences. Thus Galbraith
(1970) found that mongoloids had lower C values than nonretardates. The
observed lower C values in mongoloids are of special interest for investigating
C changes with aging because mongoloids and old people resemble one another
in several aspects (Elam and Blumenthal,1971). Behaviorally, the old re
semble mongoloids in diminished intellectual capacity and in longer RT. In
term s of brain structure there are additional sim ilarities between old and
mongoloids: reduced brain weight, loss of nerve cells, reduction of myelin,
atrophy of the cortex with necrosis and hemorrage, gliosis, pachymeningeal
and perivascular fibrosis. Since blood vessels generally are sm aller in
mongoloids than in age matched non-retarded controls it is possible that the
mongoloid brain suffers from diminished oxygen supply just as the old brain
often does. The external structure of the mongoloid and old brain also have
sim ilar characteristics, as both show flattening and distortion of the convolu
tions and poor fusion of the fissures (Benda, 1962; Brody, 1971; Elam and
Blumenthal ,1971).
Further support for considering C as a promising tool for investigating
cortical correlates of behavioral slowing with age is present as it has been
shown to increase from childhood (4-14 years) to adulthood (19-34) by
Galbraith and Williams (1972).
19
Summary of Chapter I
The literature review has provided evidence that the behavioral slowing
with age is due to changes in the central nervous system. Because of the dif
ficulty of directly investigating CNS changes with age, many investigators have
used the electroencephalogram as an indirect measure of CNS activity and have
found it to change with age. Thus a decrease in alpha and an increase in beta
activity have consistently been reported from age 50. In later years or in Ss
with vascular disease an increase in delta activity has also been observed. In
the preceeding section evidence has been presented suggesting a relationship
between age related changes in the EEG and slowing behavior with age.
For the present study it is therefore assumed that the state of the cen
tra l nervous system is important for the speed of response in a RT task. This
suggests that differing states in the CNS result in moment-to-moment fluctua
tions in RT for a given S and it also suggests that individual differences in a
state of CNS contribute to individual differences in RT. In the present study
various EEG param eters are used as measures of CNS activity in an attempt to
account for differences in RT within, as well as between, Ss by differences in
EEG prior to the presentation of the imperative stimulus.
The present study differs from previous studies in which particular
EEG param eters have been used to account for RT differences in several
aspects. Previous studies have only related alpha period to RT (Surwillo,1961
1963a; Woodruff, 1972). By doing this Surwillo obtained an average intra-S
20
correlation between alpha period and RT of . 41 and an inter-S correlation of
. 72 in 100 Ss ranging from 28 to 99 years of age. Woodruff reported an
insignificant (r=. 018) average intra-S correlation and an inter-S correlation of
. 40 across 30 Ss. These earlier studies have ignored information to be
derived from the interaction between cortical areas. In the present study RT
| and age will be related to the EEG frequencies from 1 .5hz to 20hz in five areas
of cortex as well as to measures of cortical interaction in form of cro ss-
spectra and average coherence in an attempt to account for a greater propor
tion of RT variance.
CHAPTER II I
I
Research Hypotheses
Based on the review of previous work it was the purpose of the present
i
study to further explore cortical correlates of behavioral slowing with age by
investigating (1) EEG changes with age, and (2) predictability of RT from EEG.
The EEG param eters used in the present study were auto-spectra,
cross-spectra, and average coherence (C) computed from the EEG recorded
I
; from the occipital, parietal, m otor, frontal, and temporal areas. j
I i
1 Regarding the investigation of EEG changes with age three relationships
i |
i were tested. The changes in five frequency bands of the five auto-spectra j
i
I !
I were studied in order to replicate earlier studies in which manual of electronic)
I
filtering techniques were used. j
HYPOTHESIS la: There will be an increase in beta j
|
activity and a decrease in alpha, theta, and
delta activity with age in the auto-spectrum.
The effects of age on the cross-spectral frequencies were investigated
to determine how they change with age and if they add new information about
changes in EEG with age.
21
22
HYPOTHESIS lb: There will be an increase in beta
activity and a decrease in alpha, theta, and
delta with age in the cross-spectrum .
The effect of age on C was tested. C was expected to decrease with
age. The rational for this expectation is found in a series of studies already
reviewed, EEG changes during developmental years seem to be reversed during
aging. For example, it has been established that there is an increase in alpha
frequency during development and a decrease during senescence. If C follows
a similar pattern it should decrease during aging, since it has been found to
increase during the developmental years (Galbraith, Williams, and Gliddon,
1973). Further support for a decrease in C with age is provided by the fact
that mongoloids, whose brain structure in many ways is sim ilar to that of old ^
|
people (Elam and Blumenthal,1972), have lower C than age matched normals j
i
(Galbraith, 1970).
i
HYPOTHESIS Ic: C will decrease with age.
The second objective of the study was to determine if EEG prior to the
imperative stimulus could predict RT and, more specifically, if prediction of j
i
RT from EEG could be improved by considering the full range of EEG fre
quencies and m easures of interaction between electrode locations.
HYPOTHESIS Ila: RT will be predicted more
accurately from the full EEG frequency range
and measures of cortical interaction than it has
23
been in previous studies using alpha activity
as the only predictor.
I
In the present study simple as well as complex RT paradigms were
employed. A simple visual RT task may be described as a brightness dis
crimination, since the S merely m ust distinguish between 'signal' and 'no
signal.' A complex RT task may be defined as a brightness and form discrim
ination task, since the S not only needs to detect the signal but also must
determine which of a given number of responses it is before he can respond.
Therefore, it is assumed that EEG recorded from the scalp will be more
I
| affected during a complex RT task than during a simple task. These assump-
j I
> tions led to the following hypothesis: j
HYPOTHESIS lib: The correlation between EEG and J
complex RT will be greater than the correlation
j
between EEG and simple RT.
The study also was intended to determine if the obtained predictors of
RT were unique for each individual o r common across Ss.
CHAPTER III
Methods
Subjects
Sixteen young (20-33 years) and sixteen old (60-76 years) right handed,
healthy Ss were selected. The younger Ss were selected from personnel at
Pacific State Hospital and undergraduates at nearby colleges. The older Ss
were residents of a mobile home park. It was required that none of the Ss
were currently using medication that might affect the EEG and that they had no
known brain damage beyond those physiological changes that accompany old
age. All had adequate vision.
Apparatus
BRS digital logic was used for control and timing of stimulus presen
tation and intertrial intervals. A three channel tachistoscope (Pandora's Box,
Aim Biosciences) was used for the presentation of the imperative stimuli,
which were clearly visible letters on an electroilluminescent panel. The
response apparatus consisted of three telegraph keys placed in a half circle.
They were wired such that a response turned off the tachistoscope and gener
ated pulses on a Hewlett Packard eight channel oscillograph for the
24
25
identification of the key depressed during a given trial. A pulse whose length
equalled RT was recorded on FM tape. A schematic of the RT apparatus is
shown in Figure 1.
Tektronix 2A61 differential amplifiers with 3 db attenuation at 6 and
60 Hz were used for EEG amplication before the EEG was stored permanently
on tape by means of an Ampex FM tape recorder for later data analyses on an
IBM 360/44 computer.
Procedure
Ss were seated in an electrostatically shielded room and low noise
Turskey silver-silver chloride sponge electrodes were attached to the
occipital (O ), parietal (P ), temporal (T ), motor (C ) and frontal (F )
1 O J u O
areas over the left hemisphere. In addition, linked ear reference electrodes
and mastoid ground electrodes were utilized for monopolar recordings. After
dark adaptation, S performed two RT tasks. The room was illuminated with
a red light having an intensity (4 foot-candles) sufficient for an assistant to
observe S's behavior during RT tasks. Each trial consisted of a 500 msec
low intensity but clearly audible tone was a warning stimulus (WS), followed
by an imperative stimulus (IS) after a 4.76 sec. preparatory interval (PI).
Catch trials omitting the IS were randomly interspersed. The time between
trials was 6 secs.
The IS consisted of clearly discernable capital letters, "A", "B", or
"C" on a 10 x 15 cm. electroluminescent panel of the tachistoscope presented
26
TAPE
RECORDER
TACHISTOSCOPE
BRS
DIGITAL
LOGIC
SELECTABLE
SWITCH
WAVEFORM
AND PULSE
GENERATOR
RESPONSE
KEYS
OSCILLOGRAPH
Figure 1. Reaction Time Apparatus.
27
in random order for both the simple and complex RT tasks. Care was taken
that the stimuli were well above threshold and easily seen by the older Ss.
In the simple RT task, S only has one response key available to him at
a time and thus yields the same response to all three stimuli. In the complex
RT task, S has three keys, each associated with a particular letter, and he is
required to discriminate between the three stimuli and give different respon
ses to each stimulus. Since it was possible that one of the responses, right,
left, or forward should take longer for a given S, it was necessary, during the
simple RT condition, to require that S responded on one key during the first
third of the tria ls, on another key during the second third and on the last key
during the remaining trials.
It was proposed to give each S 100 simple and 100 complex RT trials,
but after a couple of pilot Ss it was obvious that fatigue and boredom would
cause considerable artifact in the collected data. Therefore, the number of
trials were reduced to 45 simple and 45 complex trials which reduced a
recording session to 1 hour.
EEG Analysis
EEG signals were amplified by Tektronix 2A61 amplifiers and
recorded on electromagnetic tape throughout the session. Later, the tape
was reproduced at the Engineering Computer Laboratory, digitized at a rate
of 100 sam ples/sec by Adage analog-to-digital converters and analyzed by an
28
IBM 360-44 computer. Five auto-spectra from O , P , T , C , and F ,
1 O u u
ten cross-spectra and 10 C statistics (0„ -P . 0 , -T , ................C„-F„)
' 1 3 1 3 3 3
were computed for each trial and stored on a disc along with RT. Each
spectra was analyzed in the following frequency bands: Delta (1-3.5 cps),
Theta (3. 5-7.5 cps), Alpha I (7. 5-10 cps), Alpha n (10-13 cps) and Beta
(14-20 cps).
RT Analysis
The pulses representing RT were digitized and the computer computed
the length of them.
Statistical Treatment
To test Hypotheses la and lb a series of three way analyses of
variance were performed with age on one dimension, task on the second, and
cortical areas on the third. Table 1 gives an example of one of these
analyses were the above mentioned frequency bands of the five auto-spectra
and ten cross-spectra. This yielded a total of ten analyses. The statistical
model from these is a linear fixed effects model with correlated measures
on the second and third dimensions (Winer, 1962,p. 319). Significant age
effects o r significant age by area interaction with subsequent simple main
effects of age would lead to support of Hypotheses la and lb. To test for
simple main effects for age a special t_ test described by Winer (1962,p .323)
was used.
1
TABLE 1
ANALYSIS OF VARIANCE DESIGN FOR
TESTING HYPOTHESIS la AND lb
Simple RT Task Complex RT Task
P n C F T
3 3 3 3
P„ C F T
3 3 3 3
Young
S .
16
3 17
Old
32
Hypothesis Ic which stated that C would decrease as a function of age
was tested by a three way analysis of variance with age on the first dimension,
task on the second, and cortical areas on the third. The dependent variable
was C. The same statistical model as above applies. Significant main
effects of age, or age by area effects with subsequent simple main effects of
age would lead to support of hypothesis Ic. Test for simple main effects
were performed as above.
Hypothesis II which deals with the prediction of RT from C and auto-
and cross-spectra was tested by means of stepwise multiple linear regression
analyses (Guildford, 1965) and Chi square tests (Edwards, 1968,p. 62).
For each of the 32 Ss two stepwise multiple regression analyses were per
formed, one using the simple RT scores as the dependent variable, and one
using complex RT scores as the dependent variable. Since there were only
45 trials for each task it was necessary to decrease the number of possible
predictors. This was done by collapsing the Delta and Theta band into one
"low frequency band," and collapsing Alpha I and Alpha II into one Alpha band
Thus for each spectrum there were three frequency bands, low frequency,
Alpah, and Beta. A further reduction in number of predictors was achieved
by eliminating any spectra or Cs involving the temporal area. EEG from the
temporal area was eliminated because it contained some artifacts in term s
of muscle activity and blocking of amplifiers. The final set of predictors
thus consisted of four auto-spectra and their six corresponding cross-spectra
31
(each divided into three frequency bands) and six Cs, a total of 36 predictors.
Since it is questionable to use such a large number of predictors with so few
cases the results of these analyses should be viewed with caution. The
analysis can be used to determine if the predictors entered early in the
analysis were common across Ss. Also, the results could be used to test if
EEG param eters can predict complex RT better than simple RT, as stated in
hypothesis I la. Furtherm ore, the analysis could yield information regarding
which variables relate to speed of RT and if they relate positively or nega
tively.
CHAPTER IV ]
j
Results I
!
i
Reaction time as a function of age and task !
I
Figure 2 displays simple and complex RT for young and old. Mean j
simple RT for young was 465 msec and for old 680 msec which yields a dif
ference of 215 msec. Complex RT was 600 msec for young and 870 msec for
old, a difference of 270 msec. The increase from simple to complex RT was
j
thus 135 msec for young and 190 msec for the old. A two-way analysis of j
1
variance (Table 2) with repeated measures showed that the difference between ;
groups was significant (F[l ,30]=11.65,p=. 0005) and the difference between '
I
simple and complex tasks was significant (F[l,30]=209.50,p=. 0005). A !
!
significant group x task interaction showed that the increase from simple to j
i
i
complex RT was significantly greater for the old group (F[l,30]=7.57,p=. 01). j
i
Spectral frequency as a function of age j
In order to determine how age affects the different frequency bands of |
each of the auto- and cross-spectra, a series of analyses of variance were
performed. Specifically this would demonstrate if the power of delta, theta,
and alpha decreased with age while the power of beta increased with age
1000 -
SIMPLE RT COMPLEX RT
800 -
600 -
400 -
200 -
0 Y 0 Y
Figure 2. Reaction Time as a Function of Age and Complexity of Task.
(Y (cross-hatched) ■ Young; 0 (clear) - Old)
Source of Variation
Between Subjects
A (group)
Subjects within groups
Within Subjects
TABLE 2
SUMMARY OF ANALYSIS OF VARIANCE
THE EFFECT OF AGE AND TASK
COMPLEXITY ON RT
Sums of Squares Degrees of
Freedom
Mean Squares F Ratio
3501680.00 31
979424.00
2522256.00
1
30
979424.00
84075.19
11.65
525088. 00 32
B (condition)
AB
B x Subjects within groups
445248.00
16080.00
63760.00
1
1
30
445248. 00 209.50
16080.00 7. 57
2125.33
35
(hypothesis la and lb). Tables 3 through 7 are summary tables of the three
way analysis of variance having groups on the first dimension, tasks on the
second, and the five auto-spectral derivations on the third. Tables 8 through j
i
i
12 are tables from sim ilar analyses except in these the 10 paired derivations I
i
for cross-spectra are on the third dimension. The dependent variable for each
of these analyses is either the power of delta, theta, alpha I, alpha II, or beta
from auto- and cross-spectra respectively.
It is evident that there are no significant main effects of task nor any
interaction of it with any other variables. Hence, none of the EEG param eters
i distinguished between simple and complex RT task. Therefore, the data for
j the two tasks were combined for investigating the effects of age on the EEG. i
i !
i i
i Main effect of age was only present in the alpha II band (F[l,30]=3.61,p=. 01) j
but there were significant age x area interactions in the alpha and beta bands.
Tests for simple main effects of age were performed using t_ tests as de
scribed by Winer (1962, p. 324). The results of these tests are graphically
displayed in 3 through 13. These figures contain the five auto-spectra from
l
O , P , C , F , and T , and those cross-spectra that were significantly I
X 3 3 3 3 |
altered by age. |
In the occipital auto spectrum in Figure 3 it can be seen that the young
group has significantly more alpha I (t=3.295,p=.0005) and alpha II (t=4.324,
p=. 00005) than the old group. Since the degrees of freedom that were
associated with th ese ^ tests were 30 or greater, the critical values were
TABLE 3
SUMMARY OF ANALYSIS OF VARIANCE
THE EFFECT OF AGE, TASK AND ELECTRODE
DERIVATION ON AUTO-SPECTRAL POWER OF DELTA
Source of Variation
Between Subjects
A (group)
Subjects within groups (A)
Within Subjects
B (condition)
AB
B x Subjects within groups (B)
C (area)
AC
C x subjects within groups (C)
BC
ABC
BC x Subjects within groups (BC)
Sums of Squares
**************
1433141248.00
5061071240.00
* * * * * * * * * * * * * *
230621184.00
63963136.00
5811994624.00
12677021696.00
3167092736.00
39686553600.00
170917888.00
272629760.00
6269041920.00
Degrees of
Freedom
31
1
30
288
1
1
30
4
4
120
4
4
120
Mean Squares
1433141248.00
1687025408.00
230621184.00
63963136.00
193733152.00
3169255424.00
791773184.00
330721280.00
43729472.00
68157440.00
52242016.00
F Ratio
.00
1.00
0.12
9.23
2.17
0. 36
1.31
co
o>
TABLE 4
SUMMARY OF ANALYSIS OF VARIANCE
THE EFFECT OF AGE, TASK AND ELECTRODE
COMBINATION ON AUTO-SPECTRAL POWER OF THETA
Source of Variation Sums of Squares Degrees of
Freedom
Mean Squares F Ratios
Between Subjects 3510464512.00 31
A (group) 126976.00 1 126976.00 .00
Subjects within groups (A) 3510337536.00 30 117011248.00
Within Subjects 3301535744.00 288
B (condition) 618496.00 1 618496.00 0.11
AB 2306048.00 1 2306048.00 0.62
B x Subjects within groups (B) 111222784.00 30 3707426.00
C (area) 405721088.00 4 101430272.00 4. 79
AC 160882688.00 4 40220672.00 1.90
C x Subjects within groups (C) 2540384256.00 120 21169856.00
BC 593920.00 4 148480.00 0. 24
ABC 4083712.00 4 1020928.00 1.62
BC x Subjects within groups (BC) 75722752.00 120 631022.87
c o
-a
D U
Source of Variation
Between Subjects
A (group)
Subjects within groups (A)
Within Subjects
B (condition)
AB
B x Subjects within groups
C (area)
AC
C x Subjects within groups
BC
ABC
BC x Subjects within group!
TABLE 5
SUMMARY OF ANALYSIS OF VARIANCE
THE EFFECT OF AGE, TASK AND ELECTRODE
RIVATION ON AUTO-SPECTRAL POWER OF ALPHA I
b)
C)
(BC)
Sums of Squares
* * * * * * * * * * * * * *
4472696832.00
57231674880.00
* * * * * * * * * * * * * *
266354688.00
280141824.00
3674611712.00
9669824512.00
7763218432.00
90243194880.00
354811904.00
406708224.00
6665738880.00
Degrees of
Freedom
31
1
30
288
1
1
30
4
4
120
4
4
120
Mean Squares
4472696832.00
1907722496.00
266354688.00
280141824.00
122487056.00
2417456128.00
1940804608.00
752026624.00
88702976.00
101677056.00
55547824.00
F Ratio
2.34
2.17
2.29
3.21
2.58
1.60
1.83
W
00
TABLE 6
SUMMARY OF ANALYSIS OF VARIANCE
THE EFFECT OF AGE, TASK AND ELECTRODE
DERIVATION ON AUTO-SPECTRAL POWER OF ALPHA II
Source of Variation Sums of Squares Degrees of
Freedom
Mean Squares
Between Subjects
**************
31
A (group) 3844706304.00 1 3844706304.00
Subjects within groups (A) 31907888640.00 30 1063596288.00
Within Subjects
**************
288
B (condition) 2854912.00 1 2854912.00
AB 4005888.00 1 4005888.00
B x Subjects within groups (B) 332595200.00 30 11086506.00
C (area) 4560842752.00 4 1140210688.00
AC 4002922496.00 4 1000730624.00
C x Subjects within groups (C) 23603040000.00 120 196692000.00
BC 565248.00 4 141312.00
ABC 1208320.00 4 302080.00
BC x Subjects within groups (BC) 209293312.00 120 1744110.00
F Ratio
3.61
0.26
0.36
5.80
5. 09
0. 08
0.17
TABLE 7
SUMMARY OF ANALYSIS OF VARIANCE
THE EFFECT OF AGE, TASK AND ELECTRODE
DERIVATION ON AUTO-SPECTRAL POWER OF BETA
Source of Variation Sums of Squares Degrees of
Freedom
Mean Squares F Ratio
Between Subjects 167971232.00 31
A (group) 653024.00 1 653024.00 0.12
Subjects within groups (A) 167318208.00 30 5577273.00
Within Subjects 128610304.00 288
B (condition) 6352.00 1 6352.00 0.07
AB 1664.00 1 1664.00 0.02
B x Subjects within Groups (B) 2622384.00 30 87412.75
C (area) 11789088.00 4 2947272.00 3.60
AC 12152480.00 4 3038120. 00 3. 71
C x Subjects within groups (C) 98160960.00 120 818008.00
BC 88752.00 4 22188.00 0. 72
ABC 83856.00 4 20964.00 0.68
BC x Subjects within groups (BC) 3704768.00 120 30873.07
TABLE 8
SUMMARY OF ANALYSIS OF VARIANCE
THE EFFECT OF AGE, TASK AND ELECTRODE
DERIVATION ON CROSS-SPECTRAL POWER OF DELTA
Source of Variation Sums of Squares Degrees of
Freedom
Mean Squares F Ratio
Between Subjects
**************
31
A (group) 2131820544.00 1 2131820544.00 1.17
Subjects within groups (A) 54539450880.00 30 1817981696.00
Within Subjects
**************
608
B (condition) 229244928.00 1 229244928.00 1.33
AB 139132928.00 1 139132928.00 0.81
B x Subjects within groups (B) 5176950784.00 30 172565024.00
C (area) 8267038720.00 9 918559744.00 15. 52
AC 1246035968.00 9 138448432.00 2.34
C x Subjects within groups (C) 15978526560.00 270 59179728.00
BC 43450368.00 9 4827818.00 0. 49
ABC 128909312.00 9 14323256.00 1.47
BC x Subjects within groups (BC) 2637758250.00 270 9769475.00
TABLE 9
SUMMARY OF ANALYSIS OF VARIANCE
THE EFFECT OF AGE, TASK AND ELECTRODE
DERIVATION ON CROSS-SPECTRAL POWER OF THETA
Source of Variation Sums of Squares Degrees of
Freedom
Mean Squares F Ratio
Between Subjects 3499421696.00 31
A (group) 5468160.00 1 5468160.00 0.05
Subjects within groups (A) 3493953536.00 30 116465104.00
Within Subjects 1451843584.00 608
B (condition) 245760.00 1 245760.00 0.06
AB 876544.00 1 876544.00 0.23
B x Subjects within groups (B) 115986432.00 30 3866214.00
C (area) 391053312.00 9 43450368.00 13.69
AC 53080064.00 9 5897784. 00 1.86
C x Subjects within groups (C) 857243648.00 270 3174976.00
BC 700416.00 9 77824.00 0.66
ABC 684032.00 9 76003.50 0.64
BC x Subjects within groups (BC) 31973376.00 270 118419.87
TABLE 10
SUMMARY OF ANALYSIS OF VARIANCE
THE EFFECT OF AGE, TASK AND ELECTRODE
DERIVATION ON CROSS-SPECTRAL POWER OF ALPHA I
Source of Variation
Between Subjects
A (group)
Subjects within groups (A)
Within Subjects
B (condition)
AB
B x Subjects within groups (B)
C (area)
AC
C x Subjects within groups (C)
BC
ABC
Sums of Squares
* * * * * * * * * * * * * *
1763610624.00
29826485760.00
* * * * * * * * * * * * * *
109662208.00
103305216.00
1461096448.00
2058944512.00
1505058816.00
19881586080.00
100298752.00
110047232.00
BC x Subjects within Groups (BC) 1922580360.00
Degrees of
Freedom
31
1
30
608
1
1
30
9
9
270
9
9
270
Mean Squares
1763610624.00
994216192.00
109662208.00
103305216.00
48703200.00
228771600.00
167228752.00
73635504. 00
11144305.00
12227470.00
7120668.00
F Ratio
1.77
2.25
2.12
3.11
2.27
1.57
1.72
TABLE 11
SUMMARY OF ANALYSIS OF VARIANCE
THE EFFECT OF AGE, TASK AND ELECTRODE
DERIVATION ON CROSS-SPECTRAL POWER OF ALPHA II
Source of Variation
Sums of Squares Degrees of
Freedom
Mean Squares F Ratio
Between Subjects
**************
31
A (group)
2487885824.00 1 2487885824.00 3.01
Subjects within groups (A)
24763729920.00 30 825457664.00
Within Subjects
8628334592. 00 608
B (condition)
2232320.00 1 2232320.00 0.27
AB
2076672.00 1 2076672.00 0. 25
B x Subjects within groups (B) 252141568.00 30 8404718.00
C (area)
998166528.00 9 110907392.00 4.63
AC
812679168.00 9 90297680.00 3.77
C x Subjects within groups (C) 6460940288.00 270 23929408.00
BC
647168.00 9 71907.50 0.20
ABC
446464.00 9 49607.11 0.14
BC x Subjects within groups (BC) 99004416.00 270 366683.00
TABLE 12
SUMMARY OF ANALYSIS OF VARIANCE
THE EFFECT OF AGE, TASK AND ELECTRODE
DERIVATION ON CROSS-SPECTRAL POWER OF BETA
Source of Variation Sums of Squares Degrees of
Freedom
Mean Squares F Ratio
Between Subjects 96414624.00 31
A (group) 1849552.00 1 1849552.00 0. 59
Subjects within groups (A) 94565072.00 30 3152169.00
Within Subjects 28472272.00 608
B (condition) 2048. 00 1 2048.00 0.05
AB
432.00 1 432.00 0.01
B x Subjects within groups (B) 1205424.00 30 40180.80
C (area) 4877936.00 9 541992.87 7.88
AC 3162416.00 9 351379.50 5.10
C x Subjects within groups (C) 18612656.00 270 68935.75
BC
18176.00 9 2019. 56 0.95
ABC 17504.00 9 1944.89 0. 91
BC x Subjects within groups (BC) 575968.00 270 2188.01
Power (W)
400
Young
300
Old
200
100
e
5
« i a,
2
2
F ig u re 3. O c c ip ita l (Oj) A u to sp e c tra fo r Young and O ld. (W Is p r o p o rtio n a l to u V /h z )
0 5
Power (W)
400
Young
300
Old
200
100
0
e 6 a. a ■
F ig u re 4 . P a r i e t a l (P 3 ) A u to sp e c tra fo r Young and O ld. (W is p r o p o rtio n a l to u V /h z )
Power (W)
400
Young
300
Old
200
100
0
8 a.
a ,
F ig u re 5 . M otor (C3 ) A u to sp e c tra fo r Young and O ld. (W i s p r o p o rtio n a l to u V /h z )
Power (VO
400
Young
300
Old
200
100
0
0 8 a- a.
2
F ig u re 6 . F r o n ta l (F^) A u to sp e c tra fo r Young and O ld. (W i s p r o p o rtio n a l to u V /h z )
6 8
F ig u re 7. Tem poral (T j) A u to sp e c tra fo r Young and Old
— m Young
. . . Old
2
(W is proportional to u V /hz) yi
o
Power/100
250
200
Young
150
Old
100
50
0
e 8 a. a .
F ig u re 8 . O c c i p it a l - P a r i e ta l C r o s s -s p e c tr a fo r Young and O ld.
Power/100
F ig u re 9
250
200
Young
150
Old
100
50
0
e 8
P C * 2
O c c ip ita l-M o to r (Oj-C^) C r o s s -s p e c tr a fo r Young and O ld. » o
Power/100
250
200
Young
150
Old
100
50
0
B 9 a2 3
F ig u re 10. O c c ip ita l- F r o n ta l (O j-F^) C r o s s -s p e c tr a fo r Young and O ld.
U 1
C O
Power/100
250
200 -
Young
150 -
Old
100
50 -
e s
F ig u re 11. O c c ip ita l-T e m p o ra l (O j-Tg) C r o s s -s p e c tr a fo r Young and O ld.
Power/100
250
200
Young
150
Old
100
50
0
e
6
«1
a.
U 1
F ig u re 12. P a rie ta l-M o to r (P -C ) C r o s s -s p e c tr a fo r Young and O ld. w
3 3
Power/100
250 -
200
Young
150
Old
100 -
e
5
P
C X I
U 1
F ig u re 13. F ro n ta l-M o to r (F j-C ^) C r o s s -s p e c tr a fo r Young and O ld. 0 4
57
obtained from a table of the normal distribution. The power of the other
frequencies (delta, theta, and beta) for the young group also exceeded those of
the old group but none of them reached significance.
In the occipital-parietal cross-spectrum (Figure 8) the numerical
»
values are lower but the pattern is still the same, the young has more power
in all frequencies but only alpha I (tf3 .646,p=. 0005) and alpha II (t=3.864,
p=. 0005) of the young exceeded those of the old significantly.
In the occipital-motor, occipital-frontal, and occipital-temporal cro ss
spectra the overall pattern starts to change (Figures 9,10,11). The young
| still display more alpha I and alpha II than the old but the differences are less
I pronounced and only in the alpha 11 band does the difference reach significance^
i
; The Rvalues for each of the three spectra were 1.75,2.107, and 1.885,
respectively, and the associated p levels were .05, .02, and .03 (one-tail).
I
In these spectra the old group has much greater mean delta activity though
it does not reach significance.
In comparing the young and old mean auto-spectra (Figure 7) from the
!
prim ary motor area three things should be noted. First, the older group has
significantly more beta activity than the young in this area (t=1.669,p=. 05).
Second, the alpha activity in the young group is no longer predominant as it
was in Ihe occipital area and is therefore no longer significantly different
from the alpha activity in the older group. Third, the power of delta in the
older group exceeds that of the younger group though not significantly. The
increase in delta activity is not significant when using analysis of variance
because there is no overall significance. But if a planned comparison had
been performed it would have reached significance. The significantly greater ;
i
degree of beta activity in the old group is also observed in the parietal-motor
cross-spectrum (t=2.173,p=. 03).
In the frontal area the old group displays more delta activity than the
i
young group, but again since there is no overall significant main effect the
increase must be considered as insignificant. In the motor-frontal cro ss
spectrum the effects from both the motor and frontal auto spectra are present.
Thus the old group has more beta activity (t=2.209 ,p=. 02).
i
The above results can be summarized as follows. Increase in beta 1
i
activity with age is observed in the motor auto-spectrum and in two cross- I
j
spectra involving the motor area, the parietal-m otor, and the motor-frontal, j
i
Decrease in alpha activity with age is evident in the occipital auto-spectrum
and in all cross-spectra involving the occipital area. A significant increase
in delta activity is seen in the motor-frontal cross-spectrum , while theta j
does not change with age. !
I
Effects of age on C
The analysis of variance of C revealed no significant main effects of
age nor any interactions of it with any other variables. The study has thus
failed to support hypothesis Ic that C would decrease with age.
59
Prediction of RT from C. auto- and cross-spectra
Sixty four stepwise multiple regression analyses were performed on
i
!
individual S data in order to determine if it were possible to predict RT from j
the EEG recorded the last 2.56 seconds prior to the imperative stimulus.
Since a very liberal stopping rule was used (Tolerance = .01) and it resulted
in the entering of many variables that added little to the prediction of RT it !
was decided only to consider the first five steps of the analyses.
Although the choice of five variables was somewhat arbitraiy, it was
noticed that all the multiple regression coefficients had an F value of 3. 00 or
greater. Since multiple R represents the maximum correlation between a
|
j dependent variable and a weighted combination of independent variables the
i
1 multiple R is an inflated value and gives a biased estimate of the multiple
correlation of the population. To obtain a better estimate of the population
multiple correlation it is necessary to shrink R before testing it for signifi
cance (Guilford, 1965,p. 401). The corrected multiple Rs are listed in
Table 13. i
i
According to Hays, a multiple R based on five predictors and forty j
i
five cases is significant at the . 05 level if it is equal to or greater than . 49, j
!
and at the . 01 level if it equals or exceeds . 558. It can be seen in Table 13
that within S prediction of RT was significant in 40 out of 64 analyses.
The intra-S R values were transformed to Fisher's z scores and
averaged for each group. The mean intra-S R for predicting simple RT from
60
Young
TABLE 13
MULTIPLE INTRA-S CORRELATIONS
BETWEEN EEG PARAMETERS AND SIMPLE
AND COMPLEX RT FOR YOUNG AND OLD Ss
Subject Simple RT Complex RT
1 .482 .555 *
2 .512* .356
3 .471 .371
4 .465 .545
5 .572 ** .247
6 .407 .628 **
7 .589 .462
8 .386 .417
9 .532 * .700 **
10 .352 .355
11 .375 . 542 *
12 .614** .510*
13 .597 ** .502 *
14 .507 * .514 *
15 .501 * .628 **
16 . 503 * . 628 **
Mean intra-S R .49 .51 *
* p = . 05 ** p = . 01
61
Old
TABLE 13 (continued)
Subject Simple RT Complex RT
17 .464 .497 *
18 .539 * .493 *
19 .589 ** .604 **
20 .566 * .659 **
21 .328 .614 **
22 .664 ** . 481
23 .512 * .734 **
24 . 390 . 421
25 .575 * .689 **
26 .646 ** .578 *
27 .531 * . 523 *
28 . 451 .492 *
29 . 501 * .521 *
30 .410 .477
31 .492 * .295
32 .390 .431
Mean intra-S R . 51 * . 54 *
* P = . 05 ** p = . 01
6 2
EEG in the young group was . 49 and for the old group . 51 while the intra-S R
predicting complex RT from EEG was . 51 and . 54 for the young and old group j
I
i
respectively. All these mean correlations were significantly different from I
j
zero at the . 05 level. It is therefore concluded that EEG significantly j
predicts RT.
It was of interest to determine if RT was equally well predicted in old
and young. A Chi square test showed that the number of significant correla
tions were equally distributed between young and old. Likewise, the mean
intra-S R's from the young group (. 49 and . 51) were not significantly different
from those of the old group (. 52 and . 54). Therefore, it was concluded that
| ,
I prediction of RT from EEG prior to the imperative stimulus did not change |
| :
j with age. '
I |
i It was predicted that complex RT would be more accurately predicted :
!
j from EEG than would simple RT. In order to test this hypothesis the z-tran s-
I
j formed scores from simple and complex tasks, respectively, were tested for
J significance using t_ tests. There were no significant differences (t (30)=. 23). '
I |
The hypothesis was also tested by using a Chi square test, but as there were .
19 significant predictions of simple RT versus 21 significant predictions of j
i
complex RT, the obtained Chi square was insignificant ( 1 .00;df. =1). Hence j
i
the hypothesis that complex RT would be more accurately predicted from EEG|
than simple RT was not supported.
To summarize the above results regarding prediction of RT from EEG
63
it can be stated that RT is predictable from EEG from the last 2.56 sec prior
to the imperative stimulus. Significant prediction is only possible in some
cases but predictability of RT is not related to complexity of task nor to age.
Further studies are necessary to determine why RT is not predictable in all
cases.
The most important experimental question was whether prediction of
RT would improve by using the whole EEG frequency range from five electrode
positions as well as param eters reflecting interaction of cortical areas. To
determine this the mean intra-S R 's and the inter-S R's obtained in this study
was compared to correlations between alpha period and RT obtained by
! Surwillo, who has reported the highest alpha period-RT correlations in the
l
i i
!
I literature. Surwillo (1963a) obtained average intra-S correlation between ,
alpha period and RT of r=, 41 in one hundred Ss ranging from 28 to 99 years of
i
age. Each was given 90 RT trials. Since the mean intra-S multiple correla
tion between multiple EEG param eters and simple RT in the present study was
.50, it was concluded that prediction had improved. The inter individual
prediction in Surwillo's study was . 72 across the whole age range of 28 to 99
I
years. In the present study the inter-S correlation was . 77 for simple RT and!
i
. 85 for complex, so again prediction had improved.
In order to ascertain that alpha was not the main contributor to the
high correlations obtained in the present study, a Pearson r was computed
between RT and alpha for both old and young separately and jointly. Before
64
looking at the results it should be emphasized that the alpha measure in this
study reflects the power, the abundance of EEG in the frequency band 7. 5-12hz
j
and it does not reflect the period that Surwillo uses. The power is dependent |
upon the amplitude and the abundance of alpha present. j
i
The mean intra-S Pearson r between alpha and simple RT was -.14 in
the young group and -.11 in the old group, while the inter-individual correla- '
tion was -. 24 across all Ss in both the simple and complex conditions. Of
course none of these correlations were significant and it is evident that the
power of alpha contributes little to the prediction of RT. This means that
there was no relationship between degree of alpha desynchronization and RT
i j
j in the present study which is in agreement with the previous literature !
j (Thompson and Botwinick,1966; Leavitt, 1968). How this relates to Surwillo's !
j
theory that the alpha period is the m aster timing mechanism of behavior will j
j
be discussed later.
It was also the intent of the present study to determine if the entered
predictors were unique to the individual or common across individuals.
Looking at the individual analyses it was immediately evident that there were
I
no common trends. An attempt to graphically display this can be seen in the !
i
| histograms of Figures 14 and 15 . The 36 independent variables are depicted
I on the abscissa and the frequency of entering each of these variables into the
regression equations during the first five steps is on the ordinate. If the EEG
predictors were identical across Ss only five columns with 16 values each
Frequency o f Occurrence
10 -
. ^UlPl
Young
n P-.fl n-n r~P-, r1 " * !
2 6 11 16 21 26 31 36
10 -
L n n
Old
F I n-, n n J 1
2 6 11 16 21 26 31 36
F ig u re 14. Frequency o f O ccurrence o f a V a ria b le E n tered as One o f th e F i r s t F iv e P r e d ic to rs
in th e M u ltip le R e g re ssio n A n aly sis fo r Sim ple T ask. § * ■
Frequency o f Occurrence
10 -
Young
r H l
2 6 11 16 21 26 31 36
10 -
Old
n-H-u
m
Ik
2 6 11 16 21 26 31 36
Figure 15. Frequency of Occurrence of a Variable Entered as One of the F irs t Five Predictors
in the M ultiple Regression Analysis for Complex Task.
would have been present. It is evident that the selection of variables varied i
j
substantially because almost all variables were selected at least once and no j
variable was selected more than 8 tim es. An investigation of the inter -
i correlation m atrices from a number of Ss showed that the uniqueness of each
| S is even more pronounced than the figure depicts. Not only does the selection
i
! of variables vary from S to S but the sign of the correlation between the EEG
I
J param eter and RT varies. Therefore, it becomes impossible to make a
general statement regarding which variables are correlated with RT and
whether the correlation is positive or negative. j
CHAPTER V
!
I
t
: Discussion
I I
I |
I |
The present study aimed to investigate the cortical correlates of be- |
i ;
' havioral slowing with age. The problem was approached by investigating the j
j i
! differences between cortical activity in young (20-33 years) and old (60-76
i
i
j years) and by relating cortical activity to speed of RT within Ss. Cortical
i activity was defined as average coherence (C) and the power of auto- and
! cross-spectral frequencies from five areas of the dominant hemisphere. The
i
discussion will first focus on changes in RT with age, then on EEG changes
' with age, and finally on EEG correlates of speed of RT.
Age effects on RT
It is noticeable that RTs are longer than in previous studies. This can |
i
probably be attributed to the fact that the response involved a hand movement j
I
I
and not only a finger movement as is typically used. In addition, to prevent j
i muscle artifact in the EEG, S was instructed to keep his neck and shoulders
and upper arm relaxed which might have caused some slowing. However,
although the absolute RTs are longer than typically reported, the behavioral
data are still a replication of previously established results (Welford,1959).
69
Thus, the old group has longer RT than the young group. Complex RT is
longer than simple RT and the increase from simple to complex RT for the old
group was significantly greater than for the young. The fact that the behaviorai
aspects of this study are in agreement with previous literature offers assu r
ance of a number of factors. The two samples (young and old) seemed to be
representative of a young and old population. The procedures used in this
study are comparable to those used in past research of age and RT, and the
restraints that the EEG recording posed on the Ss did not affect the behavioral
responses except to increase RT.
I
! Effects of age on EEG auto-spectra
l
j Hypothesis la, stating that there would be a decrease in alpha, delta,
I and theta and an increase in beta with age, was partially supported by the
results. A significant group electrode interaction indicated that not all areas
showed sim ilar changes, and the tests for simple main effects demonstrated
that the hypothesized differences were only present in some areas. Thus,
although alpha decreased with age in all areas the decrease was only signifi- I
cant in the occipital area, where both alpha I and alpha II decreased signifi
cantly (p=.0005). The increase in beta activity with age was significant only
in the motor area (p=. 05).
The hypothesized decrease in delta and theta was not supported by the
results. The area by electrode interactions were insignificant in the analyses
70
where delta or theta were the dependent variables, and therefore no tests for
simple main effects of age were performed. The lack of change in theta in all ;
i
I
areas is evident in Figures 3 through 7. |
i
The lack of significant change in theta activity is not surprising in view j
I ;
j of past findings. Thus, increases in theta activity with age has been observed
i
t
; in aged psychiatric patients and people with cardiovascular disorders, whereas
small but significant decreases have been observed in healthy Ss (Matousek
et a l.). Matousek et al. concluded that theta decreased significantly with age
I
after obtaining a correlation of -. 207 between age and theta in 106 Ss ranging
from 17 to 64. Although the correlation is statistically significant at the . 05 '
I
I
; level it still only accounts for 4% of the variance. Another possible reason j
| |
; that theta did not change significantly might be that theta is sensitive to
i !
I emotional affect and shows great moment-to - moment fluctuations (Adey,1969). j
i i
! I
| Theta has thus been found to increase during excitement. It is possible that j
the insignificantly greater amount of theta that was present in the old group in
the present study was due to greater excitement or apprehension displayed by
older people during the test situation. In Matousek's study the EEG was
recorded in a resting state with eyes closed whereas in the present study EEG
was sampled in a test situation where the Ss were urged to perform. In
general, older Ss are more aroused by testing situations than are younger
people (Eisdorfer,1968) and this could be the reason for the higher theta
activity.
71
The suggested higher level of activation in the older group is also
supported by the significantly greater power of beta activity and less alpha
activity. Behaviorally the older Ss expressed greater anxiety regarding the j
experimental session. Most of them considered the experiment as a test of
their abilities, although I attempted to make them understand that it was not a
"test" to assess each person's ability. It is possible that the anxiety might !
have been increased because the experiment was conducted in a mental retard a
tion hospital, which was foreign to them. Many of the young Ss were employed
at the hospital and knew the experimental setting. Others had participated in
experiments at college or knew the E. All these factors are likely to have
)
I
reduced the anxiety of the young Ss during the experimental session. The j
i j
behavioral observations as well as the EEG data therefore suggest that the j
older group were at a higher level of activation during the testing situation. i
The implications of higher anxiety level in relationship to performance of RT
task will be discussed later.
I
I
The effects of age on cross-spectra
!
The effects of age on the cross-spectrum were investigated for two j
reasons. First, cross spectra are assumed to reflect interactional processes
between different areas which are important for complex cognitive tasks.
Since the greatest performance deficits with age occur in complex tasks it
seems important to investigate not only auto-spectral changes but also cro ss-
spectral changes with age. Secondly, cross-spectral data have been reported
72
to have lower variability than auto-spectral data and to be a more reliable
predictor of chronological age during the developmental years (Galbraith,
Williams, and Gliddon, 1973).
In general the cross-spectral changes were sim ilar to those of the auto
spectra reflecting an increase in CNS activation in the old group with increased
delta and beta activity and decreased alpha activity. The hypothesis that delta,
theta and alpha would decrease with age whereas beta activity would increase
with age was partially supported, but partially contradicted. As can be seen
in Figures 8 through 11 the old group displayed significantly less alpha activity'
I than the young. It is especially interesting to note, though, that while the
|
> difference was significant for the entire alpha band (7.5-12.5hz) in the occipi-
i
i tal-parietal derivation, the old group only has significantly less alpha II
j
I
| activity (9.5-12.5hz) than the young group in the occipital-motor, occipital-
I
j frontal, and occipital-temporal cross-spectra. This means that only the fast
i
I alpha activity distinguished significantly between young and old, and reflects
the often observed decrease in alpha mean frequency with age (Surwillo,1967).
This information was not obtainable from the auto-spectral data, and it
i therefore shows that the cross-spectra yield important additional information.
The hypothesized increase in beta with age was observed in the parietal-
motor, and motor-frontal cross-spectra (Figure 12 and Figure 13). Although
the differences between the young and old beta activity is less in the two cross
spectra than in the motor auto-spectrum the differences are significant at a
73
lower level. This supports the findings of Galbraith et al. (1973) that cross-
spectral data have lower variability than auto-spectral data.
I
The lower variability in cross-spectral data as compared to auto-
spectral data is also evident in the analysis of variance for the effect of age 01 1
delta (Table 3 and 8). The increase in delta activity in the motor-frontal
' cross-spectrum is less than the increase in delta activity in the frontal auto
spectrum but nevertheless only the difference in the cross-spectrum is signifi
cant.
The observed increase in delta activity was contradictory to the
j hypothesis and to Matousek et al. 's findings. However, the discrepant results
i I
| are probably due to the already mentioned differences in methodology and the j
1
I
| different anxiety levels of old and young. Thus, the relatively greater I
i I
! anxiety displayed by the older group could account for the increase in delta I
I 1
activity since delta activity has been shown to increase during anxiety (Adey,
1969).
The effect of age on C
The data failed to support the hypothesis that C would decrease with
age, since there were no significant differences between the old and young
group. It might be because the sampled old population was not old enough.
The described sim ilarities between the old brain and the mongoloid brain are
more evident in older individuals. Therefore, before rejecting that C de
creases with age a sample of people in their seventies and eighties should be
74
studied. It may also be that the predicted C change only occurs in people with
cardiovascular disorders and diminished blood flow to the brain. Further
studies are therefore indicated.
However, the auto- and cross-spectral results indicate another reason
_ !
for the lack of C changes with age. Adey (1969) reported that coherence in- |
i
creased significantly during focused attention and high anxiety. Therefore if
the older group were more anxious than the young group the old should have
higher coherence levels than the young group. However, if the hypothesized
i
decrease in coherence with age is correct the effect of anxiety and age would
| cancel each other and there would be no differences between the young and old
i i
! I
; group, which is exactly what was observed in the present study. |
I !
1 I
The presented spectral data for the old group reflects a highly 1
activated EEG pattern which can be observed in younger people when they are
autonomically aroused. The autonomic arousal in the young may be caused by
the situation (Adey, 1969) or it may be related to the person's personality and
i
possibly to some psychopathology (McCarron,1972). Thus astronauts during
launch displayed an EEG pattern of increased delta, theta, and beta activity
and very little alpha activity, but this pattern disappeared when the anxiety
producing stimulus situation had disappeared. McCarron describes a sim ilar j
I
EEG pattern in reactive depressive (1972) and in schizophrenics (personal
communication) but in these Ss the pattern was present during a resting condi
tion as well as during the execution of a task. It is well documented in the
75
literature (Kelly and W alter, 1968) that reactive depressives and schizophrenics
have high autonomic arousal level with increased heart rate, diastolic and
I
systolic blood pressure, skinconductance and forearm blood flow. These j
I
patients are in a chronic state of anxiety and display a highly activated EEG
i
pattern at all tim es, during re st as well as during the execution of a task. j
The question is now whether the observed age differences in EEG are
merely caused by a high activation level specific to a testing situation or if the
I
differences would continue to be present if the old group was not anxious about
a testing situation. The present data do not give any answers to the posed
question and neither do the previously cited studies of age changes in EEG. In
I
these studies EEG was recorded from the Ss in a resting situation without the I
!
I requirement of performing a task. However, there is data to siqpport that an
old group may still be more anxious and thereby more aroused than a young in
any experimental situation even if they do not have to execute a task (Eisdorfer,
1968). It would be of value to determine to what extent the age related EEG
changes are a function of increased autonomic arousal in an experimental situa
tion (stimulus related anxiety) and to what extent it is caused by general in
creased arousal at all tim es (trait anxiety). This could be done by obtaining
physiological as well as behavioral m easures during repeated exposures to the
experimental situation. The physiological m easures would include EEG as well
as autonomic m easures of arousal which would be related to the behavioral
m easures. The behavioral m easures would include a performance score as
76
well as scores on a series of anxiety scales reflecting situational and trait
anxiety such as the Taylor Manifest Anxiety Scale and Nowlis Mood Adjective
Check List. By continuing to expose the Ss to the situation till it no longer j
evoked situational anxiety it would be possible to obtain measures without the j
artifact of test anxiety.
i
I
Prediction of RT from EEG
It had been hypothesized that EEG would correlate higher with complex
RT than with simple RT assuming that a complex RT task requires more
cortical involvement than a simple response to a stimulus. The data, however
i
j i
j did not support the hypothesis. In retrospect, considering that the EEG was 1
| \
\
I sampled during the 2. 56 msec prior to the imperative stimulus and therefore '
I ;
| prior to the decision time that distinguishes simple from complex RT task it is
evident that there was an e rro r in the rationale for the hypothesis. Both during
the simple and complex task the EEG was sampled while the S was expecting a
signal and it is likely that the expectation state is equal for the two tasks.
Actually, the present data support the view that the expectation state is
unaffected by task complexity, since the effects of task on C, auto- and cross-
! spectra were insignificant (Tables 3 through 12). The tim e period that dis-
i
tinguishes the two tasks from each other is the decision period and it is likely
that this period would yield differential predictability between simple and
complex task. However, this time period was as short as 250 m secs in some
trials and therefore too short for computing reliable frequency spectra.
77
One of the main objectives of the study was to show that auto-spectral
frequencies beyond the alpha range, and indices of interaction between brain
sites would add valuable information for predicting RT. As mentioned in the
introduction, the best prediction of RT has been obtained by correlating RT anc
alpha period (Surwillo,1961). In that study Surwillo obtained a correlation of
. 81 and was therefore able to account for .64% of the variance. However, this
high correlation was only obtained by mixing inter- and intra-S variability. In
later studies Surwillo (1963a) separating inter-S and intra-S data obtained a
mean intra-S correlation of . 41 between alpha period and RT based on 100
trials and an inter-S correlation of . 72 in Ss ranging from 28 to 99 years.
In the present data the mean intra correlation using 5 predictors is |
. 512 which is significant at the . 05 level (Table 13). Surwillo could account |
for 16% of the variance whereas the present data can account for 25% of the j
variance. It is therefore evident that the addition of frequencies outside the
alpha band and the addition of m easures of interaction improved prediction
within Ss. j
The improvement in prediction may even be enhanced if the EEG was
sampled during the stimulus-response interval comparable to Surwillo. In
attempt to replicate Surwillo, Boddy (1971) only obtained a mean intra-S
correlation of alpha period and RT of only .10 when he sampled the EEG from
the last 5 seconds prior to the imperative stimulus. It is therefore suggested
that the stimulus-response period is m ore closely related to RT than is the
78
period prior to the imperative stimulus.
The inter-S correlation in the present study (R-. 77) was also of greater
i
magnitude than the correlation Surwillo obtained (r-. 72). Thus the present j
study accounted for 59% of the reaction time variance while Surwillo accounted |
i
for 52% of the variance. These data demonstrate that RT can be predicted j
I with greater accuracy using EEG auto- and cross-spectra as well as C as
predictors instead of alpha period.
!
The data also demonstrate an even more important point, namely that
RT can be predicted from EEG patterns that are normally present during
i problem solving tasks and not only from highly selected EEG samples contain-
I
J ing alpha activity. The need for alpha to be present for prediction of RT
I
; following Surwillo's methodology severely lim its the generalizability of
i
i
Surwillo's findings whereas the present methodology does not have such j
! |
l limitations.
What are the implications for the theory that alpha period is the master
timing mechanism that higher correlations between RT and EEG can be j
I
obtained without using alpha period as a predictive param eter? At first glance!
i
the data seem to suggest that alpha period is not the m aster timing mechanism j
of behavior because RT can be adequately predicted without considering the
alpha period. However, the data in the present study are inconclusive to make
such a statement. It is true that RT is predicted without alpha period as one
of the param eters but the study was not designed to investigate to which extent
79
the utilized predictors are correlated with alpha period. It is thus possible
that alpha period indirectly affects or even controls speed of response. It is a
I
possibility, but is not very likely when we consider that RT is predicted from j
I
I
i the EEG equally well regardless how much alpha activity is present. This is ]
indicated by the low insignificant correlations between RT and the power of
alpha in the occipital area. Intra-S correlations thus range from +. 51 to
-.3 4 with a mean intra-S correlation of -. 11 in the young group and -.1 4 in the
old group, while the inter-S correlation was -.24, all of which were significant,
The suggestion that alpha period affects RT by correlating with some of
I the predictive variables in the present study is further contraindicated by the i
I
S fact that each S had his own set of unique set of predictors of RT which would !
I I
! mean that alpha period must correlate with different variables in the different
! Ss. The relationship would have to be even more complex since the same
I
variables in different Ss sometimes correlate positively with RT and some
tim es negatively.
In order to determine if speed of RT is exclusively controlled by alpha !
i
period it would be necessary to execute a study in which the data are analyzed j
according to Surwillo's methodology as well as the present. It would then be I
i
possible to determine to which extent alpha period correlates with the different
EEG param eters used here, and by partial correlations it could be investigatea
which variables truly contribute to prediction of RT. Such a study would be
valuable since no studies have simultaneously predicted RT from alpha period
80
and power spectra and thereby evaluated the relative importance of the dif
ferent m easures.
I
Since it is not possible to determine if the alpha period is the m aster {
t
timing mechanism from the present data the author would like to suggest an j
alternative explanation of the slower responses observed in the old Ss. RT
could be multiply determined in which alpha period, powers of frequency
spectra and coherence measures are some of the variables controlling speed
of RT. Alpha period would then be one of a number of determinants of speed
of behavior. This seems to be a more reasonable hypothesis than viewing
alpha period as the m aster timing mechanism considering not only the present j
j data but also the inability of other authors to replicate Surwillo (Boddy,1970; j
I i
j Woodruff, 1972). Although Woodruff was able to change speed of RT by I
I !
i
modifying alpha frequency by biofeedback which is in support of Surwillo's j
i theory, she was unable to obtain significant intra-S correlations between alpha
period and RT. If alpha frequency is to be the m aster timing mechanism of
behavior it must be able to account for both inter-S and intra-S RT variability,
which it failed to do in Woodruff’s study.
An alternative explanation for slowing with age has been offered by
Eisdorfer (1968), who presents convincing evidence that the old are auto-
nomically over aroused and therefore prolong their decision time. The
present EEG differences between old and young indicate that the older central
nervous system is at a high level of activation with its increased delta and
81
beta activity and decreased alpha activity, so this could possibly account for
the slowing of RT with age.
I
In conclusion it can be said that the objectives of the study have been
met. Central correlates of behavioral slowing with age have been investigated.
The data supported that the older central nervous system is highly activated,
which may be responsible for slowing in behavior. Prediction of RT from j
delta, theta, alpha, and beta as well as m easures of interaction between brain
areas was more accurate than has been obtained in previous studies employing
alpha period as the only predictor. The data imply that speed of behavior is
multiply determined and that alpha most likely is not the m aster timing
mechanism of all behavior. j
i
CHAPTER V I
Summary
The present study investigated the effects of age on EEG and the pre
dictability of RT from EEG. EEG was recorded from sixteen young (20-33
years) and sixteen old (60-76 years) from the occipital (O ), parietal (P ),
X o
motor (C ), frontal (F ), and temporal (T ) areas via scalp electrodes, while
u u u
Ss had 45 simple and 45 three-choice RT trials. Both EEG and RT were
recorded on FM tape and analyzed by computer.
Auto- and cross-spectra and Weighted Average Coherence (C) were
computed from 2. 56 seconds immediately preceeding the imperative stimulus.
There was a significant decrease in alpha activity with age in the occipital
auto-spectrum and the cross-spectra involving the occipital area. A signifi
cant increase in beta activity with age was observed in the motor auto
spectrum and in the motor-frontal cross-spectrum . Delta activity increased
with age in the motor-frontal cross-spectrum , while there were no significant
changes in theta activity. Complexity of task did not affect the EEG recorded
prior to the imperative stimulus.
82
83
Stepwise multiple regression analysis was used to predict RT within S
I from the EEG. The predictors were C, auto- and cross-spectra from the
I
occipital, parietal, motor, and frontal areas. Each spectrum was divided
into three frequency bands, low (2-7.5 hz), alpha (7.5-12.5 hz) and beta
i
! (12. 5-20 hz). The average intra-S multiple R was .512 which was significant
!
at the . 05 level.
j
It is not possible to make any general statement regarding which
variables correlated highly with speed of RT because there were no common
set of predictors across Ss. The data showed, however, that the addition of
j cross-spectral data and frequencies beyond the alpha band improved prediction!
i
! of RT and was valuable for the investigation of age changes in the EEG. !
I I
i |
! i
I
I
l I
j I
84
REFERENCES
Adey, W. R. Intrinsic organization of cerebral tissue in alerting, orienting,
and discriminative responses. In G. C. Quarton, T. Melnechuk, and
F. O. Schmitt (E ds.), The Neurosciences - A Study Program . New
York: Rockefeller University Press, 1967, 615-636.
Andersen, P. and Andersson, A. Physiological Basis of the Alpha Rhythm.
New York: Appleton-Century-Crofts, 1968.
Berkhout, J . , Walter, D. O ., and Adey, W. R. Alterations of the human
electroencephalogram induced by stressful verbal activity. Electro
encephalography and Clinical Neurophysiology, 1969, 2 7 _ , 547-556.
Birren, J. E. Age differences in startle reaction time of the rat to noise and
electric shock. Journal of Gerontology, 1955, 1£, 437-440.
. Handbook of Aging and the Individual. Chicago: University of
Chicago P ress, 1959.
___________, Bick, M. W ., and Fox, C. Age changes in the light threshold
of the dark adapted eye. Journal of Gerontology, 1948 , 3 _ , 267-271.
___________ , and Botwinick, J. Age differences in finger, jaw, and foot
reaction time to auditory stim uli. Journal of Gerontology. 1955, 10,
429-432.
___________ , and Botwinick, J. Speed of response as a function of perceptual
difficulty and age. Journal of Gerontology, 1955, 10, 433-436.
___________ , Casperson, R. C ., and Botwinick, J. Age changes in pupil
size. Journal of Gerontology. 1950, 5, 216-221.
___________ , and Shock, N. W. Age changes in rate and level of visual dark
adaption. Journal of Applied Physiology. 1950, 2, 407-411.
___________ , and Wall, P. D. Age changes in conduction velocity, refractory
period, number of fibers, connective tissue space, and blood vessels
in sciatic nerve of rats. Journal of Comparative Neurology, 1956,
85
Boddy, J. The relationship of reaction time to brain wave period: A re -
evaluation. E lectroencephalography and Clinical Neurophysiology.
1970, 30, 229-235.
Botwinick, J . , and Thompson, L. W. Premotor and motor components of
reaction time. Journal of Experimental Psychology, 1966, 7 1 _ , 9-15.
Brazier, Mary A. B ., and Finesinger, J. E. Characteristics of the normal
electroencephalogram. I. A study of the occipital-cortical potentials
in 500 normal adults. Journal of Clinical Investigation, 1944, 23,
303-311.
Brody, H. Structural changes in the aging nervous system . In H. T.
Blumenthal (Ed.), Interdisciplinary Topics of Gerontology. New York:
H arper, 1971, 7, 87-117.
Busk, J. EEG correlates of visual motor coordination practice. Unpublished
m aster's thesis, University of Southern California, 1971.
Busse, E. W ., Barnes, R. H ., Friedman, E. L ., and Kelty, E. J. Psycho
logical functioning of aged individuals with normal and abnormal
electroencephalograms. I. A study of non-hospitalized community
volunteers. Journal of Nervous and Mental Disease, 1956, 124,
135-141.
___________ , and Obrist, W. D. Significance of focal electroencephalographic
changes in the elderly. Postgraduate Medicine. 1963, 34, 179-182.
Calloway, E. HI, and Yeager, C. L. Relationship between reaction time and
electroencephalographic phase. Science, 1960, 132, 1765-1766.
Creutzfeldt, O. D ., and Kuhnt, U. The visual evoked potential: Physiological
developmental and clinical aspects. Electroencephalography and
Clinical Neurophysiology. Supplement 26, 1967.
Crosby, E. C ., Humphrey, T ., and Lauer, E. W. Correlative Anatomy of
the Nervous System. New York: Macmillan, 1962.
Davis, P. A. The electroencephalogram in old age. Diseases of the Nervous
System, 1941, 2, 77.
Dustman, R. E ., and Beck, E. C. Phase of alpha brain waves, reaction time
and visually evoked potentials. Electroencephalography and Clinical
Neurophysiology, 1965, 18 , 433-440.
86
____________________. The effects of maturation and aging on the wave form
of visually evoked potentials. Electroencephalography and Clinical
Neurophysiology. 1969, 2£, 2-11.
Eccles, J. C. The dynamic loop hypothesis of movement control. In K. N.
Leibovic (Ed.), Information processing in the Nervous System.
New York: Springer Verlag, 1969, 245-269.
Eisdorfer, C. Arousal and performance: Experiments in verbal learning and
tentative theory. In G. A. Talland (Ed.), Hitman Aging and Behavior.
New York: Academic P ress, 1968, 189-216.
Elam, L. H ., and Blumenthal, H. T. Aging in the mentally retarded. In
H. T. Blumenthal (Ed.), Interdisciplinary Topics in Gerontology.
New York: Harper, 1971, 7, 87-117.
Fedio, P. M ., Mirsky, A. G ., Smith, W. J . , and P arry, D. Reaction time
and EEG activation in normal and schizophrenic subjects. Electro
encephalography and Clinical Neurophysiology, 1961, 13^ 923-926.
Galbraith, G. C. The effects of prior EEG "coupling" upon the visual evoked
response. IEEE Transactions in Biomedical Engineering, 1967, 14
223-229.
______________ . Computer analysis of the electroencephalogram and sensory
evoked response. Presented at Palo Alto, 1970.
______________ , Gliddon, J. B ., and Busk, J. B. Electrophysiological
studies of mental retardation. In R. K arrer (Ed.), Developmental
Psychophysiology of Mental Retardation, niinois: Thomas. In press.
______________ , Galbraith, G. C ., Williams, L. B ., and Gliddon, J. B.
Electroencephalographic correlates of age. In preparation. 1973.
Goldfarb, W. An Investigation of Reaction Time in Older Adults. New York:
Teachers College, Columbia University Contributors to Education,
1941, No. 831.
Gregory, R. L. "Neurological Noise" as a factor in aging. Proceedings of
the Fourth International Congress. A. Gerontology, Merano, 1957, 1 _
314-324.
Guilford, J. P. Fundamental Statistics in Psychology and Education. New
York: McGraw Hill, 1965.
87
Guth, S. K ., Eastman, A. A ., and McNelis, J. F. Lighting requirements
for older workers. Illuminating Engineering, 1956, 51, 656-660.
Haider, M ., Spong, P ., and Lindsley, D. B. Attention, vigilance, and
cortical evoked potentials in humans. Science. 1964, 145, 180-182.
Hays, W. L. Statistics for Psychologists. New York: Holt, Rinehart and
Winston, 1963.
Karlin, L. Cognition, preparation, and sensory-evoked potentials. Psycho
logical Bulletin. 1970, 73, 122-136.
Karpe, G ., Richenback, K ., and Thomasson, S. The clinical electro-
retinogram. I. The normal electroretinogram above 50 years of
age. Acta O pth., 1950, 28, 301-305.
Koga, Y ., and Morant, G. M. On the degree of association between reaction
times in the case of different senses. Biometrika, 1923, 15, 355-359.
Kumnick, L. S. Aging and pupillary response to light and sound stimuli.
Journal of Gerontology, 1956, 11, 38-45.
Lansing, R. W ., Schwartz, E ., and Lindsley, D. B. Reaction time and
EEG activation under alerted and nonalerted conditions. Journal of
Experimental Psychology, 1959, 58, 1-7.
Leavitt, F. EEG activation and reaction time. Journal of Experimental
Psychology, 1968, 77, 194-199.
Lindsley, D. B. Psychological phenomena and the electroencephalogram.
Electroencephalography and Clinical Neurophysiology, 1952, 4,
443-456.
Luders, H. The effects of aging on the wave form of the somatosensory
cortical evoked potential. Electroencephalography and Clinical
Neurophysiology, 1970, 29, 450-460.
Magladery, J. W. Neurophysiology of aging. In J . E. Birren (Ed.),
Handbook of Aging and the Individual. Chicago: University of
Chicago P ress, 1959, 173-186.
Matousek, M ., Volavka, J . , Roubicek, J . , and Roth, Z. EEG frequency
analysis related to age in normal adults. Electroencephalography
and Clinical Neurophysiology, 1967, 23, 162-167.
88
McCarron, R. T. Psychophysiological Discriminants of Reactive Depression.
1972. In press.
McFarland, R. A ., and Fisher, M. B. Alterations in darkadaptation as a j
function of age. Journal of Gerontology, 1955, 1£, 424-428.
Miles, W. R. Measures of certain human abilities throughout the life span, j
Proceedings of the Academy of Science. Washington, 1931, 17, j
622-627.
M orrell, L. K ., and M orrell, F. Evoked potentials and reaction times: A
study of intro-individual variability. Electroencephalography and I
Clinical Neurophysiology, 1966, 20, 567-575.
Mundy-Castle, A. C ., Hurst, L. M ., Beerstecker, D. M ., and Prinsloo, T.
The electroencephalogram in senile psychosis. Electroencephalo-
graphy and Clinical Neurophysiology, 1954, 6, 245-252.
I N orris, A. H ., Shock, N. W ., and Wagman, I. H. Age changes in maximum
I conduction velocity of motor fibers of human ulnar nerves. Journal of
Applied Physiology, 1953, 5, 589-593. J
j Obrist, W. D. The electroencephalogram of normal aged adults. Electro- j
j encephalography and Clinical Neurophysiology, 1954, 6, 235-244, I
j
____________. Cerebral Ischemia and the senescent electroencephalogram.
In E. Simonson and T. H. McGavack (Eds.), Cerebral Ischemia.
Springfield, Illinois: Thomas, 1964, 71-98.
____________ . Electroencephalic approach to age changes in response speed.
In A. T. Welford and J. E. Birren (Eds.), Behavior, aging and the
Nervous System. Springfield: C. C. Thomas, 1965, Chapter 13.
____________ , Henry, C. E ., and Busse, E. W. Electroencephalographic
frequency analysis of aged psychiatric patients. Electroencephalo
graphy and Clinical Neurophysiology, 1958, 10, 621-632.
____________ , Henry, C. E ., andJustiss, W. A. Longitudinal study of EEG
in old age. Excerpt a Medical International Congress, 1961, Serial
No. 37, 180-181.
Otomo, E. Electroencephalography in old age: Dominant alpha patterns.
Electroencephalography and Clinical Neurophysiology, 1966, 21,
489-491.
89
Pandya, D. N ., and Kuypers, H. G. J. M. Cortico-cortical connections in
the rhesus monkey. Brain Research. 1969, 13, 13-36.
Peterson, J. H. Die pupillographie und das pupillogram. Acta Physiologica
Scandinavia. 1956, 37 (Supplement 125), 1-141.
Rabbitt, P. M. A. Age and discrimination between complex stimuli. In
A. T. Welford and J . E. Birren (Eds.), Behavior, Aging and the
Nervous System. Springfield: Thomas, 1965, 35-53.
Rose, D ., and Lindlsey, D. B. Visually evoked electrocortical responses in
kittens: Development of specific and nonspecific systems. Science,
1965, 148, 1244-1246.
Satterfield, J. H. Evoked cortical response enhancement and attention in man
A study of responses to auditory and shock stimuli. Electroencephalo
graphy and Clinical Neurophysiology, 1965, 19^, 470-475. i
j Schenkenberg, T. Visual, auditory, and somatosensory evoked responses of '
! normal subjects from childhood to senescence. Unpublished doctoral I
| dissertation, University of Utah, 1970. j
Shagass, C ., and Schwartz, M. Age, personality, and somatosensory
cerebral evoked responses. Science, 1965, 148, 1359-1361. I
i
Singleton, W. T. Age and performance timing on simple skills, hi Old Age
and Modern World. London: Livingstone, 1955, 221-231.
Sommer, J. Synchronisierung motorischer impulse und dire Bedentung fur
} die neurophysiologische Forschung. Z. ges. N eu r., 1941, 172,
500-530.
Spong, P . , Haider, M ., and Lindsley, D. B. Selective attentiveness and
cortical evoked responses to visual and auditory stimuli. Science,
1965, 148, 395-397.
Straumanis, J. J . , Shagass, C ., and Schwartz, M. Visually evoked cerebral
response changes associated with chronic brain syndromes and aging.
Journal of Gerontology, 1965, 20, 498-506.
Suci, G. J . , Davidoff, M. D ., and Surwillo, W. W. Reaction time as a
function of stimulus information and age. Journal of Experimental
Psychology, 1960, 60, 242-244.
90
Surwillo, W. W. Central nervous system factors in simple reaction time. !
American Psychologist. 1960, 15, 419. j
______________ . Frequency of the "alpha" rhythm, reaction time and age.
Nature (London). 1960, 191, 823-824.
|
______________ . The relation of simple response time to brain-wave frequen-j
cy and the effects of age. Electroencephalography and Clinical Neuro- j
physiology. 1963, 15, 105-114. (a)
______________ . The relation of response-time variability to age and the ■
influence of brain wave frequency. Electroencephalography and I
Clinical Neurophysiology. 1963, T5, 1029-1032. (b) |
I
______________ . The relation of decision time to brain wave frequency and to
age. Electroencephalography and Clinical Neurophysiology, 1964, 16,
510-514. (a)
______________ . Some observations on the relation of response speed to j
I frequency of photic stimulation under conditions of EEG synchroniza- j
I tion. Electroencephalography and Clinical Neurophysiology, 1964, 17,!
j 194-198. (b) j
I :
j _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . Timing of behavior and central nervous system. In G. A. j
I Talland (Ed.), Human Aging and Behavior. New York: Academic,
| 1968, 1-35.
I
Thompson, L. W ., and Botwinick, J. The role of the preparatory interval in
the relationship between EEG alpha-blocking and reaction time. J
Psychophysiology, 1966, 3, 131-142. j
________________________________ . Age differences in the relationship
between EEG arousal and RT. Journal of Psychology, 1968, 68, 7-15.
Wagman, I. H ., and Lesse, H. Maximum conduction velocities of motor
fiber of ulnar nerve in human subjects of various ages and sizes.
Journal of Neurophysiology. 1952, 15, 235-244.
Walsh, E.G. Visual reaction time and the alpha rhythm, an investigation of
the scanning hypothesis. Journal of Physiology. 1952, 118, 500-508.
Walter, D. O. Spectral analysis for electroencephalograms, mathematical
determination of neurophysiological relationships from records of
limited duration. Experimental Neurology. 1963, 8, 155-181.
91
______________ , Rhodes, J. M ., and Adey, W. R. Discrimination among
states of consciousness by EEG measures: A study of four subjects.
Electroencephalography and Clinical Neurophysiology, 1967, 22, 22-29.!
i
Wayner, M. J . , J r . , and Em m ers, R. Spinal synaptic delay in young and j
aged rats. American Journal of Physiology, 1958, 194 , 403-405. |
I !
; Weiss, A. D. The relation of A-wave latency of the electroretinogram to j
j human aging. Journal of Gerontology. 1956, 1 1 _ , 448-449.
i !
; |
' ___________ . The locus of reaction time change with set, motivation, and age,|
j Journal of Gerontology. 1956, 20, 60-64.
___________ . Sensory functions. In J. E. Birren (Ed.). Handbook of Aging 1
and the Individual. Chicago: University of Chicago P ress, 1959,
503-542.
I
I Welford, A. T. Psychomotor performance. In J. E. Birren (Ed.), Handbook
of Aging and the Individual. Chicago: University of Chicago P ress, !
! l i i s r i f e s i s : --------------- :
I
j Woodruff, D. S. Biofeedback control of the EEG Alpha Rhythm and its Effect !
on Reaction Time in the Young and Old. Unpublished dissertation. '
University of Southern California, 1972.
Winer, S. N. Statistical Principles and Design. New York: McGraw Hill,
1962.
Zimkina, A. W ., Asofov, B. D ., and Kiseliva, A. M. Peculiarities of the
electroencephalogram in elderly and aged persons. Electroencephalo
graphy and Clinical Neurophysiology, 1965, 18, 107. (Abstract)

Asset Metadata

Creator Busk, Jytte (author)

Core Title Electrophysiological Properties Of Reaction Time And Aging

Contributor Digitized by ProQuest (provenance)

Degree Doctor of Philosophy

Degree Program Psychology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag oai:digitallibrary.usc.edu:usctheses,OAI-PMH Harvest,psychology, experimental

Format dissertations (aat)

Language English

Advisor Galbraith, Gary C. (committee chair), Birren, James E. (committee member), Walkon, George (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-804798

Unique identifier UC11364330

Identifier 7318800.pdf (filename),usctheses-c18-804798 (legacy record id)

Legacy Identifier 7318800

Dmrecord 804798

Document Type Dissertation

Format dissertations (aat)

Rights Busk, Jytte

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA