Content uploaded by Jan Berend Deijen
Author content
All content in this area was uploaded by Jan Berend Deijen
Content may be subject to copyright.
Psychopharmacology (1992) 109: 489-496
Psychopharmacology
© Springer-Verlag 1992
Vitamin B-6 supplementation in elderly men:
effects on mood, memory, performance and mental effort
J.B. Deijen*, E.J. van der Beck 2, J.F. Orlebeke 1, and H. van den Berg 2
1 Department of Psychology, Psychophysiology Division, Vrije Universiteit, De Boelelaan 1111, NL-1081 HV Amsterdam, The Netherlands
2 Department of Human Nutrition, TNO Toxicology and Nutrition Institute, Postbus 360, NL-3700 AJ Zeist, The Netherlands
Received March 12, 1992 / Final version May 29, 1992
Abstract.
This study evaluates the effects of vitamin B-6
supplementation (20 mg pyridoxine HCL daily for 3
months) on mood and performance in 38 self-supporting
healthy men, aged between 70-79 years. Effects were
compared with 38 controls who received placebo and
were matched for age, plasma pyridoxal-5'-phosphate
(PLP) concentration and intelligence score. Before and
after drug intervention vitamin B-6 status was determined,
and mood and performance were measured by means of a
computerized testing system. In addition, the phasic pupil
response was measured in order to assess mental effort.
Positive effects of vitamin B-6 supplementation were only
found with respect to memory, especially concerning long-
term memory. In view of the finding that mental per-
formance improvement and A PLP values were most
strongly correlated within an intermediate range of
A PLP, it is suggested that cognitive effects are primarily
associated with a certain range of vitamin B-6 status
increment. The general conclusion is that vitamin B-6
supplementation improves storage of information mod-
estly but significantly.
Key words:
Vitamin B-6 - Aging - Mood - Memory ....
Performance - Pupil size - Mental effort
Compared to younger persons, elderly people are found to
exhibit deterioration of memory functions (Gilbert and
Levee 1971; Poon 1985). Age-related changes become
manifest in a capacity decline in memory storage and
retrieval and in a reduced memory processing speed. In
addition to memory decline, other cognitive functions also
show impairment in the aging process. These cognitive
changes are partly explained by the assumption that
decrease in information processing speed results from
changes in neuro-endocrine functioning in the aging pro-
cess, causing a reduction in signal/noise ratio (Welford
1985).
Correspondence
to.' J.B. Deijen
Apart from these intrinsic age-related changes, there
are indications that extrinsic factors like food intake are
capable of influencing the speed of the aging process in
this respect. A number of studies have been performed on
the effects of nutritional status upon mental functioning.
Most of these studies investigated the relation of sub-
optimal vitamin status and behavior. For instance, in one
of the first studies on the relation between nutritional
status and cognitive functioning, healthy elderly subjects
with low blood levels (bottom 5% and 10% of the sample)
of vitamin B-12 or C scored worse on a memory and an
abstract thinking task than the remaining 90% of the
sample (Goodwin et al. 1983). In a comparable study in
healthy elderly subjects a marginal status of thiamin,
riboflavin, vitamin B-12 or C was associated with impair-
ment of psychological functioning as indicated by depres-
sion, emotional instability, fatigue and short-term mem-
ory impairment (Chom6 et al. 1986). Recently, a double-
blind supplementation study was performed in healthy,
young males. In this study thiamin, riboflavin, vitamin B-
12, C, E and folic acid status were determined. It appeared
that subjects who were deficient in one or more of these
vitamins showed an improvement in well-
being after vitamin supplementation for 8 weeks, but not
in performance. Thus, supplementation improved mood
only in persons with a suboptimal vitamin status at the
start of the study (Heseker et al. 1990).
Vitamin B-6 plays an important role in the synthesis of
a number of neurotransmitters, e.g. gamma-aminobutyric
acid (GABA), dopamine, serotonin and noradrenalin.
Hence, especially with respect to vitamin B-6, a relation-
ship between vitamin status and mental functioning can
be assumed. Moreover, defects in vitamin B-6 metabolism
are associated with several mental diseases (Merrill and
Henderson 1987).
However, the number of controlled trials which estab-
lished the relation between vitamin B-6 status and cogni-
tion in humans is very limited. In a review of controlled
trials in humans (Kleijnen and Knipschild 1991) only two
trials on the effects of vitamin B-6 on psychological
functioning in healthy adults (Molimard et al. 1980) were
reported. These studies, which used mega-dosages (500
490
and 1000 mg), failed to show positive effects. In an unpub-
lished study by Kleijnen and Knipschild, 20 elderly men
and women were supplemented for a year with a daily
dosage of 2 mg vitamin B-6, while 24 elderly (matched for
age, gender and general condition) received a placebo in a
double-blind fashion. Positive results were found on one
cognitive test ("clock test", which measures visual recogni-
tion and drawing ability) out of five, indicating a (slight)
improvement in recognition and motor ability (Tolonen
et al. 1988).
Elderly people have in general a lower food intake,
resulting in a lower intake of necessary nutrients, includ-
ing vitamin B-6 (Rose et al. 1976; Baker et al. 1979;
Guilland et al. 1984; Schrijver et al. 1985, 1987). Moreover,
it is assumed that the need for certain nutrients, like
vitamin B-6, is even higher in elderly people (L6wik et al.
1989).
In a recent study with healthy elderly (65 79 years) in
the Netherlands it was found that the mean vitamin B-6
intake was about 70% of the Dutch Recommended Diet-
ary Allowances (RDA) for men and 55% for women. In
addition, in 45% of the men and 13% of the women
studied plasma pyridoxal-5'-phosphate (PLP) levels,
which is the active form and indicator of circulating store
of vitamin B-6, were found to be below the P2.5 of healthy
young adult blood donors ( < 20 nmol/1). On the basis of
an other indicator of vitamin B-6 status, i.e. s-EAST,
which is a measure for the availability of vitamin B-6 on
cellular level, 9% of the elderly had a suboptimal vitamin
B-6 status (L6wik et al. 1989).
In the present study the effects of vitamin B-6 supple-
mentation on cognitive functioning and immune functions
were investigated. Effects on immunocompetence will be
reported elsewhere. The objective of the study was to
determine the usefulness of vitamin B-6 supplementation
in healthy elderly. Because of the low PLP levels especially
found in males, this double-blind study was carried out in
men.
In order to control for the possibility that vitamin B-6
supplemented subjects might show the same level of per-
formance while exerting less mental effort compared to
controls, the cognitive assessment included the measure-
ment of the phasic pupil response as an indicator of
mental effort (Payne et al. 1968; Peavler 1974; Beatty and
Wagoner 1978; Beatty 1982). Moreover, if the controls
made use of compensatory effort to maintain their level of
performance, this would also be reflected in the recorded
pupil size.
Materials and methods
Subjects
Subjects were recruited by means of advertisements in local papers.
From about 130 responders 82 males were selected on basis of the
in- and exclusion criteria. The following
inclusion criteria
were used:
apparently healthy/age between 70 and 79 year/restricted alcohol
use (a maximum of four glasses a day)/consent to inform the
physician/intelligence score above 80.
The
exclusion criteria were:
use of drugs affecting the vitamin B-6
metabolism, i.e. levodopa, (di) hydralazine, amitriptyline, isoniazide,
amiodarone, penicillamin and MAO inhibitors/use of drugs affec-
ting the immune reactivity i,e, corticosteroids and cytostatics/use of
vitamin B-6 supplements within 3 months before the study period/
auto-immune diseases, like rheumatoid arthritis/use of tong-acting
hypnotics, i.e. flunitrazepam, flurazepam, quazepam/use of antide-
pressants within a month before study period/addicted to drugs or
alcohol/abnormal clinical chemical-haematological profile/sensory
or motor defect which may affect performance.
Informed consent was obtained from each subject and the study
protocol was approved by both the Medical Ethical Committee of
the Free University Hospital and the TNO Toxicology and Nutrit-
ion Institute.
Half of the 82 subjects were assigned to the placebo group and
the other half to the vitamin B-6 group. Three subjects dropped out
because of illness. Because of the matched pairs procedure, which is
described in the next section, this lead to the exclusion of three other,
matched subjects. Data of the remaining 38 pairs were analysed.
At the start of the study 13 subjects (9 placebo, 4 vitamin B-6) of
the group of 76 participants (17%) were defined as marginally
deficient on the basis of plasma PLP (PLP < 20 nmol/1) and three
subjects (4%) on the basis of a-EAST ( > 1.98), all three belonging to
the placebo group. Despite this uneven distribution, the mean PLP
and a-EAST values on session 1 were nearly identical for both
groups. The most important characteristics of the subjects are
summarized in Table 1.
General procedure
Subjects who responded to the advertisements were invited for an
information meeting, where they were informed about the goal and
procedure of the study and where intelligence and vitamin B-6 status
were determined. In order to assess intelligence, a condensed version
of the Groninger Intelligence Test was administered (De Zeeuw
1971). Furthermore, blood samples were taken to determine the
vitamin B-6 status and a number of biochemical parameters. Finally,
data were collected with respect to medical status, use of drugs,
alcohol and food supplements by means of a questionnaire.
Pairs of subjects were formed, each pair matched on age, vitamin
B-6 status and IQ scores. Subsequently, each one of the two was
randomly allocated to the placebo group or the B-6 group, accord-
ing to a double-blind procedure. The mean age, vitamin B-6 status
and IQ score are shown in Table 1. Just before the start of the study
blood samples were taken again to determine vitamin B-6 status, in
addition to haematological and biochemical parameters. Subjects
also practised the test battery to get acquainted with the cognitive
tests. This was done in order to minimize learning effects. After this
practice phase the subject was invited to test session 1, followed by a
treatment period of 12 weeks. The treatment period was followed by
test session 2, on which occasion a parallel form of the test battery
was administered. Blood samples were also taken again. Treatment
consisted of vitamin B-6 or placebo. Afterwards subjects were
informed if they had been given placebo or vitamin B-6.
Dependent measures
Mood: adjective check list.
The "list of adjectives" (Janke and Debus
1978) consists of 161 items and is used to assess six major groups of
Table
l. Mean age, IQ and vitamin B-6 status (PLP, ~-EAST) of
both treatment groups
Placebo (n = 38) Vitamin B-6 (n = 38)
Mean SD Mean SD
Age (y) 73 3 73 3
IQ 11I 10 109 11
PLP 29 12 31 10
a-EAST 1.83 0.28 1.75 0.12
491
mental state aspects, i.e. vigour, inertia, extraversion/introversion,
general feeling of wellbeing, emotional irritability, and anxiety.
These mental states aspects are subdivided into 14 subscales, i.e.
active, concentrated, inactive, tired, benumbed, extraverted/intro-
verted, self-confident, elated, excited, sensitive, angry, anxious de-
pressed, dreamy. The subject has to agree or disagree with the
adjectives presented on a monitor by pressing a "yes" or "no"
button.
Memory: sensory register: Sperling whole report task.
Thirty slides of
12 random letters (3 x 4) are presented during 250, 150 and 50 ms in
blocks of 10 slides with the same presentation time. Immediately
after each flash the subject has to report as many letters as he can
remember, within an interval of 6 s. The experimenter marked the
reported letters on a response sheet. This procedure establishes
iconic memory (Sperling 1960).
Short term verbal memory: associate learning task.
Nine pairs of a
name and an occupation are displayed on a computer screen at a
constant rate of one pair per 3 s. After these pairs are presented, the
subject is prompted nine times with one of the names and the nine
occupation alternatives. Each time a name is presented, the subject
has to choose one of the nine occupations by pressing the corres-
ponding digit on the keyboard. After each answer the results are
given in the following manner: "yes, John is a grocer", or "no, Susan
is a teacher". Three trials are given in which the subject has to learn
as many paired names and occupations as possible. The number of
correct associations on each trial is stored into the computer. By
means of this recognition procedure short-term verbal learning
(Emmen et al. 1988) is measured.
Long term verbal memory: associate recognition task.
A single recog-
nition trial of the nine names to be matched with one of the nine
occupations used in the associate learning test is administered at the
end of the testing session. The delay of about 1 h with the associate
learning task allows a retention task for long-term memory. The
number of correct responses is recorded.
Long-term memory storage.
The scores on trial 3 of Associate
Learning (AL3) and on Associate Recognition (All) were combined
in a "Forget score", computed as: Forget = AL3-AR. The value of
this difference indicates the amount of information, learned before,
that is not stored in LTM. Thus, the smaller this forget value the
more information is stored in LTM.
Long term visual memory: visual memory task.
A slide of nine pictures
(i.e. chair, glasses, violin) is presented during 15 s. The subject is
instructed to memorize the items. At the end of the test session
(about 1 h later) 15 slides of a picture are presented one at a time.
After each slide the subject has to decide if the picture on the slide
was one of the memorized pictures by pressing "yes" or a "no"
button. There is no time pressure. The next slide is only presented
after the subject has responded to the former one. The number of
right responses and the response time are recorded. By this task the
recognition of visual material in long-term memory is measured.
Attention and performance:form perception: cognitrone.
In the cogni-
trone program five display fields appear on the screen, four on top
and one below. The figure displayed in the field below has to be
compared with the four figures in the top field. If the figure below is
identical to one of the four figures on top, the subject is instructed to
push a "yes" button, if there are no identical figures a "no" button
has to he pressed. The number of correct responses and reaction
time are recorded. This program measures concentration and form
perception (Ktemenjak 1983).
Perceptual-motor skill: Vienna determination unit.
The determination
unit is a complex, multiple-stimuli reaction unit. The unit allows
presentation of optical stimuli coloured white, yellow, red, green and
blue, which are presented on ten different positions. The subject has
to respond by pressing one of five reaction keys assigned to these
colours. Two additional white lamps, which are set apart from the
coloured lamps, require stepping on the left or right pedal. Two
acoustic stimuli (high tone, tow tone) are assigned to two rectangular
"tone" keys. In the time-controlled version used every stimulus is
presented for a certain time and is then followed by the next
stimulus. In the present study three subtests were used, differing in
the speed of presentation. The number of right and wrong reactions,
the number of delayed and missed reactions and reaction time are
recorded. This unit measures perceptual-motor skills, that is eye-
hand/foot coordination.
Measurement of pupil size
Pupil responses were continuously recorded while performing a
short-term memory task and a speed of processing task. Subjects
were tested individually in a dimly lit, sound attenuated room. They
were seated in an adjustable chair, 2.5 m in front of a TV screen, with
their heads fixed in a Whittaker head rest. Values of pupil size were
obtained with a frequency of 50 Hz. At the right side of the subject a
response board was positioned.
Short term memory task.
A memory set of four letters is presented on
the screen during 5 s. Following the memory set a number of 41 trials
is presented containing two or four letters, placed on the corners of
an imaginary square. On a number of trials one letter of the display
is part of the memory set ("target"). The inter-trial interval is
2500 ms. The subjects are instructed to decide if a target is presented
by pressing a "yes" or "no" button as fast as possible. The maximum
response time allowed is I s. Pupil size, reaction times and number of
correct responses are measured. This procedure makes it possible to
measure mental effort while performing a memory task.
Speed of processing task.
A cross is presented during 100 ms random-
ly in one of the corners of the screen. The subject is provided with a
square response board with four buttons at the comers. Each time
an X is presented subjects have to press as fast as possible (within a
period of 1 s after the stimulus) the button that is diagonally
opposite to the cross on the screen. There are 51 trials. The inter-trial
interval varies randomly between 300 and 1500 ms. The reaction
times and number of correct responses are measured in addition to
the pupil responses.
With respect to both tests peak latency and pupillary base level
were established for each trial. Peak latency had to be within 2 s
from stimulus onset for the mental load task and within 2.5 s for the
memory task. Pupillary base level was defined as the mean pupil size
of the inter-trial interval of the mental load task (i.e. 300 1500 ms);
for the memory task base level was the mean pupil size of the last
500 ms of the inter-trial interval. Phasic pupil response was com-
puted by subtracting the mean pupillary base level from the mean
peak level.
Apparatus
Most of tiae tasks were presented by means of the "Vienna Test
System", a computerized testing system which is connected to an
Olivetti M240 Personal Computer. The associated learning task,
associate recognition task, and the pupil registration tasks were
presented by an Olivetti M24 Personal Computer. Pupil diameter
was measured by a Whittaker 1994 - S Eye View Monitor, an
infrared video-pupillometric system.
Test procedure
Subjects were examined individually in a quiet room. The test
procedure as a whole took about 1.5 h. The two test sessions
consisted of different versions of the same test battery. The sequence
of test administration was the same for all subjects, starting with the
mood and performance tasks followed by the pupillometric proced-
ure. The pupil measurements took place in a sound attenuated
room.
492
Compliance
On two random occasions during the study period urine samples
were collected. From these samples 4-pyridoxine acid excretion was
measured. Apart from inspection of these samples, the compliance
could be judged by considering the increase in vitamin B-6 status of
each subject in the treatment group.
Vitamin B-6 status
Blood samples were taken between 9 and 10 a.m. On these occasions
subjects were instructed to eat a light breakfast (for instance, rusk
with jam and coffee or tea) with exclusion of vitamin B-6 containing
products. Written examples of possible breakfasts were given to the
subjects. Vitamin B-6 status was determined by measuring the
pyridoxal-5'-phosphate (PLP) concentration in plasma and the
basal activity of the erythrocyte enzyme aspartate aminotransferase
(EAST). EAST activation (e-EAST) with coenzyme was calculated
by measurement of EAST activity before and after in vitro stimula-
tion with excess PLP (van den Berg 1988). A low response to
stimulation with PLP indicates a high saturation with PLP, which
means a good vitamin B-6 status. Measurement of vitamin B-6
status was performed at the entry screening, within 7 days before the
study commenced and immediately after the study period was
ended.
Medication
The medication consisted of 20 mg vitamin B-6 (pyridoxine" HCL)
or placebo, administered in identical capsules. Subjects were in-
structed to ingest the pills around dinner. This dose of pyridoxine is
I0 times the recommended daily quantity in Holland (i.e.t.5-2 mg
daily). The capsules were given to the subjects on two occasions in
the study period, in envelopes containing blister cards, for a period
of 8 weeks. On receiving the second set of capsules and at the end of
the study subjects were asked to return the remaining capsules.
Data analysis
Data obtained on test sessions 1 and 2 were analysed tbr 76 subjects
(i.e. 38 in each treatment group) by means of univariate and multiv-
ariate analyses of variance and covariance. The scores on the
subscales of the adjective check list appeared not to be normally
distributed, while the distribution of the difference scores (session 2
- session 1) was normal. Hence treatment effects for mood were
studied by analysing difference scores using one-way analyses of
variance with treatment as independent factor.
With respect to the data on vitamin B-6 status and the per-
formance tasks, normal distributions were obtained. Treatment
effects for vitamin B-6 status and cognitive performance were asses-
sed by analysing these data using one-way analyses of covariance
with treatment as independent factor and the pretreatment scores
serving as covariate. If test scores were supposed to be related,
multivariate analyses were performed on clusters of scores.
Finally, some correlation coefficients (Pearson product-moment
correlation) were determined between performance difference scores
and A PLP.
Results
PLP ands-EAST
Vitamin B-6 supplementation increased PLP values in the
vitamin B-6 group significantly compared to the placebo
group [F(1,73)= 364; P < 0.001]. In addition, s-EAST
decreased significantly in the treatment group compared
with placebo t-/7(1,73) = 270; P < 0.001]. After treatment
seven subjects of the placebo group were marginally
deficient on basis of plasma PLP and five subjects of this
group on basis of c~-EAST. The mean PLP and c~-EAST
values before and after treatment are shown in Table 2.
Memo~
Regarding
short-term memory
a multivariate one-way ana-
lysis of covariance for trials 1, 2 and 3 of the Associate
Learning Task with scores on session 1 as covariates did
not show a multivariate difference between groups. How-
ever, stepdown F-tests showed a significant difference for
trial 2 [F(1,69) = 4,09; P < 0.05], indicating a better per-
formance of the vitamin B-6 group on the second trial of
the Associate Learning Task.
In order to measure
long-term memory
a visual and a
verbal memory task were used. The two treatment groups
did not respond differentially to the visual memory task,
either with respect to the response time or to the number
of recognized items. The mean number of recognized
items appeared to be very large, approaching the maxi-
mum score of 15. In addition to a small SD, this finding
suggests that for most subjects this task was quite easy to
perform.
With respect to
verbal long-term memory
a trend was
found on the Associate Recognition Task. One-way ana-
lysis of covariance with the pretreatment scores serving as
covariate revealed that the difference on this task between
the two groups failed to reach significance [F(1,73) -- 2.56;
P = 0.11]. Inspection of the data suggested an interaction
effect between treatment and session. Therefore these data
were reanalysed by means of two-way analysis of variance
(treatment × session) with session as a repeated measures
factor. Indeed, a significant interaction of treatment x
session was found [F(1,74) = 4.45; P < 0.04]. Independ-
ent t-tests revealed that neither the baseline scores of the
two groups [t(74)= 1.36; P = 0.18], nor the post-treat-
ment scores [t(74) = 0.9t; P = 0.36] differed significantly.
Correlated t-tests revealed no significant increase with
vitamin B-6 on session 2 compared to session 1 [t(37)
= 0.20; P = 0.84], but revealed a significant decrease in
performance on session 2 with placebo [t(37)= 2.82; P
= 0.008]. This interaction is depicted in Fig. t.
Table 2. Vitamin B-6 status (PLP and
:~-EAST) on session 1 and 2 of both
treatment groups Placebo
Session 1
Mean SD
PI,P 29 12
~-EAST 1.83 0.28
Vitamin B-6
Session 2 Session 1 Session 2
Mean SD Mean SD Mean SD
29 10 31 10 269 81
1.86 0.26 1.75 0.12 1.35 0.07
4.0
ILl
Z
0
B.
~: 3.5
I-,-
¢j
W
0
u. 3.0
0
m
113
'5
Z
2.5
I I
1 2
SESSION
Fig. 1. Number of correct responses on sessions 1 and 2 for both
treatment groups on the Associate Recognition Task. (Placebo:
session 1, SD = 2.1; session 2, SD = 2.0; vitamin B-6: session 1, SD
= 2.3; session 2, SD = 2.3). (*) Placebo; (A) pyridoxine
One-way analysis of covariance with the pretreatment
scores serving as covariate revealed that the "forget" score
on session 2 was significantly smaller for the vitamin B-6
group [F(1,73)= 5.01; P < 0.03] than for the placebo
group. As a low score means a better storage in LTM, this
result indicates that storage in LTM increased with vi-
tamin B-6. This finding is shown in Fig. 2.
The Pupillometer system only measures pupil values
above 2 ram. Therefore the registration of the pupil re-
sponse failed if subjects had a pupil size less than 2 mm.
Moreover, data from subjects who made too many eye-
blinks were not analysed. Thus, with respect to the (pupil-
lometric) memory task 31 cases in the placebo group and
28 cases in the vitamin B-6 group remained for analysis. In
addition, with respect to the speed of processing task,
scores were available for 49 cases, i.e. 27 cases in the
1.0
0.8
UJ
n-
O
0.6
0
¢t)
w
0.4
0
u.
0.2
0.0
1 2
SESSION
Fig. 2. Mean "forget scores" on sessions 1 and 2 for both treatment
groups. (Placebo: session 1, SD = 2.1; session 2, SD = 2.1; vitamin
B-6: session 1, SD = 1.4; session 2, SD = 1.4). (-~) Placebo; (~)
pyridoxine
493
placebo group and 22 in the vitamin B-6 group. Regarding
these mental effort measurements, the number of correct
responses and the reaction times on the speed of pro-
cessing task and the memory task were the same for both
groups. The pupil registrations during these tasks with
respect to peak level or peak latency were the same for
both groups. In addition, the pupillary base level was not
different in the two treatment groups. Peak level, peak
latency and baselevel values are shown in Tables 3 and 4.
Relations between A PLP and mental improvement
In order to investigate the possible linear relation between
the difference scores (session 2- session 1) of cognitive
tasks and A (supplementation level-baseline) PLP several
correlation coefficients were determined for data of the
vitamin B-6 suppleted group.
The lowest A PLP in this group appeared to be 100,
the highest value 412. Correlation coefficients were deter-
mined for the whole range of A PLP between 100 and 412.
In addition, separate correlation coefficients were deter-
mined for the ranges < 200, < 300, 200-300 and > 300
A PLP. These ranges were chosen for the practical pur-
pose to subdivide the whole range in about equal A PLP
ranges.
With respect to associate recall, no significant correla-
tion was found between the difference score and A PLP
100-412. For the other ranges, significant correlations
between the difference score and A PLP 100 200 (r
=0.66; P<0.02) and APLP 200-300 (r=0.65; P
= 0.001) were found. Correlations for the remaining
A PLP ranges were not significant. For the A PLP range
100-300 the correlation coefficient decreased markedly
and failed to reach significance (r = 0.29; P < 0.07).
The correlation for the difference score of the other
memory parameter, the "forget" score, and A PLP
100-200 appeared to be significant (r = - 0.65;
P < 0.03). In addition, for the range A PLP 200-300 a
Table 3. Pupillary response (mm), base level (mm) and peak latency
(ms) in the processing task
Placebo Vitamin B-6
Session 1 Session 2 Session 1 Session 2
Pupil response 0.62 0.57 0.69 0.65
Base level 3.57 3.83 3.76 3.95
Peak latency 996 1037 984 1027
Table 4. Pupillary response (mm), base level (mm) and peak latency
(ms) in the memory task
Placebo Vitamin B-6
Session 1 Session 2 Session 1 Session 2
Pupil response 0.28 0.29 0.25 0.25
Base level 4.09 4.47 4.27 4.47
Peak latency 985 790 979 854
494
significant correlation (r = - 0.54; P = 0.008) was shown.
Finally, for the range 100-300 A PLP the correlation
decreased but was still significant (r = - 0.31; P < 0.05).
Also with respect to this memory parameter, no
significant correlation with the whole A PLP 100 412
range nor with A PLP > 300 was found.
For two other difference scores the correlation coeffi-
cient was also determined, i.e. the subscale "depressed" of
the adjective checklist and the performance time on the
cognitrone task. These measures were chosen because of
the frequently reported relation between vitamin B-6 and
depression and the impression that the eognitrone task is
very sensitive in detecting perceptual changes.
For both measures highly significant correlations were
found with A PLP 200-300. The correlation of "de-
pressed" with A PLP was -0.61 (P = 0.003) and the
performance time on the cognitrone also correlated signi-
ficantly (r = - 0.62, P = 0.002) with A PLP.
The correlation coefficients and p-values are shown in
Table 5.
Finally, baseline PLP values were correlated with
PLP values on session 2 and with A PLP. Baseline PLP
correlated with PLP after treatment (r = 0.42, P = 0.004)
and with A PLP (r = 0.52, P < 0.001). This means that
PLP increase is greater for subjects with a higher baseline
PLP value, suggesting a different vitamin B-6 metabolism
in subjects with a low or a high baseline PLP concentra-
tion.
Discussion
Significant effects of vitamin B-6 supplementation on
memory were found, especially with respect to long-term
memory.
Iconic memory is supposed to be mediated by the
retina, and is thus peripherally controlled (Sakitt 1976).
The amount of vitamin B-6 is very high in the retina,
where it plays an important role in the synthesis of
neurotransmitters (Hammad and Ebadi 1984). If age-
related impairment of iconic memory (Cerella et al. 1982)
is caused by a reduction of vitamin B-6, supplementation
might be beneficial for the visual sensory register. How-
ever, additional supply of vitamin B-6 did not improve
iconic memory, indicating the absence of a relation be-
tween vitamin B-6 status and iconic memory or the fact
that supplementation does not increase the amount of
vitamin B-6 in the retina.
The present study also failed to show short-term
memory improvement by vitamin B-6 supplementation.
Maybe it is not possible to attain improvement of short-
term memory, because this kind of memory is hardly
influenced by the aging process (Craik 1968; Drachman
and Leavitt 1972; Botwinick and Storandt I974; Smith
1975; Poon and Fozard 1980).
As stated in the introduction, the component of mem-
ory that seems to be most vulnerable in the aging process
is long-term memory. The present results indicate that the
treatment group had a better verbal long-term memory on
session 2 as compared to placebo. Recognition of pre-
viously learned information appeared to be unchanged in
the treatment group, whereas the placebo group showed a
performance decrement. This reduced recognition may
reflect an age-related decline occurring within the treat-
ment period.
A composite score which relates the learned informa-
tion to its recognition afterwards shows a more pro-
nounced difference between both groups. Although the
amount of learned information (trial 3 of Associate Learn-
ing) is the same for both groups, the degree of recognition
afterwards is significantly different. This composite score,
as depicted in Fig. 2, is a measure of the amount of
information not stored in long-term memory. The post-
treatment composite score appeared to be significant
lower for the treatment group than for placebo. Moreover,
the post-treatment score was found to be zero with vi-
tamin B-6, indicating that all previously learned informa-
tion had been stored into long-term memory. The conclu-
sion seems legitimate that treatment influences storage
instead of retrieval, because a recognition procedure is
used. A decrement in recognition primarily indicates that
encoding was suboptimal.
Because post-treatment memory differences are partly
explained by a fall in performance within the placebo
group, the findings implicate that vitamin B-6 supple-
mentation inhibits age-related reduction in the storage of
verbal information in long-term memory. The fact that
this is not the case with respect to visual information may
be explained by the low degree of difficulty of the task
used.
The correlations between mental performance and
PLP values within the treatment group further support
these findings. Quite high positive (Associate Recognition)
and negative correlations (forget-score) were found be-
tween memory (difference) scores (session 2 - session 1)
and A PLP values (see Table 5). The most significant
correlations are found for a h PLP range between 200 and
300. This means that within the treatment group a clear
linear relation exists between memory and PLP increase.
In addition, a similar relation appears to exist with respect
Table 5. Correlation coefficients between
cognitive tasks (difference scores) and
vitamin B-6 status (A PLP) in the vitamin
B-6 group
A PLP 1130-200 200-300 > 300 100-300 100-412
A Associate recall 0.66* 0.65*** - 0.21 0.29 0.18
(P = 0.06)
A Forget score - 0.65* - 0.54** - 0.20 - 0.31" - 0.10
A Depressed - 0.58* - 0.61"* 0.23 0.32* - 0.29*
A Cognitron time 0.48 - 0.62** 0.50 - 0.09 0.05
(P = 0.08) (P = 0.08)
*P _< 0.05; **P _< 0.01; ***P _< 0.001
495
to "depressed" and "perceptual speed", especially for
A PLP 200 300. This relationship appears to become
opposite in three cases for the high A PLP range ( > 300).
The finding that high correlations in particular exist for
A PLP 200-300 seem to indicate that an "optimal" A PLP
range exists with respect to behavioral effects. This is in
line with studies on drug effects on learning and memory
which found U-shaped dose-response curves (Martinez
and Rigter 1980). Although these U-shaped functions are
not fully understood yet, receptor tachyphylaxis or fatigue
(Day 1979) and the opposite action of drugs on different
kinds of receptors (Mueller et al. 1982) have been pro-
posed as explanatory factors.
In agreement with this idea of U-shaped functions, two
trials, one in students and one in obese patients, found
that high doses of vitamin B-6 impaired memorization
(Molimard et al. 1980). This reduction in the memoriz-
ation process was explained as being the consequence of a
GABA-ergic effect of vitamin B-6.
The finding that a relationship exists between PLP
values and memory between treatment groups as well as
within the vitamin-supplemented group seems to justify
the conclusion that vitamin B-6 supplementation may
reduce age-related memory decline. The consistency of
related observations in the data makes it unlikely that
these effects on memory resulted from sheer chance. How-
ever, clinical importance is doubtful because the observed
memory changes are very small. Perhaps more impressive
effects are to be expected if an optimal dose of vitamin B-6
is given, thereby limiting the PLP range. The observed
negative correlations between "depressed" and "percep-
tual speed" with A PLP might indicate that an optimal
dose may also improve behavioral measures other than
memory. For the present the general conclusion is drawn
that vitamin B-6 supplementation affects storage of in-
formation in long-term memory. The data give rise to the
assumption that effects are to be expected within a certain
range of vitamin B-6 status that must not exceed a certain
upper limit. This assumption needs to be verified by
further research.
References
Baker H, Frank O, Thind IS, Jaslow SP, Loufia DB (1979) Vitamin
profiles in elderly persons living at home or in nursing homes,
versus profile in healthy young subjects. J Am Geriatr Soc
27 : 444 450
Beatty J (1982) Task-evoked pupillary responses, processing load
and the structure of processing resources. Psychol Bull
9112] : 276-292
Beatty J, Wagoner BL (t978) Pupillometric signs of brain activation
vary with level of cognitive processing. Science 199:t216 1218
Berg H van den (1988) Assessment of the B-vitamin status some
biochemcial and clinical aspects. In: Clinical chemistry--an over-
view. Plenum Press, New York, pp 545-556
Botwinick J, Storandt M (1974) Memory related functions and age.
Charles C. Thomas, Springfield Ill
Cerella J Poon LW, Fozard JL (1982) Age and iconic read-out. J
Gerontol 37:197 202
Chom6 J, Paul T, Pudel V, Bleyl H, Heseker H, H/ippe R, Kfibler W
(1986) Effects of suboptimal vitamin status on behavior. Bibl
Nutr Dicta 38:94-103
Craik FIM (1968) Two components in free recall. J Verb Learn Verb
Behav 7:996 1004
Day MD (1979) Autonomic pharmacology. Experimental and clini-
cal aspects. Churchill Livingston, New York
Drachman DA, Leavitt J (t972) Memory impairment in the aged.
Storage versus retrieval deficit. J Exp Psychol 93:302 308
Emmen HH, Hogendijk EMG, Hooisma J, Orlebeke JF, Uijtde-
haage SHJ (1988) Adaptation of two standardized international
test batteries for use in the Netherlands for detection of exposure
to neurotoxic compounds. Internal report. Rijswijk: Medisch
Biologisch Laboratorium TNO
Gilbert JG, Levee RF (1971) Patterns of declining memory. J
Gerontol 26 : 70 75
Goodwin JS, Goodwin JM, Garry PJ (1983) Association between
nutritional status and cognitive functioning in a healthy elderly
population. JAMA 249:2917-2921
Guilland JC, Bereksi-Reguig B, Lequeu B, Moreau D, Klepping
J (1984) Evaluation of pyridoxine status among aged institu-
tionalised people. Int J Vitam Nutr Res 54:185-193
Hammad HM, Ebadi M (1984) The synthesis and subcellular dis-
tribution of pyridoxal phosphate in rat and bovine retinas.
Neurochem Int 6[4] : 569-572
Heseker H, Kfibler W (1990) Psychische Ver/inderungen als Friihzei-
ehen einer suboptimalen Vitaminversorgung. Ern/ihrungs-Umsc-
hau 37 : 87-94
Janke W, Debus G (1978) List of adjectives. The PC/S Vienna Test
System. Dr. G. Schuhfried, M6dling, Austria
Kleijnen J, Knipschild P (1991) Niacin and vitamin B6 in mental
functioning: a review of controlled trials in humans. Biol Psy-
cfiiatry 29 : 931-941
Klemenjak W (1983) Das Wiener Konzentrationsger/it (Cognitrone).
Neue Methoden der apparativen Erfassung kognitieven Leistun-
gen. Kfiratorium ffir Verkehrssicherheit, Wien
L6wik MRH, Berg H van den, Westenbrink S, Wedel M, Schrijver,
J, Ockhuizen Th (1989) Dose-response relationships regarding
vitamin B-6 in elderly people: a nationwide nutritional survey.
Am J Clin Nutr 50:391-399
Martinez Jr JL, Rigter H (t980) Endorphins alter acquisition and
consolidation of an inhibitory avoidance response in rats. Neuro-
sci Lett 18 : 197-201
Merrill AH, Henderson JM (1987) Diseases associated with defects
in vitamin B6 metabolism or utilisation. Annu Rev Nutr
7:137-156
Molimard R, Marillaud A, Paille A, LeDevehat C, Lemoine A,
Dougny M (1980). Impairment of memorization by high doses of
pyridoxine in man. Biomedicine 32 : 88-92
Mueller AL, Palmer MR, Hoffer BJ, Dunwiddie TV (1982)
Hippocampal noradrenergic responses in vivo and in vitro
characterization of alpha and beta components. Naunyn-
Schmiedeberg's Arch Pharmacol 318:259-266
Payne DT, Parry ME, Harasymiw SJ (1968) Percentage of pupillary
dilation as a measure of item difficulty. Percept Psychophys
4:139-143
Pearler WS (1974) Pupil size, information overload and per-
formance differences. Psychophysiology 11 : 559-566
Poon LW (1985) Differences in human memory with aging: nature,
causes, and clinical implications. In: Birren JE, Schaie KW (eds)
Handbook of the psychology of aging. Van Nostrand Reinhold,
New York, pp 427-462
Poon LW, Fozard JL (1980) Age and word frequency effects in
continuous recognition memory. J Gerontol 35:77-86
Rose CS, Gy6rgy P, Butler M, Andres R, Norris AH, Shock NW,
Tobin J, Brin M, Spiegel H (1976) Age differences in vitamin B-6
status of 617 men. Am J Clin Nutr 29:847 853
Sakitt B (1976) Iconic memory. Psychol Rev 83 : 257-276
Schrijver J, Veelen BWC van, Schreurs WHP (1985) Biochemical
evaluation of the vitamin and iron status of an apparently
healthy Dutch free-living elderly population. Int J Vitam Nutr
Res 55 : 337 349
Schrijver J, Tolonen M, Westermarck F, Halme M, Tuominen S,
Frilander A, Sarna S (1987) Vitamin B-6 status of Finnish and
496
Dutch elderly. In: Korpela T, Christen P (eds) Biochemistry of
vitamin B-6. Birkh/iuser, Basel, pp 411 416
Smith AD (1975) Aging and interference with memory. J Gerontol
30: 319-325
Sperling G (1960) The information available in brief visual
presentation. Psychol Monogr 74 :
11
Tolonen M, Schrijver J, Westermarck T, Halme M, Tuominen SEJ,
Frilander A, Keinonen M, Sarna S (1988) Vitamin B6 status of
Finnish elderly. Comparison with Dutch younger adults and
elderly. The effect of supplementation. Int J Vitam Nutr Res
58 : 73-77
Welford AT (1985) Changes of performance with age: an overview.
In: Charness N (ed) Aging and human performance. Wiley, New
York pp 333 369
Zeeuw J de (1971) Algemene psychodiagnostiek. Swets and Zeitlin-
ger, Lisse, p 133
