Content uploaded by Paula J Durlach
Author content
All content in this area was uploaded by Paula J Durlach
Content may be subject to copyright.
Abstract Rationale: Previous investigations have dem-
onstrated increased performance after the administration
of a glucose-load on certain aspects of cognitive func-
tioning in healthy young adults. Generally these studies
have used a procedure where participants were tested in
the morning after an overnight fast. Objective: The aim
of the present study was, for the first time, to investigate
the glucose cognitive facilitation effect under more natu-
ral testing times and with shorter duration of the previ-
ous fast. Methods: Measures of verbal and non-verbal
memory performance were compared under different
fasting intervals (2-h fast versus overnight fast), times
(morning versus afternoon) and glycaemic conditions
(glucose versus aspartame drinks) in healthy young par-
ticipants. Results: There was a significant glucose facili-
tation effect on long-term verbal memory performance.
In addition, glucose significantly enhanced long-term
spatial memory performance. The effect of glucose was
essentially equivalent whether it was given after an over-
night fast or a 2-h fast following breakfast or lunch.
There was no effect of drink and time of day on working
memory performance. Conclusions: The results of this
study further support the hypothesis that glucose admin-
istration can enhance certain aspects of memory perfor-
mance in healthy young adults. More significantly, the
findings indicate that this cognitive facilitation effect
persists under more naturalistic conditions of glucose ad-
ministration and is not restricted to long fast durations or
morning administration.
Keywords Glucose · Short-term memory · Long-term
memory · Verbal memory · Non-verbal memory
Introduction
There is considerable evidence that modest increases in
circulating blood glucose concentrations regulate several
brain functions, including particular enhancement of
learning and memory in rodents and humans. The exami-
nation of glucose effects on cognitive functions did not
grow from a specific interest in nutrition, but rather from
attempts to understand the mechanisms by which hor-
mones, released when animals are trained in a memory
task, regulate memory formation. Animal studies have
indicated that adrenaline enhances consolidation of a
memory (see, for example, Gold et al. 1986; McGaugh
1989; Gold 1992). As peripheral adrenaline is largely ex-
cluded from the central nervous system, it is likely that
adrenaline does not improve memory performance by di-
rectly stimulating brain mechanisms. Instead, a number
of findings suggest that increases in blood glucose levels
subsequent to peripheral adrenaline actions contribute to
enhancement of memory storage processing (Wenk
1989; Gold 1991, 1992; White 1991).
Glucose is the major source of energy for the brain
and is essential for the normal functioning of the central
nervous system (Sieber and Trastman 1992). Neurones
share the biochemical machinery of all other living cells,
including the necessity to liberate chemical energy
through the oxidation of glucose. In addition, glucose is
necessary as the substrate for a wide variety of simple
and complex molecules. The brain uses a large quantity
of glucose for its size and volume. Although other tis-
sues can utilise glucose, it is not the main metabolite of
most organs; for example, heart, renal cortex, and even
the liver derive most of their energy from fatty acids.
During hypoglycaemia, these other tissues stop meta-
S.I. Sünram-Lea (
✉
)
Department of Psychology, University of Manchester,
Oxford Road, Manchester M13 9PL, UK
e-mail: sunram@fs4.psy.man.ac.uk
Tel.: +44-1524-593834, Fax: +44-1524-59344
J.K. Foster
Department of Psychology, University of Western Australia, Perth,
Western Australia 6907, Australia
P. Durlach · C. Perez
Unilever Research Colworth Laboratory, Sharnbrook,
Bedford MK44 1LQ, UK
S.I. Sünram-Lea
Department of Psychology, Fylde College,
University of Lancaster, Lancaster LA1 4YW, UK
Psychopharmacology (2001) 157:46–54
DOI 10.1007/s002130100771
ORIGINAL INVESTIGATION
Sandra I. Sünram-Lea · Jonathan K. Foster
Paula Durlach · Catalina Perez
Glucose facilitation of cognitive performance in healthy young adults:
examination of the influence of fast-duration,
time of day and pre-consumption plasma glucose levels
Received: 23 August 2000 / Accepted: 14 March 2001 / Published online: 8 June 2001
© Springer-Verlag 2001
47
bolising glucose in order to increase its availability to the
brain (Thompson 1967). Since relatively little glucose
can be stored, the brain is reliant on a continuous supply
of glucose as its primary fuel, delivered from the blood-
stream via a regulatory carrier mechanism (Wenk 1989).
A decline in blood glucose levels to hypoglycaemic lev-
els has a profound and rapid impact on normal brain
functioning; including disrupted neuronal activity, altera-
tions in cortical cellular functions and changes in EEG
activity (Thompson 1967; Holmes et al. 1983).
Although the interrelationship between normal brain
functioning and glucose metabolism has been well estab-
lished for many years, it was not until relatively recently
that the neural effects of everyday variations in blood
glucose levels were investigated. Our understanding
about the role of glucose in cognitive functioning has
been derived mainly from animal studies and from re-
search into human memory deficits (see Messier and
Gagnon 1996 for a recent review). Previous research has
demonstrated that glucose administration can enhance
memory in rats (see, for example, Gold 1986; Lee et al.
1988; Winocur 1995; Kopf and Baratti 1996). Moreover,
glucose administration and/or regulation can enhance
learning and memory in healthy aged humans (see, for
example, Gonder-Frederick et al. 1987; Hall et al. 1989;
Craft et al. 1994; Messier et al. 1997) and may improve
several cognitive functions in subjects with severe cogni-
tive pathologies, including individuals with Alzheimer’s
disease (Craft et al. 1992; Manning et al. 1993) and
Down’s syndrome (Manning et al. 1998). Facilitation of
cognitive performance induced by elevations in plasma
glucose and/or insulin levels has also been reported
in patients with schizophrenia (Fucetola et al. 1999;
Newcomer et al. 1999). Furthermore, impairments in
certain cognitive skills have been recognised as a possi-
ble complication of long-standing non-insulin and insu-
lin dependent diabetes mellitus, and evidence exists that
improved glycaemic control can ameliorate performance
on selective areas of cognition in these populations
(Gradman et al. 1993; Meneilly et al. 1993).
The administration of glucose can also facilitate cog-
nitive functioning in healthy young adults under certain
laboratory conditions (Hall et al. 1989; Benton and
Owens 1993; Benton et al. 1993; Craft et al. 1994;
Owens and Benton 1994; Parker and Benton 1995;
Foster et al. 1998; Martin and Benton 1999; Messier et
al. 1998; Messier et al. 1999; Kennedy and Scholey
2000; Metzger 2000). Generally, these studies have used
a procedure where participants were tested in the morn-
ing after an overnight fast (e.g. Hall et al. 1989; Foster et
al. 1998), following a standardised breakfast (Azari
1991) or no dietary restrictions were implemented (e.g.
Benton et al. 1993). So far, only one study specifically
investigated possible differences in the degree of cogni-
tive facilitation after glucose ingestion between a group
of young adult females who had or had not eaten break-
fast (Martin and Benton 1999). This study demonstrated
that glucose administration improved working memory
performance after an overnight fast. However, glucose
did not influence performance of participants who had
eaten breakfast. Furthermore, in those who had fasted,
glucose ingestion resulted in memory performance com-
parable to those who had consumed breakfast.
In the present study, we decided to investigate further
the significance of an overnight fast on the facilitative ef-
fect of glucose, by comparing cognitive performance on a
range of memory tasks between participants who either
fasted overnight or who were administered a standardised
breakfast 2 h before testing commenced. Given that all
previously reported studies in the literature were conduct-
ed in the morning, we decided not only to manipulate the
fast-duration (overnight versus 2-h fast), but also to carry
out cognitive test sessions at different times of the day
(morning versus afternoon). Given that people may con-
sume a glucose-containing product as a morning and/or
afternoon snack after a period of post-breakfast/post-
lunch fasting, the aim of this study was to increase the
generalisability of the previously observed glucose facili-
tation effect (Foster et al. 1998) and to establish its eco-
logical validity. It was intended that this design would en-
able us to investigate further the impact of these parame-
ters on the glucose facilitation effect. For this reason,
cognitive performance was assessed using the same pro-
tocol that has previously proved sensitive to glucose after
overnight fasting (Foster et al. 1998). The effect of glu-
cose under each of the meal conditions was compared
with the effect of a sweetness-matched aspartame drink.
Materials and methods
Participants
Sixty participants with a mean age of 21 years (range 18–28) were
randomly assigned to one of six conditions. None of the partici-
pants was diabetic, obese, or underweight (mean body mass index
was 21.87). Prior to the start of the study, the Ethics Committee of
the Department of Psychology, University of Manchester ap-
proved the experimental procedure.
Memory tests
i) Modified version of CVLT (Californian Verbal Learning Test):
measuring immediate, short delay and long delay long-term
memory (free recall, cued recall, and recognition) for a supra-
span word list (Delis et al. 1987). A (previously tape-recorded)
list of 24 words was presented. The original word list had been
modified to suit the needs of the experiment in order to avoid
possible ceiling effects. Thus, the word list had been extended
from 16 words to 24 words, each new word being chosen to fit
to the design of the original CVLT. In each trial the list was
read at a word frequency of one word every 2.5 s. A complex
hand motor sequence was performed at the same time as the
list was presented to participants, in order to reduce the possi-
bility of performing at or near ceiling levels on the memory
task (inasmuch as carrying out an additional task divides the
participant’s allocation of cognitive resources between the
hand sequence and the memory tasks).
ii) Rey-Osterrieth complex figure drawing: used to evaluate long-
term memory for non-verbal materials (Osterrieth 1944).
iii) Modified digit span: used to evaluate working memory span
(Wechsler 1987) (for a more detailed account of the tests used
see Foster et al. 1998).
Treatment
Participants received one of the following: Rusco Pharmaceuticals
pure powdered glucose (25 g) or Boots aspartame tablets (5 tab-
lets).
Glucose was used in doses of 25 g, as a previous study in our
laboratory observed this to be the most effective glucose dosage for
memory enhancement (Foster et al. 1998). Five tablets of aspartame
were used since, when dissolved in 300 ml of water, the resulting
sweetness was rated as equivalent to that of the glucose solution.
Standardised meal
i) One bagel with cream cheese (one tablespoon)
ii) One yogurt
iii) One glass of orange juice (170 ml)
Design
A between-participants 2×3 design (ten participants per cell) was
employed in order to investigate the effect of drink administration
Drink (glucose, aspartame) on memory performance under the dif-
ferent meal conditions Cond (fasting, breakfast, lunch), to which
participants were randomly assigned. The administration of the
drinks (glucose, aspartame) followed a double-blind procedure.
Procedure
Each participant attended one test session that lasted approximate-
ly 60 min.
●
Condition “Fasting”. Participants were informed that they
should not eat or drink (except water) from midnight prior to
testing. Testing was carried out between 0900 and 1200 hours
on the following day.
●
Condition “Breakfast”. Participants were informed that they
should not eat or drink (except water) from midnight prior to
testing. The next morning, participants were given a standardi-
sed breakfast of a bagel with cream cheese, yogurt and a glass
of orange juice by the experimenter. Testing was carried out
2 h later. During the delay between receiving the meal and test-
ing, participants were instructed to stay in the test room. Test
sessions took place between 0900 and 1200 hours.
●
Condition “Lunch”. Participants were instructed to have their
normal breakfast but not to eat anything 3 h before arriving in
the laboratory. They were then given the standardised meal
(see Breakfast meal condition). Testing was carried out 2 h
later, during which time participants were instructed to stay in
the room. Test sessions were carried out between 1400 and
1700 hours.
All participants were informed that they would undergo cognitive
testing relating to human memory performance, and that they were
required to drink a non-harmful, non-intoxicating liquid. Partici-
pants were asked to give information about their age, weight, and
height and whether they were taking any medication. At the begin-
ning of each session (before the administration of the drink), base-
line glucose levels were measured. All participants agreed to have
their blood glucose levels monitored, and they were ensured that
they were permitted to withdraw from the experiment without
prejudice if they were not willing to have small samples of blood
taken during the experiment. Blood glucose readings were ob-
tained using the ExacTech blood glucose monitoring equipment
(supplied by MediSense Britain Ltd, 16/17 The Courtyard, Gorsey
Lane, Coleshill, Birmingham B46 1JA, UK), following the recom-
mended procedure.
Immediately after the first blood glucose reading was obtained,
participants received one of the two drinks (glucose and aspar-
tame). These drinks were randomly allocated to participants as
they entered the laboratory. The administration of the glucose and
48
the aspartame drink followed a double-blind procedure. Partici-
pants were instructed to consume the drink within a period of
10 min. This was directly followed by the administration of the
first cognitive test. Twenty-five minutes post-consumption, the
second blood glucose reading was carried out. The final blood
glucose measurement was taken approximately 45 min post-con-
sumption (for sequence of testing, see Fig. 1).
Statistical analyses
Blood glucose values were examined using a three-way analysis
of variance (ANOVA), with repeated measures on one factor (time
of blood sampling). The three different factors were Cond (Fast-
ing, Breakfast, Lunch)×Drink (aspartame versus glucose)×Time
(time when blood glucose was measured), followed by Tukey
honest significant difference (HSD) testing.
The results of all cognitive tests were analysed using a two-way
between subjects ANOVA, with factors of Cond (Fasting, Breakfast,
Lunch) and Drink (aspartame versus glucose). Significant effects
revealed by ANOVA were further analysed using the HSD test.
Covariate analyses were also performed in order to control
for immediate memory performance on group differences on
other parts of the CVLT. This was done by conducting further
ANCOVAs on memory effects that were significant in the previ-
ously conducted ANOVAs (using immediate free recall perfor-
mance on the CVLT as a covariate).
Correlations were also carried out in order to compare perfor-
mance on various subtests of the CVLT against the glycaemic re-
sponse (i.e. correlating subtest memory performance with those
blood glucose values closest to the time of administration of that
subtest).
Fig. 1 Sequence of testing
49
Results
Analysis of blood glucose levels (BGLs)
This analysis showed that there was a significant effect
of Cond [F(2,54)=4.54, P<0.05]. Post hoc comparison
showed that blood glucose levels are significantly higher
under Cond B (Breakfast) and Cond C (Lunch) than
under Cond A (Fasting), when examining averaged
glycaemic responses (Cond B versus Cond A: P<0.05;
Cond C versus Cond A: P<0.05).
Additionally, there was a significant effect of Drink
(collapsed across all Conds) [F(2,54)=14.69, P<0.001].
Post hoc comparisons showed that, as expected, and
analogous to our previous experiments, participants in-
gesting the glucose drink displayed significantly higher
blood glucose levels than the aspartame group.
The main effect of Time also reached significance
[F(2,108)=56.56; P<0.0001]. Post hoc comparison of
means revealed that the averaged glycaemic response
taken at t25 and t45 differed significantly from baseline
blood glucose levels (t25 versus t0: P<0.001; t45versus
t0: P<0.001). The two-way interaction of Cond and
Time also reached significance [F(4,108)=3.43; P=0.01],
indicating that, as expected, the three conditions dis-
played different trajectories of glycaemic response.
These observations were confirmed by post hoc testing,
which showed that baseline blood glucose levels started
off lower after an overnight fast and continued to rise
throughout the experiment, whereas after a 2-h fast
blood glucose levels were generally higher but did not
continue to rise from t25 to t45. In addition, as expect-
ed, there was a significant two-way interaction between
Drink and Time [F(2,108)=32.63; P<0.001], which indi-
cated that blood glucose levels followed a different tra-
jectory under two different drinks (aspartame versus
glucose). Again, this observation was confirmed by post
hoc testing, which showed that blood glucose levels did
not differ significantly from baseline levels after the in-
gestion of an aspartame drink. However, after glucose
administration there was a significant rise of blood glu-
cose levels from baseline to t25 (see Fig. 2 for mean
BGLs).
Cognitive performance
California Verbal Learning Test (CVLT)
Immediate free recall list A (IFRCa). A three-way
ANOVA (Cond, three levels; Drink, two levels; Trial,
five levels) revealed that there was a significant trial ef-
fect upon performance on this task [F(4,216)=253.55;
P<0.0001]. Further analysis showed that, as anticipated,
there was a steady and significant increase in the number
of words that were remembered in the course of the five
trials (all P<0.0001). Moreover, there was a significant
effect of Drink [F(1,54)=63.47; P<0.0001] on immediate
free recall performance. Post hoc comparison showed
that participants receiving the glucose drink performed
significantly better than the aspartame group. However,
the main effect of Cond failed to reach significance at
this stage of recall [F(2,54)=2.88; P=0.06; NS]. There
was no interaction between Cond and Drink [F(2,54)=
0.40; P=0.67 NS], nor did the factors Cond and Trial
interact [F(8,216)=1.08; P=0.38 NS]. The interaction
of Drink and Trial did reach significance, however
[F(4,216)=8.67: P<0.00001]. Post hoc testing indicated
that not only does performance on the free recall task get
better over the course of the trials, but after the ingestion
of a glucose drink a faster learning trajectory is observ-
able.
Fig. 2 Graph illustrating
glycaemic response across con-
ditions over the course of the
experiment
50
Immediate free recall list B (IFRCb). There was a signif-
icant Drink effect on performance of list B [F(1,54)=
9.19; P<0.01], with the glucose group outperforming the
aspartame group. There was, however, no significant
effect of Cond on word recall performance of the inter-
ference list [F(2,54)=1.61, P=0.21 NS], and the factors
of Drink and Cond did not significantly interact with
each other [F(2,54)=1.57, P=0.22 NS] (see Fig. 3).
Short delay free recall (SDFRC). There was a significant
effect of Drink [F(1,54)=114.98; P<0.0001] and of Cond
[F(2,54)=6.47; P<0.01] on performance of this part of the
CVLT. Further analysis revealed that the glucose groups
remembered significantly more than those receiving the
control substance (aspartame), and that the Breakfast con-
dition (B) outperformed the Lunch (C) and the Fasting
condition (A) (condition B versus condition A: P<0.001;
condition B versus condition C: P<0.001). There was no
significant interaction between the factors of Cond and
Drink [F(2,54)=0.59; P=0.56 NS] (see Fig. 3).
The ANCOVA, in which immediate free recall perfor-
mance (total score) was used as a covariate, revealed
that both main effects on short delay performance were
preserved when immediate memory recall was taken
into account (i.e. main effect of Cond: [F(2,53)=
3.34; P<0.05], main effect of Drink: [F(1,53)=23.56;
P<0.0001]).
Short delay cued recall (SDCR). There was a significant
glucose facilitation effect on performance on this task
[F(1,54)=129.04; P<0.0001]. Again, the glucose groups
performed significantly better than the aspartame groups.
The effect of Cond failed to reach significance
[F(2,54)=3.13; P=0.05 NS], and there was no significant
interaction between the factors Cond and Drink
[F(2,54)=1.58; P=0.22 NS] (see Fig. 3).
Long delay free recall (LDFRC). There was a significant
effect of Drink on performance of the long delay free re-
call component of the CVLT [F(1,54)=99.12; P<0.0001].
Comparison of the means revealed that participants re-
ceiving the glucose drink performed significantly better
than those receiving the aspartame drink. There was no
significant effect of Cond on long delay free recall per-
formance [F(2,54)=2.60; P=0.083 NS], and no signifi-
cant interaction between Drink and Cond [F(2,54)=1.95;
P=0.15 NS] (see Fig. 3).
The ANCOVA showed that the Drink effect on
long delay performance was sustained when immediate
memory recall was taken into account [F(1,53)=17.03;
P<0.0001].
Long delay cued recall (LDCR). There was a significant
Drink effect on long delay cued recall performance
[F(1,54)=106.93; P<0.0001]. Again, after the ingestion
of the glucose drink performance was significantly better
than after the administration of the control drink. In
addition, there was a significant effect of Cond [F(2,54)=
4.15; P<0.05], with condition B (Breakfast) outperform-
ing condition C (Lunch) (P<0.05). There was no signifi-
cant interaction between the factors of Drink and Cond
[F(2,54)=2.02; P=0.14 NS] (see Fig. 3).
Long delay recognition (LDRecog). In terms of percent-
age correct recognition, there was a significant effect of
Drink on this subtask of the CVLT [F(1,54)=47.21;
P<0.0001]. Subsequent comparison of the means showed
that the glucose group performed significantly better
than the aspartame group. There was no effect of
Cond on percentage correct recognition on this task
[F(2,54)=1.15; P=0.32 NS], nor did the two factors
interact with each other significantly [F(2,54)=1.45;
P=0.24 NS] (see Table 1).
Fig. 3 Graph illustrating CVLT
performance under the two dif-
ferent glycaemic conditions.
The tasks are: IFRCa immedi-
ate free recall list A, IFRCb im-
mediate free recall list B (inter-
ference list), SDFR short delay
free recall, SDCR short delay
cued recall, LDFR long delay
free recall, LDCR long delay
cued recall
51
The use of a signal detection parameter showed that
there were significant differences in the amount of di-
scriminability, depending on Drink [F(1,54)=57.11;
P<0.0001] and Cond [F(2,54)=7.66; P<0.01] (see
Table 1). The glucose group outperformed the aspartame
group and those performing the test in the morning (con-
dition A, condition B) did significantly better than those
who were tested after lunch (condition C) (Fasting ver-
sus Lunch: P<0.03; Breakfast versus Lunch: P<0.001).
The interaction of the factors of Drink and Cond did not
reach significance, however [F(2,54)=2.85; P=0.07 NS].
There were no significant differences in response bias
shown under each Cond [F(2,54)=0.75, P=0.48 NS] or
Drink [F(1,54)=3.93; P=0.05 NS] on this task.
Rey-Osterrieth complex figure
There was a significant effect of Drink on reproduction of
the Rey-Osterrieth figure [F(1,54)=8.96; P<0.005], with
the glucose group performing significantly better on this
task than the aspartame group. However, no significant
differences in figure reproduction scores emerged be-
tween the three conditions [F(2,54)=1.48; P=0.24 NS].
Moreover, no significant interaction of Cond and Drink
was observable [F(2,54)=0.14; P=0.87 NS] (see Table 1).
Digit span
There was no significant effect of either Cond [F(2,52)=
0.57; P=0.57 NS] or Drink [F(1,54)=0.16; P=0.69] on
performance of this task (see Table 1).
Correlations
Short and long delay recall
Analysis of the Pearson’s product moment correlation
coefficient (one-tailed) across all participants (all drinks,
all conditions) showed that performance on short delay
free and cued recall correlated significantly with blood
glucose levels at t25 (r=0.45; P<0.05, r=0.46; P<0.05,
respectively).
Similarly, there was a significant positive correlation
between blood glucose levels at t40 and free and cued re-
call performance after the long delay (r=0.51; P<0.05,
r=0.55; P<0.05, respectively).
Recognition and discriminability
There was a significant positive correlation (across all
participants) between blood glucose levels at t45 and
long delay recognition performance (r=56; P<0.05).
Other measures of memory performance did not dis-
play a significant correlation with blood glucose levels.
Discussion
The results of this study further support the hypothesis
that glucose administration enhances memory function-
ing. More significantly, the findings indicate that this
cognitive facilitation effect persists under more naturalis-
tic conditions of glucose administration (i.e. as little as
2 h after a meal).
In regard to the glycaemic response, as expected the
ingestion of 25 g glucose resulted in a significant rise in
blood glucose levels. By contrast, there was no signifi-
cant rise in blood glucose levels after the ingestion of an
aspartame drink. Fast duration did affect baseline blood
glucose levels, as these were higher after the administra-
tion of a standardised meal. It is important to note, how-
ever, that the slightly different baseline in blood glucose
levels did not alter the cognitive facilitation effects of the
subsequently administered glucose drink.
Regarding cognitive performance, a significant effect
of glucose administration on immediate, short and long
delay memory performance was observable. The admin-
istration of the glucose drink resulted in a significantly
better immediate recall performance compared with the
aspartame drink, indicating that already at this early
stage a cognitive benefit from the consumption of glu-
cose was observable. In contrast to our previous findings
(Foster et al. 1998), a significant drink effect was also
observable on immediate recall of the interference list
(list B). There was no significant effect of condition on
the list B task. However, examination of Fig. 3 indicates
Table 1 Results of several memory tests for all conditions. Re-
sults are shown as mean and standard deviation and the units are
absolute scores for recall of interference list B, working memory
(digit span) and spatial memory performance (Rey-Osterrieth) and
percentages for the recognition accuracy and discriminability
Fasting Breakfast Lunch
GLUC ASP GLUC ASP GLUC ASP
List B recall 8.90 (±1.66) 8.40 (±2.17) 10.40 (±3.63) 7.00 (±1.83) 8.50 (±1.51) 6.30 (±3.77)
Recognition percentage correct 93.00 (±6.75) 84.00 (±10.22) 96.50 (±4.12) 82.50 (±6.77) 94.50 (±4.38) 77.50 (±10.34)
Recognition discriminability 94.32 (±4.58) 87.73 (±5.16) 97.28 (±2.79) 88.64 (±2.40) 93.86 (±3.04) 80.00 (±8.82)
Recognition response bias –0.06 (±0.22) –0.13 (±0.41) –0.03 (±0.11) –0.33 (±0.42) –0.05 (±0.39 –0.12 (±0.53)
Digit span 6.00 (±2.75) 6.30 (±2.82) 7.10 (±3.14) 6.40 (±2.41) 7.20 (±2.04) 6.80 (±2.25)
Rey-Osterrieth 26.50 (±4.99) 23.20 (±5.06) 29.10 (±3.25) 24.70 (±2.98) 26.10 (±6.01) 23.20 (±4.39)
that the effect was stronger after the administration of a
standardised meal. This could explain why a glucose ef-
fect on list B was not observed in earlier studies, in
which a standard meal was not administered. However, it
would seem more plausible that the glucose facilitation
effect is seen only sporadically on List B because this list
is administered only once, whereas for list A any drink-
related differences are magnified ×5 (corresponding to
5×administration of list A).
When comparing recall performance after the short
and the long delay, elevated plasma glucose levels due to
the administration of the glucose drink enhanced the per-
formance of participants on all delayed recall (cued and
free recall) and recognition components of the CVLT.
Differential facilitating effects of the three meal condi-
tions were only observable on short delay free recall and
long delay cued recall: both tasks were performed signif-
icantly better after the administration of a standardised
breakfast, as revealed by post hoc testing.
Long-term non-verbal memory performance was also
significantly increased after raising blood glucose levels
by means of a glucose drink. However, analogous to pre-
vious findings, the glucose drink failed to display a ben-
eficial effect on working memory performance, as evalu-
ated by the Digit Span test.
The observation that glucose administration repeated-
ly fails to exert any beneficial aspects on the Digit Span
test adds support to the notion of fractionation of memo-
ry enhancement effects. It appears plausible to relate the
observed glucose memory facilitation effect to the iden-
tification of insulin receptors in the hippocampus, as this
is the brain region most strongly implicated in long-term
memory performance (see Aggleton and Brown 1999,
for a recent review). Insulin receptors in the brain are
densely represented in the hippocampus, hypothalamus,
and olfactory bulb (Unger et al. 1989). As noted by
Unger et al., this distribution displays a remarkable simi-
larity to the primary areas of neuropathology seen in
Alzheimer patients. In addition, moderate doses of insu-
lin have been shown to increase firing of hippocampal
neurones (Palovcik et al. 1984). Given the established
role of the hippocampus in long-term declarative memo-
ry performance, elevated insulin in response to hyper-
glycaemia may boost glucose utilization in the hippo-
campus and result in improved memory performance
(Craft et al. 1993). Although the mechanisms through
which insulin affects brain function are, as yet, still un-
clear, variations in the concentration of insulin could
affect glucose utilisation in specific brain regions and,
consequently, the functions supported by those areas.
The failure to demonstrate a glucose facilitation effect
on digit span performance might also relate to the prop-
erties of the cholinergic system. The anti-cholinergic
drug scopolamine induces a cholinergic blockage that
impairs long-term memory acquisition, retrieval and de-
layed free recall. However, scopolamine seems to have
little effect on short-term memory (Wesnes et al. 1988).
Glucose acts in a manner suggesting increased activation
of cholinergic receptors (perhaps located in or near the
hippocampus). It has also been demonstrated that glu-
cose attenuates scopolamine-induced amnesia (Gold
1992). This notion might therefore offer some explana-
tion for the failure to observe glucose-induced improve-
ment on short-term memory performance. It has to be
noted, however, that facilitation of working memory
performance has been reported previously following glu-
cose administration (Martin and Benton 1999; Kennedy
and Scholey 2000). Further delineation of the aspects
of cognition that are facilitated by glucose is clearly
needed.
The main aim of this study was to increase the gener-
alisability of our previous findings and establish their
ecological validity. Demonstrating that the glucose facil-
itation effect is observable after a standardised meal fol-
lowed by a two hour fast has a significant consequence
for the relevance of our findings to everyday life. People
are likely to have a morning and/or an afternoon glu-
cose-containing snack or meal after a period of fasting of
just a few hours’ duration, as in this study, and then carry
out cognitively demanding tasks in their work environ-
ment (office, schools, university).
Previous studies did demonstrate differences in the
susceptibility to glucose memory facilitation between
participants who had or had not eaten breakfast immedi-
ately prior to glucose consumption and testing (Benton
and Parker 1998; Martin and Benton 1999). In those ex-
periments (Benton and Parker 1998; Martin and Benton
1999), fasting (no breakfast) was associated with poorer
performance on a working memory task. Although the
glucose drink improved working memory performance
of those who had fasted, it did not influence the perfor-
mance of those who had eaten breakfast. In those who
had fasted, the glucose drink resulted in working memo-
ry performance comparable to those who had consumed
breakfast. However, it is important to note that such rela-
tionship could not be found on a verbal memory task
(word list recall), when looking at the number of words
recalled. An influence of fasting was only found for the
time taken to complete this task, inasmuch as those who
had breakfast performed this task significantly faster. It
is arguable inasmuch the time taken to perform the word
list recall task (especially as participant were not in-
structed to perform this task as fast as possible and the
answers were written down) indicates better memory
performance on this task.
Moreover, it is important to note that in the present
study, 2 h intervened between meal consumption and
drink administration. The reason why previous studies
(Benton and Parker 1998; Martin and Benton 1999)
reported no facilitative effect of glucose administration
after the ingestion of a breakfast meal might be due to
the fact that, in their experiment, glucose administration
took place directly after breakfast ingestion. Taking into
account that the dose-response curve of glucose upon
memory performance follows an inverted U-shape,
further immediate post-meal glucose administration
might result in an “overload”, thus exceeding the opti-
mum dosage.
52
However, from a more theoretical viewpoint, the fact
that resting blood glucose levels were generally higher
after a 2-h fast compared with an overnight fast, but that
resting blood glucose levels did not generally influence
cognitive performance after administration of the glu-
cose drink, is also of great interest. These findings indi-
cate that the actual resting blood glucose level may not
be critical for the effect of glucose on facilitating memo-
ry performance, but rather what is critical is the actual
rise in blood glucose levels following administration of
the glucose drink.
The current consensus (such as it is) is that, in the nat-
ural environment, the effects of glucose administration
on memory performance are due to elevated blood glu-
cose levels within normal physiological limits, which are
raised in response to increased adrenaline output from
the adrenal medulla. This is thought to represent a mech-
anism that also mediates the effects of adrenaline on
memory (see, for example, Gold 1992). However, the
identification of insulin receptors in the brain (and spe-
cifically in the pyramidal cell layer and dentate gyrus of
the hippocampus; Unger et al. 1989; Schwartz et al.
1992) identifies another possible route by which glucose
might facilitate memory performance. Although there
seems to be a lack of consensus regarding the role of in-
sulin on brain glucose metabolism, it has been shown
that insulin can stimulate brain glucose uptake (Clarke
et al. 1984; Dringen and Hamprecht 1992) via an in-
crease in glucose transporter (GLUT1) mRNA (Werner
et al. 1989).
Taking into account the possible underlying mecha-
nisms of the glucose memory facilitation effect, it seems
reasonable that the positive cognitive effects of glucose
administration should also be observable under slightly
higher blood glucose levels, as observed in the present
study. Otherwise, the physiological basis of memory en-
hancement via increases in glucose levels (possibly sub-
sequent to peripheral adrenaline actions) would be re-
stricted, in the natural environment, to a sub-sample of
humans (and, presumably, other species) in whom base-
line blood glucose levels were within a limited range at
the time when the to-be-remembered episode occurred.
References
Aggleton JP, Brown WB (1999) Episodic memory, amnesia and
the hippocampal-anterior thalamic axis. Behav Brain Sci
22:425–486
Azari NP (1991) Effects of glucose on memory processes in
young adults. Psychopharmacology 105:521–524
Benton D, Owens DS (1993) Blood glucose and human memory.
Psychopharmacology 113:83–88
Benton D, Parker PY (1998) Breakfast, blood glucose, and cogni-
tion. Am J Clin Nutr 67:772S–778S
Benton D, Owens DS, Parker PY (1993) Blood glucose influences
memory and attention in young adults. Neuropsychologia
32:595–607
Clarke DW, Boyd FT, Kappy MS, Raizada MK (1984) Insulin
binds to specific receptors and stimulates 2-deoxyglucose up-
take in cultured glia cells from the rat. J Biol Chem 259:
11672–11675
Craft S, Zallen G, Baker LD (1992) Glucose and memory in mild
senile dementia of the Alzheimer type. J Clin Exp Neuropsy-
chol 14:253–267
Craft S, Dagogo-Jack SE, Wiethop BV, Murphy C, Nevins RT,
Fleischman S, Rice V, Newcomer JW, Cryer PE (1993) Effects
of hyperglycemia on memory and hormone levels in dementia
of Alzheimer Type: a longitudional study. Behav Neurosci
207:926–940
Craft S, Murphy C, Wemstrom J (1994) Glucose effects on com-
plex memory and nonmemory tasks: the influence of age, sex,
and glucoregulatory response. Psychobiology 22:95–105
Delis DC, Kramer JH, Kaplan E, Ober BA (1987) CVLT: Califor-
nian Verbal Learning Test. Psychological Corporation/
Harcourt Brace Jovanovich, New York
Dringen R, Hamprecht B (1992) Glucose, insulin and insulin-like
growth factor I regulate the glycogen content of astroglia-rich
primary cultures. J Neurochem 58:511–517
Foster JK, Lidder PG, Sünram SI (1998) Glucose and memory:
fractionation of enhancement effects? Psychopharmacology
137:259–270
Fucetola R, Newcomer JW, Craft S, Melson AK (1999) Age- and
dose-dependent glucose-induced increases in memory and
attention in schizophrenia. Psychiatr Res 18:1–13
Gold PE (1986) Glucose modulation of memory storage process-
ing. Behav Neural Biol 45:342–349
Gold PE (1991) An integrated memory regulation system: from
blood to brain. In: Frederickson RCA, McGaugh JL, Felten
DL (eds) Peripheral signaling of the brain: role in neural-im-
mune interactions and learning and memory. Hogrefe and
Huber, Toronto, pp 391–419
Gold PE (1992) Modulation of memory processing: enhancement
of memory in rodents and humans. In: Squire LR, Butters N
(eds) The neuropsychology of memory, 2nd edn. Guilford
Press, New York, pp 402–414
Gold PE, Vogt J, Hall JL (1986) Posttraining glucose effects on
memory: behavioral and pharmacological characteristics. Be-
hav Neural Biol 46:145–155
Gonder-Frederick L, Hall JL, Vogt J, Cox DJ, Green J, Gold PE
(1987) Memory enhancement in elderly humans: effects of
glucose ingestion. Physiol Behav 41:503–504
Gradman TJ, Laws A, Thompson LW, Reaven GM (1993)Verbal
learning and/or memory improves with glycaemic control in
older subjects with non-insulin-dependent diabetes mellitus.
J Am Geriatr Soc 41:1305–1312
Hall JL, Gonder-Frederick LA, Chewning WW, Silveira J, Gold
PE (1989) Glucose enhancement of performance on memory
tests in young and aged humans. Neuropsychologia 27:1129–
1138
Holmes CS, Hayford JT, Gonzalez JL, Weydert JA (1983) A sur-
vey of cognitive functioning at different glucose levels in dia-
betic persons. Diabetes Care 6:180–185
Kennedy DO, Scholey AB (2000) Glucose administration, heart
rate and cognitive performance: effects of increasing mental
effort. Psychopharmacology 149:63–71
Kopf SR, Baratti CM (1996) Effects of posttraining administration
of glucose on retention of a habituation response in mice: par-
ticipation of a central cholinergic mechanism. Neurobiol Learn
Mem 65:253–260
Lee MK, Graham SN, Gold PE (1988) Memory enhancement with
posttraining intraventricular glucose injections in rats. Behav
Neurosci 102:591–595
Manning CA, Ragozzino ME, Gold PE (1993) Glucose enhance-
ment of memory in patients with probable senile dementia of
the Alzheimer’s type. Neurobiol Aging 14:523–528
Manning CA, Honn VJ, Stone WS, Jane JS, Gold PE (1998) Glu-
cose effects on cognition in adults with Down’s syndrome.
Neuropsychology 12:479–484
Martin PY, Benton D (1999) The influence of a glucose drink on a
demanding working memory task. Physiol Behav 67:69–74
McGaugh JL (1989) Involvement of hormonal and neuromodula-
tory systems in the regulation of memory storage. Annu Rev
Neurosci 12:255–287
53
Schwartz MW, Figlewicz DP, Baskin DG, Woods SC, Porte D Jr
(1992) Insulin in the brain: a hormonal regulator of energy
balance. Endocrinol Rev 13:387–414
Sieber FE, Trastman RJ (1992) Special issues: glucose and the
brain. Crit Care Med 20:104–114
Thompson RT (1967) Foundations of physiological psychology.
Harper International, New York, pp 111–128
Unger J, McNeill TH, Moxley RT III, White M, Mosi A,
Livingstone JN (1989) Distribution of insulin-like-immunore-
activity in the rat forebrain. Neuroscience 31:143–157
Wechsler DA (1987) Wechsler memory scale – revised. Harcourt
Brace Jovanovich, NewYork
Wenk GL (1989) An hypothesis on the role of glucose in the
mechanism of action of cognitive enhancers. Psychopharma-
cology 99:431–438
Werner H, Raizada MK, Mudd LM, Foyt HL, Simpson IA,
Roberts CTJ, LeRoith D (1998) Regulation of rat brain/HepG2
glucose transporter gene expression by insulin and insulin-like
growth factor-I in primary cultures of neuronal and glia cells.
Endocrinology 125:413–320
Wesnes K, Simpson PM, Kidd AG (1988) An investigation of the
range of cognitive impairments induced by scopolamine
0.6 mg. Hum Psychopharmacol 3:27–43
Winocur G (1995) Glucose-enhanced performance by aged rats on
a test of conditional discrimination-learning. Psychobiology
23:270–276
White NM (1991) Peripheral and central memory enhancing
actions of glucose. In: Frederickson RCA, McGaugh JL,
Felten DL (eds) Peripheral signaling of the brain: role in
neural-immune interactions and learning and memory.
Hogrefe and Huber, Toronto, pp 421–441
54
Meneilly GS, Cheung E, Tessier D, Yakura C, Tuokko H (1993) The
effect of improved glycaemic control on cognitive functions in
the elderly patient with diabetes. J Gerontol 48:M117–M121
Messier C, Gagnon M (1996) Glucose regulation and cognitive
functions: relation to Alzheimer’s disease and diabetes. Behav
Brain Res 75:1–11
Messier C, Gagnon M, Knott V (1997) Effect of glucose and pe-
ripheral glucose regulation on memory in the elderly. Neuro-
biol Aging 18:297–304
Messier C, Pierre J, Desrochers A, Gravel M (1998) Dose-depen-
dent action of glucose on memory process in women: effect on
serial position and recall priority. Con Brain Res 7:221–233
Messier C, Desrochers A, Gagnon M (1999) Effect of glucose,
glucose regulation, and word imagery value on human memo-
ry. Behav Neurosci 113:431–438
Metzger MM (2000) Glucose enhancement of a facial recognition
task in young adults. Physiol Behav 68:549–553
Newcomer JW, Craft S, Fucetola R, Moldin SO, Selke G, Paras L,
Miller R (1999) Glucose-induced increase in memory perfor-
mance in patients with schizophrenia. Schizophr Bull 25:321–
335
Osterrieth PA (1944) Le test de copie d’une figure complexe. Arch
Psychol 30:206–356
Owens DS, Benton D (1994) The impact of raising blood glucose
on reaction times. Neuropsychobiology 30:106–113
Palovcik RA, Phillips MI, Kappy MS, Raizada MK (1984) Insulin
inhibits pyramidal neurons in hippocampal slices. Brain Res
309:401–404
Parker PY, Benton D (1995) Blood glucose levels selectively in-
fluence memory for word lists dichotically presented to the
right ear. Neuropsychologia 33:843–854