Biological PsychoIogy 24 (1987) 95-100 95

North-Holland

GLUCOSE IMPROVES ATTENTION AND REACTION

TO FRUSTRATION IN CHILDREN

David BENTON, Veronica BRETT and Paul F. BRAIN

Departments of Psychology and Zoology, University College, Swansea SA2 8PP, U.K.

Accepted for publication 21 October 1986

There have been many suggestions that both dietary sugar and the level of blood glucose may
be associated with abnormal behaviour. To examine this question 60 six- and seven-year old
children were given a drink that contained either 25 grams of glucose or a placebo towards the end
of the school day. They were then subjected to one test that required sustained attention and to a
second that was intentionally frustrating. Those given the drink containing glucose were more
capable of sustaining attention and showed fewer signs of frustration.

1. Introduction

The relationship between dietary sugar, blood glucose levels and the be-
haviour of children is a cause of continuing controversy. There is a popular
notion that sugar is the cause of adverse behaviour in children (Crook, 1974),
although an increasing number of double-blind studies that have challenged a
child with sucrose or glucose suggest that this is not the case (Behar, Rapo-
port, Adams, Berg, & Comblath, 1984; Ferguson, Stoddart, & Simeon 1986;
Mahan et al., 1985; Rapoport, 1986). Prinz, Roberts, and Hantman (1980)
asked parents to keep a diary of the food eaten by their child; in hyperactive
children a correlation was found between sugar consumption and destructive/
aggressive behaviour. In children without behavioural problems sugar con-
sumption has been reported to be related to motor activity but not to
aggressive behaviour (Wolraich, Milich, Stumbo, & Schultz, 1985). Children
above the 75th percentile for dietary sugar intake have also been reported to
have problems of sustaining attention (Prinz & Riddle, 1986). As such correla-
tions between diet and behaviour do not allow a statement to be made
concerning causal relationships, the present study challenged a group of
children with glucose and examined several behavioural indices. To date much
of the work has concentrated on hyperactive children although the weight of
evidence suggests that few hyperactive children respond to sugar. The present
study has, therefore, extended the data by examining a sample of children not
pre-selected in this way.

0301-0511/87/$3.50 0 1987, Elsevier Science Publishers B.V. (North-Holland)

96 D. Benton et al. / Glucose and behauiour

2. Method

2.1. Subjects

The subjects were 60 children who attended two primary schools. The
socio-economic backgrounds of the children were not systematically explored
but appeared to be fairly homogenous; the catchment areas were described by
the schools as largely middle class, and none of the pupils had a history of
behaviour problems that had led to referral to child guidance or similar
services. They were aged 6 to 7 years and exactly half were female: Their
parents provided informed consent.

2.2. The ability to sustain attention

The paradigm of Shakow (1962) was used. Reaction times were measured
using a timer (ms); a light came on and the time taken to push a button was
recorded. To measure the ability to sustain attention, subjects were given a test
of reaction times that started after either a 3 or 13 s delay. Following a verbal
warning, the children were told to press a button when a light appeared. The
test consisted of four blocks of six trials. In the first block, the delay between
the warning and the appearance of the light was 3 s, in the second and third it
was 13 s, and in the final block of trials it was again 3 s. The data were
analyzed using a four-way analysis of variance (glucose/placebo x first or
second block of delay of a particular duration x delay for 3 or 13 s x trials).

2.3. Reaction to frustration

The second test used a commercially available television computer game,

Tele-match 4 model 6600, in which an electronic representation of a ball
moves from the left towards the right of a television screen. None of the
subjects were familiar with the game. By moving a knob the subject could
attempt to place an electronic bat in the way of the ball. The difficulty of the
task could be altered by adjusting the speed of the ball and the angle that it
leaves the side walls. The conditions were chosen to make the task very
difficult (fast ball and 40 angle). Thus the child could be expected to
experience failure; the aim was to study the reaction to the resulting frustra-
tion. Each subject performed 10 sessions of 15 trials during which ethological
descriptions of behaviour were recorded. The machine automatically delivered
15 balls in 19 s. In many of the blocks of 19 s the child paid attention to the
task and was categorized as quietly concentrating. If other behaviour was
observed, it was recorded as falling into one of the other categories:
(1) Quietly concentrating: This was by far the most common response in
which the child paid attention to the task and did not display the
behaviours in the other categories.
 D. Benton et al. / Glucose and behaviour 97

(2) Fidgeting: Any restless movements of the hands or arms, shuffling of the
body in the chair or feet on the floor.
(3) Signs of frustration: Roughly handling the controls, kicking the feet,
auditory expressions of annoyance of a non-verbal nature (e.g. sighing).
(4) Talking: Any verbal statement.
The data were analyzed using the Mann-Whitney U Test.

2.4. Procedure

Randomly, and under a double-blind procedure so neither the observer nor

the child knew which drink had been administered, the children were given a
drink of diet orange cordial (diluted with water and sweetened only with
saccharine by the manufacturers, Boots of Nottingham) to which either 25
grams of glucose (on average 1.14 g/kg) or additional saccharine had been
added so that in preliminary tests children reported a similar degree of
sweetness. The children ate lunch from 1230 to 1300, supplied in some cases
by the school and in others by the parents. They received the drink at 1430
and completed the two tests from 1445 to 1530. All testing was carried out by
the same individual.

3. Results

The reaction time data were associated with problems of heterogeneity of

variance; in some cases the longer periods of delay resulted in a failure of

.80 -
-I- 0 Glucose
i? Placebo
._
,E .75 -
i,
:
LT

z
-r
3 .70-
::
0
5
$
.65=

0T 3 second 13 second
delay delay

Fig. 1. Reciprocal of mean reaction time at two warning intervals for glucose and placebo groups.
The data, mean seconds k s.d., are reciprocals of reaction times following a delay of either 3 or 13
s after an initial warning. Those receiving glucose show significantly faster reaction times.
98 D. Benton et al. / Glucose and behaviour

Table 1
The behaviour of children playing with a television game deliberately designed to be frustrating a

Trials 1-5 Trials 6-10

Placebo n = 30 Glucose n = 30 Placebo Glucose
Quietly
concentrating 2(0-5) q2-5) p < 0.00015 1 (O-5) 2.5(0-5) p < 0.0002
Fidgeting O(O-4) 0(0-l) n.s. 0.5(0-3) 0 (O-2) p i 0.04
Signs of
frustration l(O-3) O(O-2) p < 0.034 2 (O-5) 0 (O-3) p < 0.0006
Talking l(O-5) l(O-5) n.s. 1.5(0-5) 0 (O-4) p < 0.01

a Each of ten trials consisted of the delivery of a series of 15 balls. During each trial the childs
behaviour was observed and described as falling into one of the above four categories. The data
are medians and ranges compared using the Mann-Whitney U Test.

attention that produced particularly long reaction times. The problem was
solved satisfactorily by transforming the data by taking the reciprocal (l/( Y +
OS), where Y was reaction time in s). Fig. 1 illustrates the reaction time data.
The glucose/placebo main effect was significant (F(l,%) = 5.27, p < 0.05);
those who had received the glucose had significantly faster reaction times. Fig.
1 also illustrates that the delay of 13 s resulted in slower reaction times
(F(1,58) = 69.51, p < O.OOl), evidence that the test measured the ability to
sustain attention. None of the higher order interactions involving the
glucose/placebo dimension reached statistical significance.
Table 1 shows the behaviour of children during the frustrating television
game. The administration of glucose was associated with marked differences in
the response to the test. Those having the glucose drink were more likely to
quietly concentrate during both the first five (p < 0.00015) and second five
trials (p -c 0.00025). During trials 6-10 the children receiving the placebo were
more likely to fidget (p < 0.04), to show signs of frustration ( p < 0.0006) and
were more likely to talk ( p < 0.01).

4. Discussion

The present results inevitably lead to the controversial suggestion that the
beneficial influence of orally administered glucose may to some extent result
from the brains of these children being starved of glucose. Many popular
books (Duffy, 1975; Fredericks & Goodman, 1969; Martin, 1970; Steincrohn,
1973), although not many of the informed scientific community (Statement,
1973) propose that some individuals react to sugar intake by the prolonged
production of insulin to the extent that hypoglycaemia results; glucose values
fall to the extent that brain functioning is disrupted. The fact that blood
 D. Benton et al. / Glucose and behaviour 99

glucose levels were not measured in the present study makes any discussion of
this topic very speculative. Meal tolerance tests suggest that sugars, as part of
a mixed meal, are associated with a relatively slow increase in blood glucose
levels that remain at lower levels for longer periods than when glucose is given
by itself (Buss, Kansal, Roddam, Pino, & Boshell, 1982; Charles et al., 1981;
Hogan, Service, Sharbrough, & Gerich 1983). Thus, in the present sample who
had eaten recently, it is perhaps unlikely that many, if in fact any, were
hypoglycaemic. A more speculative suggestion is that some individuals may
develop locally low levels of intracellular glucose within the brain (Pardridge,
1986); this awaits the development of the technology that would allow
investigation of this question. It should not be assumed that the presumed
increase in the levels of blood glucose is the only possible mechanism that may
result in these findings. The ingestion of glucose prompts the secretion of
several gut hormones, some of which are believed to effect brain metabolism,
as well as having a direct effect upon the autonomic nervous system. Logically,
it is possible that the results reflect a negative reaction to saccharine rather
than a positive reaction to glucose.
When challenged with glucose hyperactive children have been found to
display decreases in motor activity (Behar et al., 1984) although others have
failed to find changes in either activity (Rapoport, 1986) or cognitive function-
ing (Behar et al., 1984). Thus the literature does not lead to the expectation
that the present very striking results would be obtained. Various methodologi-
cal factors may account for the powerful influence of glucose found here: The
ages of the children, the tasks, and the time of day when they were tested are
possibly important variables. The present study used a much larger sample
size than is common in this area and the ages of the children were homoge-
neous to an extent that has been unusual in previous work. The studies of
putatively sugar reactive children have very often used small samples whose
ages vary by as much as 10 years. Because the cognitive functioning of
children changes very markedly with age, it could well be that the impact of
sugars at critical stages has been missed. Most previous studies have con-
centrated on ethological descriptions of play or measures of gross movements,
and no other study has systematically subjected their subjects to a frustrating
task. The time of day is another factor that may be important in the present
study; the impact of oral glucose may be most apparent later in the day when
a range of factors predispose the child to act poorly.
Both parents and teachers will find unremarkable the observation that
children tend to be more difficult in the late afternoon. Many factors are
potentially involved; the child may be tired, hungry, bored or in need of
exercise. The present data very clearly illustrate the ability of a glucose drink
to facilitate the ability of children to concentrate and to react positively when
frustrated at this time of day. The question as to whether these findings can be
extended to other tasks and other situations awaits future study.
100 D. Benton et al. / Glucose and behaviour

References

Behar, D., Rapoport, J.L., Adams, A.J., Berg, C.J. & Cornblath, M. (1984). Sugar challenge testing
with children considered behaviorally sugar reactive. Nutrition and Behavior, I, 277-288.
Buss, R.W., Kansal, P.C., Roddam, R.F., Pino, J., & Boshell, B.R. (1982). Mixed meal tolerance
test and reactive hypoglycemia. Hormones and Metabolic Research, 14, 281-283.
Charles, M.A., Hofeldt, F., Shackelford, A., Waldeck, N., Dodson, L.E., Bunker, D., Cogins, J.T.
t Eichneer, H. (1981). Comparison of oral glucose tolerance tests and mixed meals in patients
with apparent idiopathic postabsorptive hypoglycemia: Absence of hypoglycemia after meals.
Diabetes, 30, 465-470.
Crook, W.G. (1974). An alternative method of managing the hyperactive child. Pediatrics, 54, 656.
Duffy, W. (1975). Sugar blues. Radnor, PA: Chilton.
Ferguson, H.B., Stoddart, C., & Simeon, J.G. (1986). Double-blind challenge studies of behavioral
and cognitive effects of sucrose-aspartame ingestion in normal children. Nutrition Reviews, 44,
144-150.
Fredericks C., & Goodman, H. (1969). Low blood sugar and you. New York: Constellation
International.
Hogan, M.J., Service, F.J., Sharbrough, F.W., & Gerich, J.E. (1983). Oral glucose tolerance test
compared with a mixed meal in the diagnosis of reactive hypoglycemia. Mayo Clinic Proceed-
ings, 58, 491-496.
Mahan, L.K., Chase, M., Furukawa, CT., Shapiro, G.G., Pierson, W.E., & Bierman, W. (1985
March). Sugar AZlergv and childrens behavior. Paper presented at the 41st Annual Meeting of
the American Academy of Allergy and Immunology, New York.
Marks, V., & Rose, F.G. (1981). Hypoglycaemia (2nd ed.). Oxford: Blackwell Scientific Publica-
tions.
Martin, C.G. (1970). Low blood sugar - the hidden menace of hypoglycaemia. New York: ABC
Books.
Pardridge, W.M. (1986). Blood-brain barrier transport of nutrients. Nutrition Reviews, 44, 15-24.
Prim, R.J., & Riddle, D.B. (1986). Association between nutrition and behavior in five-year-old
children. Nutrition Reviews, 44, 151-157.
Prinz, R.J., Roberts, W.A., & Hantman, E. (1980). Dietary correlates of hyperactive behavior in
children. Journal of Consulting and Clinical Psychology, 48, 760-769.
Rapoport, J.L. (1986). Diet and hyperactivity. Nutrition Reviews, 44, 8-162.
Shakow, D. (1962). Segmental set. Archives of General Psychiatry, 6, 1-17.
Statement (1973). Statement of the Ad Hoc Committee on hypoglycaemia. Annuals of Internal
Medicine, 78, 300-301.
Steincrohn, P.J. (1973). Low blood sugar. The most common misdiagnosed disease. London: Allison
t Busby.
Wolraich, M., Milich, R., Stumbo, P., & Schultz, F. (1985). Effects of sucrose ingestion on the
behavior of hyperactive boys. Journal of Pediatrics, 106,, 675-682.