Journal of Pediatric Psychology, Vol. 25, No. 7, 2000, pp. 515-520
© 2000 Society of Pediatric Psychology
Brief Report: Speed of Information Processing in Children With Insulin-Dependent Diabetes Mellitus
1 University of Maryland, 2 University of Houston, 3 University of Michigan
All correspondence should be sent to Robert Kail, Department of Psychological Sciences, Purdue University, West Lafayette, Indiana 47907-1364. E-mail: rk{at}psych.purdue.edu .
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionReferences |
|---|
Objective: To determine whether children with insulin-dependent diabetes mellitus (IDDM) process information more slowly than children who do not have diabetes.
Methods: We tested 31 children with early onset and longer duration of IDDM, 35 with later onset and briefer duration of IDDM, and 36 comparison children without diabetes. They were administered five tasks requiring rapid responding that assessed a range of cognitive processes.
Results: On most tasks, children in the three groups were quite similar in the accuracy and speed of performance. Furthermore, for children in the diabetic groups, disease-related variables were unrelated to accuracy and speed of performance.
Conclusions: The results suggest that children with IDDM do not have a pervasive deficit in speed of information processing, although more circumscribed deficits in processing speed are possible.
Key words: diabetes; processing speed; cognitive development.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Children who develop insulin-dependent diabetes mellitus (IDDM, also known as Type 1 diabetes) before 5 years of age often have impaired performance on cognitive tasks. For example, children who develop IDDM before age 5 often have lower scores on measures of intelligence, attention, learning, memory, and visual spatial processing (Hagen et al., 1990
; Ryan, Vega &
Drash, 1985). This impaired performance in children with IDDM
might reflect slower speed of cognitive processing because, in children
without IDDM, increased processing speed is associated with more accurate
retention and more effective reasoning and problem solving
(Fry & Hale, 1996). If
children who develop IDDM before age 5 process information more slowly than
children without IDDM, their performance on memory and cognitive tasks may be
sometimes impaired.
In fact, the extant literature on speed of processing in children with IDDM
is sparse and inconsistent. In the most extensive study
(Ryan et al., 1985), children
who developed IDDM before age 5 responded more slowly on two tasks: (1) a
symbol substitution task in which a code consisting of digits paired with
unique symbols appeared at the top of the page and children printed the digit
adjacent to symbols at the bottom of the page, and (2) a pegboard task in
which children moved pegs between two sets of holes. Children with IDDM did
not, however, respond more slowly on a third task, trails, in which they
quickly drew lines linking letters and numbers in a specified order.
Two other studies included a single measure of processing speed. On the one
hand, Sansbury, Brown, and Meacham
(1997) found that, among a
group of children with IDDM, those who developed IDDM early tended to respond
more slowly on a visual search task. On the other hand, Holmes, Dunlap, Chen,
and Cornwell (1992) found that
children with IDDM did not respond more slowly than children without IDDM on a
symbol substitution task like that used by Ryan et al.
(1985).
Given these inconsistent findings and the critical role that processing
speed plays in memory and cognitive functioning in healthy children
(Kail, 1995), this study was
designed to provide a comprehensive evaluation of speed of processing in
children with IDDM. Our sample included children who developed IDDM before age
5, a separate group of children who developed IDDM at age 5 or older, and a
comparison group of children who did not have IDDM. Because a deficit in
processing speed may appear in some but not all speeded tasks, we used five
tasks that tap different cognitive processes and that reveal age differences
in processing speed (Kail,
1995). Three tasks (mental rotation, form board, analogies) relied
heavily on visual-spatial processing; these were chosen because investigators
have reported impaired performance by children with IDDM in the visual-spatial
domain (Ryan et al., 1985). The
other tasks (addition and visual search) relied more on verbal skills. We
hypothesized that, if processing speed is impaired in children who develop
IDDM early (and thus have had the disease for a longer duration), they should
respond more slowly on some if not all these speeded tasks.
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Participants
The results described here were obtained from children participating in a comprehensive study of psychological and behavioral characteristics of children with IDDM (Wolters, Yu, Hagen, & Kail, 1996
). We tested 77 children with IDDM and 36 children
without IDDM. However, 11 children with IDDM were excluded because they did
not meet the age of onset and duration of illness requirements described
below. Thus, 66 children with IDDM and 36 comparison children were included.
The 49 boys and 53 girls ranged in age from 9.1 to 16.7 years, with a mean of
12.6 years (SD = 2.24 years). Children with IDDM were categorized
into two groups: 31 children diagnosed before their fifth birthday and at
least seven years before testing constituted the earlier onset/longer duration
group; 35 children diagnosed after their fifth birthday and no more than seven
years before testing constituted the later onset/briefer duration group.
(Hereafter, for simplicity these are referred to as early and late onset
groups, respectively.) All children with IDDM were diagnosed at least two
years before testing. The early onset, late onset, and comparison groups
included 1, 1, and 6 African American children, respectively; all remaining
children were white. Details concerning the sample can be found in Wolters et
al. (1996).
Procedure
Children and at least one parent came to our laboratory for two 1
-
to 2-hour sessions separated by at least 30 days. The general procedures,
described in Wolters et al.
(1996), included cognitive and
memory tasks and a parental interview. Blood glucose was assessed in children
with IDDM; children were not tested if they were hypoglycemic or
hyperglycemic.
Stimuli were presented on a computer monitor (mental rotation, addition,
visual search) or were prepared as slides and projected onto a
8x 10 in. screen (analogies, form board). Children responded by
pressing one of two keys on a keyboard with their right index finger. Due to
space limits, details of specific tasks are not presented here; they are
available upon request from the authors. The order of tasks was rotation,
addition, and visual search in session 1 and analogies and form board in
session 2.
Mental Rotation. On each of 80 trials, children judged whether a stimulus presented in different orientations was a letter or the mirror image of a letter.
Analogies. On each of 48 trials, there were two pairs of rectangles, with geometric figures inside each rectangle. Children determined the relation between the figures in the first pair of rectangles, then decided whether the figures in the second pair were related in like manner.
Form Board. On each of 64 trials, a problem consisted of a completed figure (e.g., square) and an array of pieces; children judged whether the pieces in the array could make the completed figure.
Addition. On each of 60 trials, children determined the truth of addition statements in the form m + n = y, where m and n ranged from 1 to 9.
Visual Search. On each of 36 trials, children first studied a target letter approximately 4 sec. Then 1, 3, or 5 letters were presented, and children decided if the target letter was present.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Children performed quite accurately. Children in the early onset and comparison groups tended to perform at about the same level of accuracy, and children in the late onset group tended to be slightly less accurate. This difference in accuracy was significant on the rotation and analogies tasks, Fs
3.21, p <.05 (rotation: early onset, 94.4%; late
onset, 89.5%; comparison, 94.9%; analogies: early onset, 86%; late onset, 81%;
comparison, 87.7%). Each task included several within-subject conditions, each of which included 6-8 response times (RTs). For example, mental rotation included 10 within-subject conditions, reflecting orthogonal combination of 5 stimulus orientations and 2 responses (identical, mirror image). Within each of these conditions, a mean RT was computed from times on correct responses. To reduce the impact of outlying RTs, any RT within a condition that was twice the mean was deleted and the mean was recomputed. Fewer than 1% of RTs were deleted in this fashion. To account for age-related variance in performance, the samples were divided at the median chronological age to create "younger" and "older" groups. The data from each task were analyzed with analysis of variance (ANOVA) in which there were two between-group factors (group, age) and two within-group factors that defined the structure of the task (e.g., for mental addition, size of the stated sum and whether the stated response was true or false).
Results of primary importance were the main effects of groups (early onset,
late onset, comparison) and interactions between groups and other variables.
On four tasks, these effects were all nonsignificant, Fs < 1 for
main effects; Fs 
2.18, ps >.05 for interactions. As
shown in the left panel of Figure
1, on the analogies, form board, mental addition, and visual
search tasks, children in the three groups responded in approximately the same
amount of time. On mental rotation, the main effect of group was not
significant, but the interaction of age, group, and orientation was,
F(8, 174) = 2.19, p <.05. As shown in the right panel of
Figure 1, for older children,
RTs increased as a function of orientation at about the same rate for all
groups. For younger children (Figure
1, center panel), this was true for orientations between 0 and 90
degrees. However, at 120 degrees, the mean RT for children in the early onset
group was unexpectedly less than the mean at 90 degrees, whereas the mean RTs
at 120 degrees for children in the late onset and comparison groups were
greater than the corresponding means at 90 degrees.
|
In a final analysis, we computed correlations between performance on speeded tasks and several diabetes-related variables: age of onset of IDDM, parent's rating of control (as a global measure of the child's long-term control of the disease), and blood glucose levels at each testing session (as a measure of the child's current level of control). In these analyses, mean level of accuracy and mean RT for each task were calculated separately for each child. In addition, overall measures of accuracy and RT were determined by calculating, for each child, a standard score on each task and summing these standard scores across tasks. For example, the overall measure of accuracy was simply the sum of each child's standard scores for accuracy on the five tasks. Of the 48 rs between diabetes-related variables and the 12 performance measures, only one was significant, approximately the number expected by chance.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
The primary outcome of the current study was that, contrary to our hypothesis, children who developed IDDM early (and thus had the disease for a longer duration) performed at a similar rate as children in the late onset group and children without IDDM. Because group differences were not significant, it is important to show that our study was sufficiently sensitive to find such differences had they existed. Four factors suggest that the tasks and design were sufficiently sensitive. First, tasks such as these have been used previously to reveal deficits in clinical groups, such as children with specific language impairment (Kail, 1994
). Second, we found the usual pattern of age differences.
Consistent with previous research on developmental change in speed of
cognitive processing, older children responded more rapidly than younger
children on each task, and the magnitude of these age differences resembled
those found previously (Kail,
1995). Third, within each task, we found the usual effects of task
complexity or difficulty. Children responded more slowly when completing the
more complex portions of each task. For instance, on mental addition, children
responded more slowly on trials with greater sums. Fourth, power analyses
indicate that our study would have detected differences of the size found
previously. Specifically, the differences reported by Ryan et al.
(1985) and by Sansbury et al.
(1997) ranged from
"medium" to "large" effects (as defined by Cohen,
1979). The power of this design to detect such differences between children
with IDDM and those without IDDM was.71 and.98, respectively. Thus, it is
unlikely that clinically important differences between groups would have gone
undetected in this design. Although most analyses showed no differences between groups, in three instances children with IDDM had impaired performance. There were group differences in accuracy on two of the tasks, with the late onset group less accurate than the other two. This outcome was unexpected based on previous work, in which impaired performance is more likely when onset of IDDM is early rather than late. However, overall accuracy was reasonably high, especially on the rotation task. More important, given the special interest in the early onset group, it is noteworthy that children in this group were just as accurate as children in the comparison group; children in the early onset group did not sacrifice accuracy to achieve processing speed comparable to children in the comparison group.
The third instance of a group difference occurred on the rotation task. The interaction of age, group, and orientation was significant, reflecting the fact that children in the early onset group responded more rapidly when stimuli appeared at 120 degrees than at 90 degrees. This result is unusual on this task and we have no obvious explanation. More important, however, is that at both ages children in the early onset groups responded more rapidly on the rotation task than children in the comparison group (though not significantly so). Thus, none of the findings of group differences suggests a pattern of serious impairment in children with IDDM.
These findings of unimpaired processing speed in children with IDDM differ
from those of Ryan et al.
(1985). One explanation is that
the tasks used by Ryan et al. included a more demanding response component
than the tasks we used. In symbol substitution, children moved the pencil from
one symbol to the next, then printed the correct digit; in pegboard, children
moved their hand to a peg, grasped it, then moved the peg to a new location.
In contrast, in the tasks used here children simply pushed a button to
respond. Perhaps children with IDDM have slower motor but not cognitive
processing. This deficit would have produced group differences in the Ryan et
al. study but not in these studies because of the much simpler response (and
because motor processing time was small relative to the cognitive processing
time). A variant of this explanation is that the tasks in Ryan et al.
emphasize efficient serial processing—children had to rapidly plan and
execute a series of movements—whereas such serial processing was
relatively unimportant in our tasks.
However, both of these explanations lead to the predictions of (1) group
differences on the trail making test used by Ryan et al. and (2) no
differences on the visual search task used by Sansbury et al.
(1997). These predictions
counter the results obtained in those studies, which suggest that additional
explanations are needed to account for the inconsistent results in the
literature.
Regardless of the specific explanation for the difference between tasks that yield group differences and those that do not, it seems clear that, overall, children with IDDM are not consistently slower in their information processing. Perhaps future research will reveal a more circumscribed processing speed deficit, one linked to motor processing in the manner suggested previously, but it seems unlikely that early onset of IDDM is associated with a pervasive deficit in processing speed.
The finding of no differences in speed of processing is a beneficial result
for children with IDDM. Because speed of processing predicts many important
developmental outcomes, including memory and problem solving, had children
with IDDM processed information more slowly than average, the expectation
would have been for less skilled reading and problem solving. Because children
with IDDM processed information at age-appropriate rates, processing speed
should not hinder their ability to achieve normal levels of reading and
problem solving. When children with IDDM do not achieve normal levels of
remembering, reading, and problem solving, other factors besides
diabetes-related deficits in processing speed are apparently responsible.
Executive processing and memory, for example, may contribute to impaired
performance by children with IDDM (Hagen et
al., 1990; Ryan et al.,
1985). Furthermore, although each of these deficits may be
relatively small, in aggregate they can lead to below-average achievement,
particularly when children are exposed to other risk factors unrelated to
diabetes.
Finally, there was no evidence that metabolic control-related aspects of
IDDM were related to children's speed of cognitive processing. Neither
metabolic history (age at onset, parental report of control) nor current
control (blood glucose levels) was related to performance. At first glance,
these findings may appear inconsistent with previous work showing that (1)
individuals with a history of poor metabolic control complete speeded tasks
more slowly than those with better control
(Ryan & Williams, 1993) and
(2) speed of processing is slower during hypoglycemic states
(Holmes, Hayford, Gonzalez, & Weydert,
1983). However, our measures of metabolic history were limited
(e.g., parents' ratings of overall metabolic control may be inaccurate or
unreliable). Furthermore, because children were not tested if they were either
hyperglycemic or hypoglycemic, our results simply show that blood glucose
levels within the normal range are unrelated to speeded performance.
|
Acknowledgments |
|---|
This research was funded by a grant (DK43725) from the National Institute of Diabetes and Digestive and Kidney Diseases. We thank Douglas Henry, Alvah O. Conley, Amy Gendleman, Andrew Frey, Christopher Houck, Matt Jaffe, and John Lundin for their contributions to this research.
Received November 30, 1998; revision received April 15, 1999; accepted September 21, 1999
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Cohen, J. (1977). Statistical power analysis for the behavioral sciences. New York: Academic Press.
Fry, A. F., & Hale, S. (1996). Processing speed, working memory, and fluid intelligence: Evidence for a developmental cascade. Psychological Science, 7, 237-241.[Web of Science][Medline]
Hagen, J. W., Barclay, C. R., Anderson, B. J., Freeman, D. J., Segal, S. S., Bacon, G., & Goldstein, G. W. (1990). Intellective functioning and strategy use in children with insulin-dependent diabetes mellitus. Child Development, 61, 1714-1727.[Web of Science][Medline]
Holmes, C., Dunlap, W. P., Chen, R. S., & Cornwell, J. M. (1992). Gender differences in the learning status of diabetic children. Journal of Consulting and Clinical Psychology, 60, 698-704.[Abstract]
Holmes, C., Hayford, J., Gonzalez, J., & Weydert J. (1983). A survey of cognitive functioning at different glucose levels in diabetic persons. Diabetes Care, 6, 180-185.
Kail, R. (1994). A method for studying the generalized slowing hypothesis in children with specific language impairment. Journal of Speech and Hearing Research, 37, 418-421.
Kail, R. (1995). Processing speed, memory, and
cognition. In W. Schneider and F. E. Weinert (Eds.), Memory
performance and competencies (pp.
71-80). Hillsdale, NJ: Erlbaum.
Ryan, C., Vega, A., & Drash, A. (1985). Cognitive deficits in adolescents who developed diabetes early in life. Pediatrics, 75, 921-927.[Web of Science][Medline]
Ryan, C., & Williams, T. (1993). Effects of insulindependent diabetes on learning and memory efficiency in adults. Journal of Clinical and Experimental Neuropsychology, 15, 685-700.
Sansbury, L., Brown, R. T., & Meacham, L. (1997). Predictors of cognitive functioning in children and adolescents with insulin-dependent diabetes mellitus: A preliminary investigation. Children's Health Care, 26, 197-210.[Web of Science][Medline]
Wolters, C. A., Yu, S. L., Hagen, J. W., & Kail, R. (1996). Short-term memory and strategy use in children with insulin-dependent diabetes mellitus. Journal of Consulting and Clinical Psychology, 64, 1397-1405.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. C. Ferguson, A. Blane, J. Wardlaw, B. M. Frier, P. Perros, R. J. McCrimmon, and I. J. Deary Influence of an Early-Onset Age of Type 1 Diabetes on Cerebral Structure and Cognitive Function Diabetes Care, June 1, 2005; 28(6): 1431 - 1437. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||