Attention in children with insulin-dependent diabetes mellitus


Material Information

Attention in children with insulin-dependent diabetes mellitus
Physical Description:
Engel, Nicole A
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 28703674
oclc - 50543966
System ID:

Table of Contents
    Title Page
        Page i
        Page ii
        Page iii
        Page iv
    Table of Contents
        Page v
        Page vi
    List of Tables
        Page vii
        Page viii
        Page ix
        Page x
        Page xi
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
    Method and materials
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
        Page 159
        Page 160
        Page 161
        Page 162
        Page 163
        Page 164
        Page 165
        Page 166
    Biographical sketch
        Page 167
        Page 168
        Page 169
Full Text








This is dedicated in loving memory of my grandmother,
Sabina Knorr Engel (1905-1991), an unflagging supporter of
higher education who remembered the way the world was long
before cars and televisions were invented, and who always
marvelled at the accomplishments of science.


I would like to thank my chair, Eileen Fennell, Ph.D.,

for her ideas, support, and encouragement from the very

beginning of this research project to the very end. I also

would like to thank my cochair, Suzanne B. Johnson, Ph.D.,

for her sharing of her expertise in diabetes research and

for her careful reading of the drafts of this document. I

know that, as a result of their guidance, I have a better

eye for research, issues in diabetes, and statistics.

Additionally, I would like to express my gratitude to

Janet Silverstein, M.D., for her willingness to allow me to

collect data in the Pediatric Diabetes Clinic, and to the

Clinic staff for their assistance in this process. I also

would like to thank Rus Bauer, Ph.D., for being a supportive

and encouraging committee member, and Joyce Stechmiller,

Ph.D., for her willingness and enthusiasm to become involved

with this research project on short notice.

Furthermore, I am very appreciative of the Clark-Ryans

and of the Center for Pediatric Psychology Research for

their endorsement and generous financial support of this

research project. I also would like to thank my research


assistants, Sharmayn Sayers, Kimberly Harrell, and April

Barrow, for their enthusiastic help with data collection.

And last, but certainly not least, I would like to express

my appreciation to all the families who participated in this



ACKNOWLEDGEMENTS ........ ................... iii

LIST OF TABLES ........ .................... vii

ABSTRACT .......... ....................... ix

INTRODUCTION ......... ..................... 1

Insulin-Dependent Diabetes Mellitus .. .. ....... 1
Physical Complications of IDDM ...... .......... 2
Acute Hyperglycemia ....... ............. 3
Acute Hypoglycemia ....... ............. 4
Central Nervous System Complications of IDDM 4
Electroencephalographic Findings ... ...... 5
Evoked Potential Findings .. .......... 8
CNS Lesions ........ ................ 10
Neuropsychological Findings in Children with IDDM 11
Early Research (1922-1961) .. ......... 11
Current Research (1981-present) ........ 23
Acute Hypo- and Hyperglycemia and Performance 41
Summary of Neuropsychological Findings 47
Models of Attention ..... ............... 48
Brain Development ...... ................ 58
Brain Volume ...... ................ 59
Myelogenesis ...... ................ 60
Growth Spurts ...... ................ 61
Critical Periods ..... .............. 62

PURPOSE ........... ........................ 64

HYPOTHESES ......... ...................... 67


Subjects ......... ..................... 69
Assessment Instruments ..... .............. 74
Selective Attention .... ............. 76
Alternating Attention .... ............ 79
Sustained Attention .... ............. 81
Receptive Vocabulary and Academic Performance 81
Parent Report ...... ................ 83
Procedure ........ .................... 84
Recruitment ...... ................. 84
Blood Glucose Level .... ............. 85
Test Administration .... ............. 86
Test Order ...... ................. 87
Parent and Medical Information ....... 87

RESULTS ........... ........................ 89

Preliminary Analyses ..... ............... 89
Standardization of Attention Scores ..... 89
Normality ....... .................. 90
Relationship of Scores Within and Across
Attentional Domains .. .......... 91
Descriptive Analyses ..... ............... 93
Demographic Characteristics .. ......... 94
Diabetes Variables .... ............. 97
Testing Conditions .... ............ 99
Hypothesis 1: Performance on Measures of
Attention .................... 100
Analyses of Variance ... ........... 100
Chi-Squared Analyses ... ........... 102
Hypothesis 2: Predictors of Performance ..... ..103
Selective Attention .... ............ 106
Alternating Attention ... ........... 106
Sustained Attention .... ............ 107
Overall Mean of Composite Scores ..... 107
Hypothesis 3: Parent Ratings and Attentional
Ability ......... ..................... 107
Hypothesis 4: Achievement and Attentional Ability 109
SNAP-R Ratings and Achievement Scores ..... 113

DISCUSSION ......... ...................... 139

REFERENCES ......... ..................... 157

BIOGRAPHICAL SKETCH ....... .................. 167


Possible causes of CNS dysfunction in diabetes . .. 6

Demographics of Sample ...... ................ 70

Diabetes Characteristics ..... ............... 75

Tests of Attention ....... .................. 77

Correlation Matrix of Attention Scores .. ........ 114

Averages of Correlations Within and Between Domains 116

Correlations of Attentional Domain Composite Scores 116

Correlations of Age with Standardized Scores (z-scores) 117

Correlations of PPVT-R with Attentional
Composite Scores ...... ................. 119

Correlations of Diabetes Variables .. ......... 120
Means (Standard Deviations) of Attentional
Scores (z-scores) by Group ... ........... 121

Analysis of Variance Results of Domain Composite Scores 124

Chi-Square Results ...... ................. 128

Correlations of Age, Onset Age, and Disease Duration 132

Correlations of Disease Duration and Composite Scores
Controlling for Age ..... .............. 133

Hierarchical Multiple Regression .. .......... 134

Means (S.D.) of SNAP-R ADHD and Achievement Scores
by Group ........ .................... 136


Correlations between SNAP-R ADHD Scale and Composite
Scores Controlling for Age ... ........... 137

Correlations between Attention and Achievement Scores 137

MANOVA Results for Achievement Scores ......... 138


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Nicole A. Engel

December 1995

Chairperson: Eileen B. Fennell
Major Department: Clinical and Health Psychology

Attentional ability was examined in 50 children with

insulin-dependent diabetes mellitus (IDDM; 20 with early-

onset diabetes [EOD; < 6 yrs] and 30 with late-onset

diabetes [LOD; > 6 yrs]) and 50 age- and sex-matched

nondiabetic control subjects. Following current

neuropsychological theory which views attention as a

multidimensional construct, the domains of selective,

alternating, and sustained attention were assessed. Disease

variables including onset age, disease duration, and history

of seizures and ketoacidotic episodes were evaluated as

possible risk factors for poorer performance. Attentional

ability was also related to standardized achievement scores

and parent ratings of child attention on the SNAP-R.

The results showed that there were significant

differences in performance in the alternating domain based

on group membership. The children with diabetes obtained

significantly lower scores than the controls, and the EODs

obtained significantly lower scores than the LOD children

and the controls. Hierarchical regression showed that the

selective domain was predicted by a seizure by gender

interaction with boys with seizures performing worse than

girls with seizures. The alternating domain was predicted

by chronological age, while the sustained domain was

predicted by chronological age and metabolic crises. The

overall mean of the attentional domain scores was best

predicted by disease onset age (as a continuous variable)

and history of severe hypoglycemia.

Better performance in the selective domain of attention

significantly correlated with higher language achievement,

and better performance in the alternating domain

significantly correlated with higher scores in math

achievement. In language achievement, boys with diabetes

performed more poorly than girls with diabetes and the

control children. This finding remained even when the

higher number of school absences and the lower PPVT-R scores

for children with diabetes were taken into account.

Parent rating of poorer attention significantly

correlated with poorer performance in the selective domain,

but was not significantly related to group membership or to

the achievement scores.

These findings are consistent with the bulk of the

current research on cognitive functioning in children with

diabetes which shows a higher likelihood of cognitive

decrements and/or deficits in children with early disease

onset as compared to control children and children with

later disease onset. It is suspected that the underlying

mechanism for these differences is related to a greater

susceptibility in younger children to metabolic disturbance,

specifically severe hypoglycemia. In this study, as in

others, hypoglycemic seizures were more likely to have

occurred in children who acquired diabetes at a younger age

and in children who had had diabetes for a longer period of

time. This information has a number of implications,

especially for parents, clinicians, and educators who need

to take children's attentional ability and achievement into

account in their work and planning for the child's care and



Insulin-Dependent Diabetes Mellitus

Insulin-dependent diabetes mellitus (IDDM; also called

Type I Diabetes) is a common, chronic, metabolic disease

which can occur at any age, but is most commonly diagnosed

during childhood and adolescence (Tsalikian, 1990) and

occurs in more than 1 out of every 500 children under the

age of eighteen (Drash, 1986). IDDM has been found to be

more common in white children and in boys (Drash, 1986).

Children may also have non-insulin-dependent diabetes

mellitus (NIDDM), but this is rare (Tsalikian, 1990). This

form of diabetes is more commonly found in adults and will

not be examined in this study.

Insulin-dependent diabetes mellitus is caused by a

pancreatic disorder in which the beta cells of the

pancreatic islets are destroyed due to what is now thought

to be an autoimmune reaction (Eisenbarth, 1986). This

process is likely activated by environmental factors, such

as a virus, in genetically prone individuals (Cahill &

McDevitt, 1981). Consequently, there is a reduction or

deficit in insulin production resulting in decreased

carbohydrate utilization and increased lipid and protein

metabolism which creates an accumulation of glucose in the

blood stream and urine (Hensyl, 1990). Blood glucose levels

may be measured on a daily basis with a home monitoring

device using capillary blood. Long-term metabolic control

is assessed at doctor visits with glycosylated hemoglobin

(HA1C) values which provide an index of average blood

glucose levels over the preceeding 2-3 months (Holmes,


Diabetes is regulated through the use of exogenous

insulin, which was discovered in 1922 (Banting, 1922),

appropriate diet, and exercise (Rovet, Ehrlich, & Czuchta,

1990). Nevertheless, there are typically acute and chronic

physical, neurological, and neuropsychological sequelae due

to hyper- and hypoglycemia since maintaining perfect

metabolic balance over an extended period of time is

virtually impossible.

Physical Complications of IDDM

The general consensus has been that when IDDM is

properly controlled, there are few serious complications

until at least 5-10 years after diagnosis or around early to

mid-adulthood (Rovet, Ehrlich, & Czuchta, 1990). The more

common serious long-term complications which develop include

neuropathy, retinopathy, nephropathy, increased

susceptibility to infection, and generalized changes in the

vasculature predisposing those with diabetes to an increased

risk of stroke and cardiovascular problems (Berg & Linton,

1989; Hensyl, 1990). Additionally, limb amputations may

result from peripheral vascular occlusion, and blindness and

renal failure may result from microvascular changes

(Tsalikian, 1990).

These serious physical disorders are caused by chronic

hyperglycemia (Cahill, Etzwiler, & Freinkel, 1976; Drash,

1986; Hazlett, 1993). Because of this link, those with

diabetes, or their parents in the case of young children

with diabetes, are instructed to maintain strict metabolic

control to prevent episodes of hyperglycemia (Cahill et al.,

1976; Reich et al., 1990; Rovet, Ehrlich, & Czuchta, 1990).

This tight control leads to increased episodes of

hypoglycemia (Hazlett, 1993; Reich et al., 1990) which may

be treated through ingestion of carbohydrates.

Acute Hyperglycemia

Hyperglycemia may occur due to an inadequate amount of

insulin and/or the ingestion of an excessive amount of food.

Moderate hyperglycemia (bG = 300 mg/dl) may cause slowed

reaction time and memory impairment possibly because of

decreased nerve conduction velocity (Mooradian, 1988).

Severe hyperglycemia may cause hyperosmolarity (cerebral

edema; Mooradian, 1988) and a lack of insulin may cause

ketoacidosis (excessive accumulation of ketones, a by-

product of fat breakdown), both of which left untreated,

cause unconsciousness, coma, and eventually, death (Holmes,

1990; Rovet et al., 1988, 1990).

Acute Hviooclycemia

Hypoglycemia (bG less than 60 mg/dl, Reich et al.,

1990) may develop as a result of too much insulin and/or too

little food. Symptoms of hypoglycemia entail adrenergic and

neuroglucopenic responses. Adrenergic symptoms may include

tremulousness, weakness, palpitations, sweating, nausea, and

irritability. Headache, confusion, poor concentration,

slowed mentation and reaction time, and visual changes may

also occur. Seizures, followed by coma and death ensue with

severe hypoglycemia (bG below 15 mg/dl) without adequate

medical attention (Mooradian, 1988).

Central Nervous System Complications of IDDM

Confusion, slowed mentation, decreased reaction time,

and seizures, as described above, which accompany acute

hyper- and hypoglycemia are the most frequent and well-known

of the central nervous system (CNS) symptoms. Additionally,

there are a number of other possible changes in CNS

functioning as a result of short- and long-term changes in

blood glucose metabolism in diabetes. Acute hypoglycemia

has been documented as causing specific CNS impairments

including hemiplegia, cortical blindness, and color vision

dysfunction (Mooradian, 1988). The long-term CNS


complications associated with diabetes have received limited

acknowledgement likely because they are typically subtle in

nature (Mooradian, 1988). Mooradian (1988) outlines many

possible causes of CNS dysfunction in those who have had

diabetes for an extended period of time including changes in

brain metabolism, neurochemical changes, neuroendocrine

changes, alterations in the blood brain barrier, structural

(vascular) differences, and other influences including

uremic or hypertensive encephalopathy (see Table 1).

Moreover, abnormalities have been found in

electroencephalographic and evoked potential studies in

children and adults, and actual CNS lesions have been

reported in adults with diabetes. These findings are

discussed below.

Electroencephalographic Findings

A small number of electroencephalographic (EEG) studies

have been done outside the United States on children with

IDDM (e.g., Eeg-Olofsson & Petersen, 1966; Halonen,

Hiekkala, Huuponen, & Hakkinen, 1983; Haumont, Dorchy, &

Pelc, 1979) and only six of them are reported in English

language journals. All the EEG studies have reported that

children with diabetes have significantly more EEG

abnormalities than normal subjects (Halonen et al., 1983;

Haumont et al., 1979). The percentage of abnormalities

found varies greatly from study to study from a high of 76%

(Heik, Schadlich, & Warnle, 1962 as cited in Haumont et al.,

Table 1

Possible Causes of CNS Dysfunction in Diabetes

A. Neurological causes

1. Degeneration of ganglion cells

2. Demyelinization of axons

3. Loss of axons

B. Vascular cause

1. Cerebrovascular accidents

2. Alterations in BBB transport carriers

3. Impaired BEE function

C. Metabolic causes

1. Hyperglycemia/hyperosmolarity

2. Hypoglycemic reactions

3. Ketosis

4. Acidosis

5. Hypoxia

6. Electrolyte abnormalities

7. Neurobiochemical changes

8. Neuroendocrine changes

D. Other contributing causes

1. Peripheral and autonomic neuropathy

2. Hypertension

3. Renal failure

from Mooradian (1988).


1979) to a low of 19% (Laron, Karp, & Freinkel, 1972, as

cited in Haumont et al., 1979) with as many abnormal values

above 50% as below. In a normal population, 10-15% of EEG's

have nonspecific abnormalities (Haumont et al., 1979). The

abnormalities reported in children with diabetes when

compared with controls included significantly more positive

spike-potentials (Eeg-Olofsson & Petersen, 1966),

significantly more non-rhythmic slowing and paroxysmal

activity (Eeg-Olofsson, 1977; Haumont et al., 1979), and

significantly lower alpha frequency and amplitude (Eeg-

Olofsson, 1977; Eeg-Olofsson & Petersen, 1966).

Most of the studies examined the relationship of some

diabetic variables to EEG abnormalities. Only one of the

eleven studies reported a significant relationship between

EEG abnormalities and age or duration of illness in children

with diabetes (Lerman et al., 1977, as cited in Haumont et

al., 1979). However, all the studies indicated that a

history of hypoglycemic seizures and/or unconsciousness were

significantly related to the presence of abnormalities on

the EEG examination (Halonen et al., 1983; Haumont et al.,

1979). A history of mild hypoglycemia was related to

unspecified EEG abnormalities in one of the five studies

which examined this variable (Abramowicz, Margolis, &

Wawrzynkiewicz, 1969, as cited in Haumont et al., 1979), as

was a history of diabetic ketoacidosis in the only study

which examined this variable (Halonen et al., 1983).

All these studies were performed by researchers in

countries other than the United States and all were

conducted a decade or more ago. These factors raise the

question of whether the results can be generalized to

children with diabetes who have had medical care in the

United States within the last decade. Nevertheless, the

consistency of the finding that children with diabetes are

significantly more likely to have abnormalities in their

EEGs, especially if they have a history of severe

hypoglycemic episodes, is compelling.

Evoked Potential Findings

Evoked potentials have been used as a method of early

detection of the development of nerve anomalies with

visually evoked potentials being used to detect

demyelinization of optic pathways (Cirillo et al., 1984).

Few evoked potential studies have been done with children

with diabetes, however. Cirillo et al. (1984) compared 30

children with diabetes (7-22 years old) to 28 healthy

controls and found a significant difference in the latency

of the P100 wave of visually evoked potentials with subjects

with diabetes responding more slowly. Closer examination of

this finding revealed that it was due to a gender

difference. Although the latency time for girls with

diabetes was not significantly different from that of the

boys with diabetes, the girls with diabetes had a longer

mean latency than the controls. No significant correlations

were found between the latency and age, duration of

diabetes, or mean glycosylated hemoglobin.

The results of auditory evoked potential studies done

with adults with IDDM and healthy controls support the

findings of the child study. Dejgaard et al. (1991)

reported abnormal brain stem auditory evoked potentials in

40% of those who had had diabetes for a long duration

(duration = 13-46 years; age = 25-66 years) as compared with

5% in those who had had diabetes for a short duration

(duration = 0-6 years; age = 18-50 years). Other studies

found significant differences in a number of auditory brain

stem evoked response values with subjects with IDDM having

longer latencies than age- and sex-matched controls (Donald

et al., 1984; Fedele et al., 1984). Results suggested that

the dysfunction was located in the central auditory

projections and not in the acoustic nerve (Donald et al.,

1984). Neither study found correlations between impaired

responding and disease duration, metabolic history, or

peripheral neuropathy (Donald et al., 1984; Fedele et al.,

1984). Donald et al. (1984) suggested that the increased

latencies in brain stem response emerged early in the course

of diabetes. Evidence to support this finding may be found

in a study of normals (age = 20-24 years) who manifested

changes in auditory brain stem and cortical evoked

potentials during induced mild hypoglycemia (Jones et al.,


CNS Lesions

CNS lesions have been described in those with chronic

diabetes (Dejgaard et al., 1991) and have been reported as

directly related to severe acute hypoglycemia (Chalmers et

al., 1991). Wredling, Levander, Adamson, and Lins (1990)

described the mechanism for these lesions as being a result

of the brain's decreased ability to utilize oxygen during

hypoglycemia despite normal arterial oxygen tension. This

impairment may create a hypoxic condition during which

neuronal death (brain damage) may occur. The hippocampus

and the frontal lobes, in addition to other brain areas,

have been implicated as being sensitive to blood glucose


Chalmers et al. (1991) described a man with diabetes

who experienced an episode of severe hypoglycemia which

resulted in coma, and, following resolution of the coma,

significant memory deficits. Magnetic resonance imaging

(MRT) documented lesions of the left hippocampus and the

left inferomedial temporal lobe. While Erickson (1990)

portrayed the hippocampus as being particularly sensitive to

the effects of hypoglycemia to the extent that recurrent

episodes may lead to an amnestic syndrome, Chalmers et al.

(1991) reported that "usually, there is a distribution of

lesions involving principally the cerebral cortex,

hippocampus, and basal ganglia" (p. 924). However, Wredling

et al. (1990) suggested that patients with recurrent severe

hypoglycemia may have bilateral frontal lobe dysfunction.

Holmes (1990) offered evidence in support of frontal lobe

damage following episodes of acute hypoglycemia based on

neuropsychological testing and cerebral blood flow studies.

Clearly, this is an area which deserves further


Lesions have also been described in other areas of the

brain. MRI results revealed from 1 to 10 lesions (average

5) in eleven of sixteen (69%) patients with diabetes

considered to have a long disease duration (13-46 years).

These lesions were 2-10 mm in size and were distributed

throughout the cortex, cerebellum, and the brain stem. In

contrast, only 12% of the healthy age-matched controls had

lesions, and the lesions were less frequent, smaller, and

located only in the cerebral hemispheres (Dejgaard et al.,

1991). Evidence was not available to determine whether the

etiology of these lesions was due to severe hypoglycemia,

cerebrovascular accidents, or another cause.

Neuropsychological FindinQs in Children with IDDM

Early Research (1922-1961)

Early research on the cognitive functioning of children

with IDDM is comprised of less than a dozen studies which

appeared in journals during a forty year period between 1922

and 1961. The very first studies examining cognitive

functioning in subjects with diabetes were done around the

same time period that exogenous insulin was first isolated

(Miles & Root, 1922; Dashiell, 1930). Miles and Root (1922)

tested 40 subjects with diabetes who were "free from

acidosis" after breakfast making the possibility of

hypoglycemia unlikely. They found that the subjects with

diabetes were impaired in a number of areas including

memory, attention, motor speed, and reaction time compared

to controls who did not have diabetes. Dashiell (1930)

tested a medical student with diabetes during what they

presumed to be periods of euglycemia and hypoglycemia. He

found slowed reaction time, decreased grip strength, and

slowed ability to name colors as well as reported declines

in memory and attention during periods of hypoglycemia.

Apparently, these studies were overlooked (Ryan, 1990)

because subsequent studies on children examined only global

intelligence until the 1980s when researchers (e.g., Rovet,

Gore, & Ehrlich, 1983; Ryan, Vega, & Drash, 1981) again

examined a range of cognitive functions in children with


The studies completed prior to the 1980s primarily

utilized a version of the Stanford-Binet to measure

intelligence. Their conclusions were contradictory as they

declared that children with IDDM were either of higher than

average intelligence (e.g., Grishaw, West, & Smith, 1939),

of lower than average intelligence (e.g., Shirley & Greer,

1940), or of average intelligence (e.g., Ack, Miller, &

Weil, 1961) depending on the study.

A few early researchers concluded that children with

diabetes were of higher than average intelligence (Grishaw

et al., 1939; Joslin, 1935, 1937 as cited in Brown &

Thompson, 1940, Ryan, 1990, and Teagarden, 1939; West,

Richey, & Eyre, 1934, and White, 1932 as cited in Kubany et

al., 1956 and Teagarden, 1939). The assumption as to why

children with diabetes would be more intelligent was based

on the idea of compensation. This argument stated that, as

a result of their physical limitations, children with

diabetes focused more on intellectual pursuits, thus

developing themselves in this area (Ack et al., 1961).

Grishaw et al. (1939) presented data collected on 341

children with diabetes which he and his colleagues had

treated in private hospital based practices between 1920 and

1938. Of this group, 62 "unselected" children were given

the Stanford-Binet. The mean of the IQ scores was 105.4

(SD=ll.7) Forty-two percent of the children obtained IQs

above 110, 47.4% of the children obtained IQs between 90 and

110, and 10.5% obtained IQs below 90. In comparing these

results with Terman's normative sample for the Stanford-

Binet (20.6% above 110, 60% between 91 and 110, 19.4% below

90), the authors concluded that children with diabetes were

"well in advance of the normal distribution" (p. 799).


According to Teagarden (1939), Joslin (1937) and White

(1932) described the same 169 children with diabetes who

were patients at a specialty clinic in Boston and likely of

a higher SES (Brown & Thompson, 1940; Kubany et al., 1956;

Ryan, 1990). Teagarden (1939) reported that Joslin and

White's findings were that 15% of the children had IQs below

90; 55% had IQs between 90 and 110; and 30% had IQs of 110

and over. As there was a greater percentage of children in

the higher IQ range than in Terman's sample, Joslin and

White concluded that children with diabetes possessed higher

than average intellectual abilities (Teagarden, 1939).

Teagarden (1939) also described a study of the

intelligence of 76 children with diabetes conducted by West,

Richey, and Eyre (1934). Reportedly, 9.1% had IQs below 90;

47.3% had IQs between 90 and 110; and 43.6 had IQs above

110. These results lead the authors to conclude that

children with diabetes were mentally precocious (Teagarden,


In contrast to the above findings for higher IQs in

children with diabetes, other researchers suggested that

children with diabetes were actually of lower than average

intelligence (Boulin, 1950; Shirley & Greer, 1940;

Teagarden, 1939). Teagarden (1939) postulated that this may

be due to a combination of coma, "insulin shock"

(hypoglycemia), and restricted diet, all of which may affect

mental processes and possibly lead to neurological damage.

Teagarden (1939) described six case studies of children

ranging in age from 20 months to 13 years who had diabetes.

Intelligence testing revealed IQs of 45, 64, 73, 85, 95, and

128 for each of the children, respectively. Teagarden

stated that no generalizations could be made because these

results were based on case studies. An additional reason

for not generalizing to other children with diabetes is that

the children she described had other significant factors in

their histories which likely impacted on their IQ scores far

more than diabetes alone would have. Included in the

children's histories were: severe developmental delay,

generalized endocrine dysfunction with an indication of

possible neurological problems, macrocephaly and chorea in a

child who had had numerous febrile illnesses, and the fact

that one child came from a French speaking home.

Nevertheless, this study has been cited in at least two

subsequent articles as support for the conclusion that

children with diabetes may have lower than average IQs (Ack

et al., 1961; Kubany et al., 1956).

The study by Shirley and Greer (1940) also supported

the conclusion that children with diabetes possessed less

than average intellectual abilities. They studied 155

children with diabetes who were referred to a department of

psychiatry. The children ranged in age from below 5 to 18

years old and the majority were considered to be indigent

cases. In comparing the distribution of IQs they obtained

to those in Terman's standardization sample, Shirley and

Greer indicated that the children with diabetes were of

slightly lower intelligence than average. Kubany et al.

(1956) cited Boulin (1950), a French researcher, as also

describing children with diabetes as having lower than

average intelligence.

The majority of the studies examining general

intelligence level in children with diabetes determined that

their intelligence was normally distributed (Ack, Miller, &

Weil, 1961; Brown & Thompson, 1940; Fischer & Dolger, 1946;

Kubany, Danowski, & Moses, 1956; McGavin, Schultz, Peden, &

Bowen, 1940). McGavin et al. (1940) included 49 children

with diabetes of varying socioeconomic backgrounds in their

sample to study physical growth, intelligence, and

personality. According to the researchers, the majority of

the children were healthy except for having IDDM. The

children were given the Revised Stanford-Binet Intelligence

Scale (Terman & Merrill, 1937) in addition to achievement

and personality inventories. Their results yielded a mean

intelligence quotient of 103 (SD = 16.8) and a distribution

of IQ scores which they described as being similar to that

obtained by Terman and Merrill (1937) in the standardization

sample of the test. McGavin et al. (1940) concluded that

the children with diabetes were "not significantly brighter

or duller" (p. 122) than the standardization sample.

Brown and Thompson (1940) reached the same conclusion

as McGavin et al. (1940) based on the results from their

sample of 60 subjects with diabetes ages 22 months to 20

years. The authors asserted that the varying socioeconomic

backgrounds of the subjects was representative of the

general Minnesota population from which the sample drawn.

For controls, 28 slightly older healthy siblings of the

children with diabetes were included. Results from the

Revised Stanford-Binet Intelligence Scale reportedly showed

that the group with diabetes had normally distributed IQ

scores. Brown and Thompson reported no relationship between

duration or severity of diabetes with regards to

intelligence scores. Based on parent and teacher reports,

they found the ability and achievement of children with

diabetes to be similar to that of their siblings. The

authors also commented that the IQs obtained by the children

with diabetes showed "no deviation from the average and no

significant deviation from that of their sibling controls or

from the average of Minneapolis children" (p. 254). It is

not clear whether the lack of significant deviation between

the diabetic children's mean IQ score of 100.65 and their

siblings' mean IQ score of 107.00 was determined

statistically or by clinical judgement. Additionally, no

mention of the relationship between disease onset to

intelligence was made despite the opportunity to do so given

that one third of the sample was comprised of children who

developed diabetes prior to 5 and 1/2 years of age.

In a longitudinal study of children with diabetes,

Fischer and Dolger (1946) followed 43 outpatients with

diabetes from childhood through adolescence and early

adulthood. Most of the children were from low or lower

middle class homes. Of the twenty-four outpatients that

were given the Revised Stanford-Binet-Simon Intelligence

Test, 20 earned IQs of 90 or above. As these results were

comparable to the scores obtained in some unspecified past

testing with patients without diabetes, but of the same

ages, the authors inferred that the results "demonstrated

that diabetics were neither 'brighter' nor 'duller' than

normals" (p. 725). Ten of the subjects with diabetes were

retested after an unspecified number of years. Overall, the

second set of IQ scores was deemed to be not significantly

different than the previously obtained scores.

In their study of intelligence and psychopathology in

those with diabetes, Kubany et al. (1956) used a sample of

40 children and young adults with diabetes with the average

age of 18.8 years who had been outpatients at a hospital

clinic. The authors stated that the sample was

representative of the socioeconomic range of the children

with diabetes treated at the hospital. For intelligence

testing, 18 adults with diabetes were added to the sample

described above yielding a total sample of 58 subjects with


diabetes who ranged in age from 7 to 63 years with a mean of

20.8 years. As this sample obtained the mean IQ score of

103 on the Stanford-Binet, Kubany et al. (1956) concluded

that there were no differences in IQ between children with

diabetes and normals.

Ack et al. (1961) used 38 children with diabetes

ranging in age from 3 years, 1 month to 18 years, 6 months

with a mean age of 10 years and compared them to sibling

controls. The children were outpatients of a university

hospital or private patients of one of the authors and were

considered to represent a range of socioeconomic levels.

The subjects were administered the Stanford-Binet

Intelligence Scale, Form M. The difference in IQ between

the child with diabetes and his/her sibling (diabetic IQ

minus sibling IQ) was used in the statistics, but no direct

comparisons were made. Ack et al. concluded that there were

no significant differences between the groups as the overall

IQ differences were not significantly different from zero.

However, the results indicated that children who developed

diabetes before age 5 had IQs significantly lower than their

siblings (mean difference score = -10.15) as compared with

the IQ differences between those children who acquired

diabetes after age 5 and their siblings (mean difference

score = +0.72). The number of severe hypo- and

hyperglycemic episodes (i.e., those requiring emergency room

treatment or hospitalization) tended to be related to lower


IQs in the children with earlier onset, but this result was

not found to be statistically significant. Finally, no

relationship was reported between difference scores of IQs

and the duration of the disease.

As Ryan (1990) delineated in his book chapter, there

are a number of methodological problems associated with

these past research studies which cast doubt on the validity

of their conclusions. These methodological concerns include

sampling biases (Ack et al., 1961; Kubany et al., 1956;

Ryan, 1990), lack of adequate control subjects, and an

inattention to disease variables such as duration of

illness, degree of metabolic control, and age at onset (Ack

et al., 1961; Ryan, 1990). Additionally, the statistics

employed were largely descriptive (Ryan, 1990). Sampling

biases in the socioeconomic statuses of the subjects being

examined was present in the majority of the studies cited

above. Socioeconomic status is an important variable to

consider when examining intelligence as it is significantly

correlated with IQ scores (Sattler, 1988). Grishaw et al.

(1939), Joslin (1937), West et al. (1934), and White (1932),

who found subjects with diabetes to be of higher than

average intelligence, selected their subjects from private

practices. As these subjects would tend to be of a higher

socioeconomic class and better educated, the finding of

above average intellectual ability would be expected (Ryan,

1990). Additionally, Shirley and Greer (1940) sampled


indigent patients who were psychiatric referrals and likely

biased their sample toward lower IQs (Ryan, 1990).

Furthermore, those researchers who included subjects who

were considered to be representative of the larger

population from which they were drawn (Ack et al., 1961;

Brown & Thompson, 1940; Kubany et al., 1956; McGavin et al.,

1940) determined that the intellectual abilities of those

with diabetes were normally distributed. Again, according

to Ryan (1990), this finding could have been anticipated

based on the relationship of socioeconomic status and

intelligence. The socioeconomic status of the subjects

studied did not need to be representative of the general

population if control subjects were used. However, only

three of the studies used controls: one had controls which

were aged-matched only (Fischer & Dolger, 1946), and two

used sibling controls (Brown & Thompson, 1940; Ack et al.,

1961). These studies all concluded that children with

diabetes were of average intelligence.

In addition to the over-reliance on comparisons to

normative samples, the early research on cognitive

functioning in children with diabetes as a whole did not

examine disease variables such as duration of illness,

degree of metabolic control, and age at onset. In

neglecting to investigate these factors, early researchers

were implying that they considered the children with

diabetes to be a homogeneous group when this is known not to

be the case. Only two studies analyzed diabetic variables.

Brown and Thompson (1940) found duration of disease and

degree of metabolic control not related to IQ scores;

however, they questioned the reliability of the reporting of

metabolic control. Ack et al. (1961) examined not only

duration of disease and degree of metabolic control, but

also analyzed the relationship of onset age with performance

on the Stanford-Binet finding only onset age to be

significantly related to IQ.

An additional concern regarding these early studies,

according to Ryan (1990), is the lack of sophistication of

the statistics. Only Ack et al. (1961) reported statistics

more complicated than means and percentages. However, as

Ryan (1990) pointed out, this study neglected to directly

compare the mean IQ of the children with diabetes to the

mean IQ of their sibling controls, opting instead to examine

IQ difference scores in each of the 38 subject pairs

separately. As using difference scores truncates the range

of values, the results could be influenced by the phenomenon

of regression to the mean. Because of these methodological

problems and contradictory results described above, Ryan

(1990) concluded that these early studies provided no

substantive information on the cognitive functioning of

children with diabetes.

Current Research (1981-present)

Based on the earlier studies described above,

researchers generally concluded that children with diabetes

did not differ from those without diabetes with regards to

intellectual functioning (Ryan, 1988). It was not until the

1980s when the study of cognitive effects of diabetes in

children reemerged as a research question. Ryan and his

colleagues led the resurgence in examining the cognitive

functioning of children with IDDM after reconsidering the

topic in light of neurological (e.g., Eeg-Olofssen &

Petersen, 1966) and adult (e.g., Bale, 1973) findings

suggesting brain dysfunction and impairments in those with


Ryan, Vega, and Drash (1981) examined the performance

of 50 adolescents (mean age = 15) on a battery of

neuropsychological tests which assessed language skills,

attention, memory, problem-solving, visuospatial skills, and

motor coordination. The subjects comprised one of three

subgroups: children with IDDM who developed the disease

prior to age 3, children with IDDM who developed the disease

after age 3, and children without diabetes. The number of

subjects in each subgroup was not reported. Adolescents who

developed diabetes before age three were significantly

impaired on nearly all the tests compared with controls. In

contrast, the children with diabetes who were diagnosed

after age 3 were virtually unimpaired, and performed

similarly to the controls. Unfortunately, neither the

measures utilized nor the definition of "impaired" was

specified, nor was information provided on which tests the

adolescents with diabetes were impaired.

Ryan, Vega, Longstreet, and Drash (1984) assessed

several aspects of neuropsychological functioning in 40

adolescents with IDDM and 40 demographically similar

controls. The focus of the study was to determine any

patterns of neuropsychological impairment in the adolescents

with IDDM, and to learn if there was a relationship between

metabolic control and test performance.

The WISC-R or WATS subtests of Information,

Comprehension, Similarities, Digit Span, Vocabulary, Picture

Completion, and Block Design assessed verbal and nonverbal

intelligence. Associative learning and memory were assessed

by the Symbol-Digit Paired-Associate Learning Test, the

Verbal Paired-Associate Learning Test, Four-Word Short-Term

Memory Test, and the Wechsler Memory Scale (WMS) Visual

Reproductions. Visuospatial ability was measured by the

Boston Embedded Figures Test, the Hooper Visual Organization

Test, and the Road Map Test. Speed, dexterity, and

visuomotor integration was assessed by the Digit Symbol

Substitution Test, the Trail Making Test (TMT), and the

Grooved Pegboard Test. Self-concept was assessed with the

Piers-Harris Children's Self Concept Scale. Critical

flicker frequency was included as it was thought to be

sensitive to diffuse brain dysfunction.

The results showed that while the subjects with

diabetes achieved scores within normal limits on all the

tests, they performed significantly lower on verbal

intelligence (VIQ), fine motor coordination (Grooved

Pegboard Test), critical flicker threshold, and on tests

requiring attention and visuomotor coordination (Digit

Symbol Substitution Test, the TMT, Part B) No significant

differences were found between the subjects with diabetes

and controls on visuospatial abilities, verbal and nonverbal

memory, or self-concept and anxiety. Furthermore, duration

of disease and metabolic control (HA1C) were not found to

significantly correlate with the lower test scores.

In a study comparing 125 adolescents with IDDM to 83

demographically similar controls, Ryan, Vega, and Drash

(1985) assessed multiple areas of cognitive functioning.

The subjects were Caucasian, between the ages of 10 and 19,

and had had diabetes for at least 3 years. The group with

diabetes was subdivided into early (<5 years old) and late

( 5 years old) disease onset groups following Ack et al.

(1961). Virtually the same tests were used as in the

previously described study with the addition of Digit Span

from the WISC-R or WAIS to assess attention and the Wide

Range Achievement Test (WRAT) to assess school achievement.

The results revealed that the adolescents with early

onset diabetes (EOD) performed more poorly on all the

measures. The adolescents with EOD scored significantly

lower than the control subjects on the WISC-R or WAIS verbal

subtests of Information, Vocabulary, and Comprehension.

They scored significantly lower than both the adolescents

with late onset diabetes (LOD) and the controls on WRAT

reading and on tests requiring visuospatial and

visuoconstructional abilities including tests of visual

memory (Block Design, Embedded Figures, Road Map, immediate

and delayed Visual Reproduction). Additionally, adolescents

with EOD performed more poorly on tests of attention and

motor speed (Digit Span, Digit Symbol Substitution, Grooved


Despite these significant differences, the mean scores

for all the tests for each of the three subject groups were

within the normal range. Because of this, cutting scores at

2 standard deviations below the mean were established. A

subject was considered to be "clinically impaired" if they

performed below the cutting score on three or more of the

tests. These additional analyses revealed that 24% of the

adolescents with EOD as compared with only 6% of the

adolescents with LOD and 6% of the controls were judged to

be clinically impaired. Through multiple regression, onset

age was found to best predict performance on nonverbal

tests; whereas disease duration best predicted scores on


verbal tests. A history of hypoglycemic seizures, which was

assessed in a dichotomous manner (presence/absence), was not

found to be significantly relate to the results. Despite

this finding, Ryan et al. (1985) hypothesized that a history

of hypoglycemia was related to the poorer scores of the

adolescents with EOD. They cited as support the "well-

known" finding that CNS impairments were related to a

greater incidence of seizures (hypoglycemic and

nonhypoglycemic) especially when the seizures occurred at a

younger age (p. 926).

Rovet, Gore, and Ehrlich (1983) were interested in

determining if the cognitive deficits found in adolescents

who developed diabetes early in life also were present in

younger children. Their study compared children with EOD (N

= 19), children with LOD (N = 12), and sibling controls (N =

16) in children between 6 and 12 years old. They classified

early onset as developing diabetes before the age of three

and late onset as developing diabetes after the age of

three. The measures used were not specified. Differences

in performance were found to be related to onset age and to

gender. Their results yielded lower Performance IQ (PTQ)

and perceptual-motor scores for EOD girls, but no

significant differences in scores between EODs, LODs, and

controls for boys. Additionally, PIQ correlated with

duration of disease in EODs, but not in LODs. Reportedly, a

history of hypoglycemic convulsions was not found to be

significantly related to the lower scores in EODs.

Rovet, Ehrlich, and Hoppe (1987, 1988) assessed both

intellectual performance and school achievement in an

attempt to replicate Ryan et al. (1985) with a younger

sample. Their subjects included 27 children with EOD (< 4

years old), 24 children with LOD (> 4 years), and 30 sibling

controls. The children were between the ages of 6 and 13

with a mean of 9.8 years of age.

The WTSC-R subtests of Similarities, Vocabulary, Block

Design, and Object Assembly were given to all the children;

the children with diabetes also received the subtests of

Information, Picture Completion, Coding, and Digit Span.

Full Scale IQ (FSIQ), VIQ, and PIQ were estimated for each

subject. It is of concern that the IQ scores for the

control subjects were estimated from four subtests. While

the FSIQ score can be estimated from the Vocabulary and

Block Design subtests alone, the fewer the number of

subtests used in this estimate, the less reliable and stable

the IQ is (Sattler, 1988). The Primary Mental Abilities

(PMA) verbal meaning and spatial reasoning subtests and the

Beery-Buktenica Developmental Test of Visual Motor

Integration (VMI) were also given to assess verbal and

nonverbal abilities. The WRAT assessed school achievement.

The results showed that the children with EOD performed

significantly lower than the children with LOD on Picture

Completion and Coding. These subtests were not given to the

controls, so no comparison could be made to their

performance. Reportedly, the EODs obtained significantly

lower PIQs than the LODs or the controls. However, since

all the IQ scores were estimated, and the PIQ scores of the

controls were derived from only two subtests, this result

should be viewed with caution. As with Rovet et al. (1983),

they reported gender differences in their results. EOD

girls obtained lower PIQ and Block Design scores; EOD girls

and LOD boys obtained lower Object Assembly scores; and all

boys obtained lower scores on the Coding subtest and on the

WRAT reading and spelling subtests.

Overall, the children with diabetes, especially the EOD

girls, were more likely to have difficulty at school (e.g.,

failing a grade, being in special education, academic

difficulties) than the controls. With regard to metabolic

control, the EODs, especially the girls, were most likely to

have had hypoglycemic seizures. Both onset age and a

history of seizures, independent of gender, were found to be

significantly associated with poorer performance on

visuospatial tasks (the PMA spatial test, WISC-R Picture

Completion) and WISC-R Coding. Duration of disease and

history of ketoacidosis were not significantly correlated

with neuropsychological performance.

Rovet, Ehrlich, and Czuchta (1990) examined

neuropsychological functioning in children with diabetes at

diagnosis and one year later. Prior research had been

cross-sectional and retrospective, therefore, as Rovet et

al. (1990) pointed out, it was not possible to determine

whether the lower scores found on neuropsychological tests

of children with diabetes were a result of their disease or

existed prior to diagnosis. Their prospective study

intended to ascertain whether children with diabetes have

deficits at the time of diagnosis, in addition to

identifying the risk factors associated with impairments

should any be found.

Their sample consisted of 63 newly diagnosed children

with diabetes and 40 siblings without diabetes. The

children ranged in age from 5 months to 12.5 years with the

mean age of 7.3 years. The children with diabetes diagnosed

prior to age 5 were placed in the early onset group whereas

those diagnosed later comprised the late onset group. The

cognitive domains examined were intelligence, verbal and

spatial abilities, memory, and achievement. The measures

given included the Griffiths, the WPPSI, the WISC-R, the

Reynell, the VMI, the PMA Spatial, the Woodcock-Johnson

Spatial Relations, and the WRAT depending upon the age of

the child.

Testing revealed that none of the children displayed

any evidence of deficits and that, at diagnosis, the

children with diabetes did not significantly differ on any

of the measures except for the WISC-R Similarities subtest.

(Their performance was superior to their siblings'.) Only

the children with diabetes were retested. One year later,

significantly improved performances for them were found on

Griffiths Hearing and Speech, WISC-R Digit Span, Woodcock-

Johnson Spatial Relations, and WRAT Reading. A significant

decline was documented on the WISC-R vocabulary although the

mean performance was still within the average range.

Overall, younger children improved in their verbal

functioning while older children declined in this area. No

clarification was provided to indicate whether the

improvements were above the normal advances in ability

expected with age and whether the decreases in performance

reflected an actual loss of ability.

With respect to diabetic variables, EODs tended to have

lower PIQ scores than the LODs at diagnosis and one year

later. LODs had significantly lower VIQ scores than EODs

one year after diagnosis. A history of ketonuria and

hospitalizations were significantly associated with negative

spatial change scores (score at one year minus score at

diagnosis). Furthermore, a trend was noted for children who

were ketoacidotic at diagnosis to perform somewhat more

poorly on spatial tasks one year after diagnosis. As the

subjects' age range spanned from 5 months to 12.5 years, the

relationships between the EODs and LODs must be viewed with

discretion because direct comparisons of abilities were not

possible. This age range meant that the children were

assessed by different measures and in different cognitive

domains depending on their age.

Holmes and Richman (1985) also assessed cognitive

abilities and academic achievement in their sample of 42

children with diabetes. The children, who ranged in age

from 6 to 16, were given the WISC-R (all subtests except

Digit Span and Mazes) and word list reading from either the

WRAT or the Standard Reading Inventory. Children with

evidence of reading impairment, defined by a reading score

of 2 years or more below their grade level, were given

additional tests. These included the Bender Visual Motor

Gestalt Test, the Rey Auditory Verbal Learning Test (RAVLT),

and the WISC-R Digit Span.

In order to analyze the results, the children were

grouped according to onset age (early onset = < 7 years old;

late onset = > 7 years old) and duration of illness (short =

< 5 years; long = > 5 years). Ultimately, four groups

resulted to examine interaction effects: early onset-short

duration, late onset-short duration, early onset-long

duration, late onset-long duration. The results revealed

significantly lower PTQ scores for the early onset-long

duration group compared with the other groups. Slower

performances on the timed items were hypothesized as

contributing to the lower scores. Further analyses

suggested that onset age was a more potent predictor of

performance than duration. With regard to reading

achievement, those children with early onset were

significantly more likely to be classified as impaired than

those in the late onset group (16/22 vs. 6/20). The

children classified as "reading impaired" also tended to

have impaired memory for words and digits. However, since

no information was gathered on memory functioning in the

non-reading impaired group, this finding cannot be fully

interpreted with regard to diabetic variables. No measures

of metabolic control were taken into account.

Anderson et al. (1984) examined cognitive abilities in

addition to scholastic achievement and variables related to

school performance. Their sample consisted of 15 EODs

(onset < 4 years old), 15 LODs (onset > 4 years old), and 30

matched controls. Their performance on cognitive,

achievement and school-related variable, memory,

neurological, and perceived self-competence measures were

evaluated. The specific tests were not described. Although

the overall findings suggested that there were no

significant differences between children with diabetes and

controls in the area of cognitive functioning, "modest

deficiencies" on subtests of the WISC for the EODs were

reported. The nature of these deficits and on which

subtests they occurred was not specified, however. EODs

were found to have significantly more difficulty at school

in areas including: "remedial services, repeating grades,

absences, and problems with teachers" (p. 21A) However,

these children demonstrated confidence in their intellectual

and social abilities. Furthermore, the EODs had more

difficulty with hypoglycemia according to medical records

which documented more convulsions and seizures for this

group. Unfortunately, this finding was not analyzed with

regard to test performance.

These researchers elaborated on their findings in a

later journal article (Hagen et al., 1990). The measures

utilized included the WISC-R Vocabulary, Information,

Comprehension, Digit Span, and Block Design subtests; the

reading comprehension and mathematics subtests from the

Peabody Individual Achievement Test (PIAT); two unnamed

memory tasks, one involving a word list and one involving a

visual-pictorial task; the Harter Perceived Competence Scale

for Children; the Conners Behavioral Symptom Scale (CBSS).

Hagen et al. (1990) reported that all the children with

diabetes obtained lower estimated WISC-R IQ scores than the

controls, and that the children with diabetes performed

significantly lower on Information and Vocabulary.

Additionally, EODs performed significantly lower than

controls on Block design and Digit Span. On the PIAT

reading comprehension section, the LODs performed more

poorly than the controls. While none of the groups differed

significantly on their overall performance on the two memory

tasks, differences were found with regard to memory

strategies, even when using age as a covariate. EODs

obtained higher scores on incidental memory suggesting less

focused attention. Moreover, evidence appeared that EODs

had a reduced primacy effect compared to LODs and controls.

Parent report on the CBSS showed significantly more symptoms

of inferior concentration and problems in completing tasks

for the EODs.

Furthermore, EODs experienced significantly more

metabolic disturbances than LODs, having greater incidence

of hypoglycemia convulsions and seizures with loss of

consciousness and/or hospitalizations. Although the results

are similar to those reported in other studies, they must be

viewed with some caution as 20% of the EODs reportedly

sustained a "serious" head injury compared with none ("0%")

of the LODs and 7% of the controls. No specific information

was given as to the nature of the head injuries. Research

on head injuries suggests that the major effects of closed

head injuries include memory and attentional problems

(Fennell & Mickle, 1992).

The relationship of metabolic control to cognitive

functioning was the focus of the study by Golden et al.

(1989). Twenty-three children ranging in age from 3-9 years

old, all of whom developed IDDM before age five, were given

the Stanford-Binet Intelligence Scale: Fourth Edition

(SB:FE). Testing was done immediately after a meal which

followed a documented blood glucose level above 100 mg/dl,

presumably to control for the effects of hypoglycemia.

Information was collected on HAjC level, the number of

severe hypoglycemic episodes (defined as a change in or loss

of consciousness), and the percentage of asymptomatic

hypoglycemic episodes (< 50 mg/dl) which were determined

from self-monitoring over the past several months.

The results revealed average intellectual abilities

(composite IQ mean = 109.9 + 10.2). Duration of disease and

HAIC level were not significantly correlated with scores on

the SB:FE. However, children who developed diabetes

relatively later performed significantly poorer on short-

term memory (Bead Memory, Sentence Memory). History of

asymptomatic hypoglycemia, but not history of severe

hypoglycemia, was significantly correlated with lower scores

on the abstract/visual reasoning scale, especially the

copying subtest.

Clark (1989) was interested in determining if the

neuropsychological deficits found in children ages 6-16 with

EOD would be found if acute hypoglycemia, which may impair

functioning, was controlled. Her sample of 30 EODs

(diagnosis at < 5 years old), 30 LODs (diagnosis at > 6

years old), and 20 sibling controls were given tests

examining eight cognitive areas. Intelligence was estimated

with the WISC-R subtests of Information, Comprehension,

Similarities, Digit Span, Picture Completion, Block Design,

and Coding. The WRAT was used as a measure of school

achievement. Verbal memory was assessed with the RAVLT and

the Logical Memory subtest of the Wechsler Memory Scale

(WMS); spatial memory was assessed with the Visual

Reproduction subtest of the WMS and the Rey-Osterreith

Complex Figure (ROCF) Visuomotor skills were evaluated by

the WISC-R Coding subtest, the Symbol Digit Modalities Test

(SDMT), the Cancellation of Rapidly Recurring Target Figures

Test (CRRTFT), and the Trail Making Test (TMT). Attentional

abilities were assessed with the SDMT, the CRRTFT, the TMT,

the WISC-R Digit Span subtest, and the Seashore Rhythm Test.

Motor ability was evaluated using the Grooved Pegboard Test.

Frontal lobe/executive functioning was measured with the

Controlled Oral Word Association Test, the TMT, Part B,

Luria Figures and Motor Programmes, and the Wisconsin Card

Sorting Test (WCST).

Clark found that the EODs performed significantly lower

than the LODs on visuomotor skills, and significantly lower

than controls and LODs on attentional functioning. The

interpretation of this finding is obscured by the overlap of

the measures used for assessing these two cognitive areas.

Impartial raters determined that significantly more EODs

than LODs or controls performed in the impaired range on

tests of attention. Disease duration, history of severe

hypoglycemia, and onset age were predictors of

neuropsychological performance.

Cognitive ability in a group of 85 adolescents ranging

in age from 14 to 16.5 years, who had been diagnosed with

IDDM for at least 12 months, was also examined by Northam,

Bowden, Anderson, and Court (1992). Intelligence was

assessed with the WISC-R subtests of Similarities,

Vocabulary, Digit Span, Block Design, Object Assembly,

Coding, and Mazes. The RAVLT provided information on verbal

memory and learning and the ROCF was used to evaluate visual

memory. The TMT was used to assess visuomotor tracking, and

the WCST was used to examine problem-solving. Additionally,

the Child Behaviour Checklist and the Youth Self-Report were

used to assess social competence and behavioral adjustment.

Northam et al. (1992) found a trend (p < .06) for onset

age to predict lower verbal scores when, unlike in other

studies, it was treated as a continuous variable. Their

subjects with diabetes performed below age norms on all the

neuropsychological tests, and were characterized as having

more behavior problems (according to maternal report) than

the norm. As no control group was examined and as it is

unclear as to whether the norms utilized were based on

Australian children similar to the subjects, it is difficult

to interpret these results. Data was collected on metabolic

control (HA1C), episodes of ketoacidosis and severe

hypoglycemia which was subjectively defined as "an incident

of altered state of consciousness requiring external

assistance" (p. 888). Blood glucose levels prior to and

after testing were recorded, but not controlled for. No

relationship between test scores and current or past HA1C,

current blood glucose level, and severe metabolic crises was


Holmes, Dunlap, Chen, and Cornwell (1992) also used the

WISC-R in their evaluation of learning difficulties in 95

children with diabetes and their 97 grade- and gender-

matched controls, all of whom were 8 to 16 years old. They

found that, while all mean scores were within a normal

range, boys with diabetes performed more poorly on the

Freedom From Distractibility subtests (Arithmetic, Coding,

and Digit Span) than the girls with diabetes and the

controls. Additionally, the boys with diabetes had lower

Perceptual Organization scores compared to the boys in the

control group. Based on school records, the children with

diabetes required more extra instruction than the controls

(24% vs. 13%), with the boys with diabetes requiring more

than the girls with diabetes (40% vs. 16%). The

investigators interpreted these findings as indicating that

children with diabetes, especially boys, had more learning

problems than their healthy peers. These differences were

not significantly associated with age at disease onset,

duration, or HA1C level. Acute blood glucose level was not


Vocabulary subtest scores on the WISC-R and the WAIS-R

were assessed over an eight year period in 57 children with

diabetes who ranged in age from 8 to 13 at the beginning of

the study (Kovacs, Ryan, & Obrosky, 1994). The researchers

detected a significant decline in verbal performance and

school grades over time. They attributed these changes to

poor short-term memory and learning ability. Memory and

learning were assessed by the Four-Word Short-Term Memory

Test, and the Logical Memory and Verbal Associative Learning

Test on the WMS Test. From these results, they conclude

that a decline in verbal intellectual performance mediated

by verbal learning and memory impairment led to a decline in

school grades. Disease duration, but not metabolic control,

influenced these results. The relationship of school

absences to verbal and school performance were not examined.

The most recent study on cognitive performance in

children with diabetes examined 124 children ages 3-14 who

were evaluated three months after diagnosis (Northam,

Anderson, Werther, Adler, & Andrewes, 1995). They, along

with 129 healthy children matched on age, gender, and SES,

were assessed on general intelligence, attention, memory,

new learning, executive functions, and educational

achievement as part on the base-line testing for a

longitudinal study. Blood glucose levels at testing and

HAjC values were also recorded.

General intelligence was assessed by the WPPSI-R and

the WISC-R, the Sentences subtest on the WPPSI-R assessed

attention and short-term memory in young children, Digit

Span and Corsi Blocks measured attention and immediate

memory, Coding examined sustained memory and psychomotor

speed, the RAVLT and Lhermitte Board evaluated new learning,

the COWAT and the ROCFT measured executive functioning, and

the WRAT-R assessed academic achievement. Additionally, the

CBCL and the Teacher Report Form (TRF) were given to

evaluated psychological adjustment.

The results showed that there were notable differences

between the children with diabetes and their non-diabetic

peers only in the area of attention. On the Freedom From

Distractibility (FFD) subtests on the WISC-R there was a

tendency for the controls to perform better than the

children with diabetes with the difference being significant

on Coding. (Incidentally, girls overall had significantly

higher scores on the FFD subtests.) Additionally, on the

TRF, teachers reported that the children with diabetes had

significantly more problems with attention than their peers.

Acute Hypo- and Hyperglycemia and Performance

The studies of cognitive functioning in children with

IDDM over the past decade do not have the methodological

problems characteristic of the early studies as they

typically have employed control subjects, evaluated diabetic

variables, and utilized appropriate modern statistics.

However, despite documentation from adult studies that

neuropsychological performance is impaired during acute

hypo- and hyperglycemia (c.f., Herold, Polonsky, Cohen,

Levy, & Douglas, 1985; Holmes, Hayford, Gonzalez, & Weydert,

1983; Pramming, Thorsteinsson, Theilgaard, Pinner, & Binder,


1986), only three studies (Clark, 1989; Golden et al., 1989;

Northam et al., 1995) have reported controlling for

hypoglycemia and none have controlled for hyperglycemia.

Clark (1989) and Northam et al. (1992, 1995) measured pre-

and post-testing blood glucose levels. However, the blood

glucose levels in Clark's study were high with a pre-test

mean of 300 mg/dl (SD = 118 mg/dl) and a post-test mean of

255 mg/dl (SD = 93 mg/dl). And, although Northam et al.

(1992) found no significant correlation between test

performance and blood glucose levels, they admitted that two

children were hypoglycemic to the degree that cognitive

impairment was possible and that most of their subjects had

blood glucose levels "higher than clinically desirable" (p.


Only a few studies by Puczynski, Puczynski, and

colleagues, which were reported in 1990, have specifically

examined the effect of various blood glucose levels on

cognitive functioning in children (Puczynski, Puczynski,

Reich, Kaspar, & Emanuele, 1990; Reich, Kaspar, Puczynski,

Puczynski, Cleland, Dell'Angela, & Emanuele, 1990; Ryan,

Atchison, Puczynski, Puczynski, Arslanian, & Becker, 1990).

Twenty-four children with IDDM 7 to 15 years old attending

summer camp were the subjects in a study examining the

relationship of hypoglycemia and performance on

neuropsychological tests of motor speed, attention and

concentration, and memory (Puczynski et al., 1990; Reich et

al., 1990). The measures included the Finger Tapping Test,

the TMT, the WISC-R Digit Span subtest, and the Klove-

Matthews Coordination Test. The children were also asked to

write their names with their dominant and non-dominant


The experimental group consisted of 14 children with

diabetes who were tested 10-40 minutes after an episode of

documented hypoglycemia (bG between 30 and 60 mg/dl

accompanied by physical symptoms) as soon as physical

symptoms had subsided and again 4-6 hours later. Their test

scores were compared to those of another group of 10

children with diabetes who were initially tested during

documented euglycemia (bG between 80 and 240 mg/dl with no

physical symptoms). The control group, which was described

by Reich et al. (1990), was comprised of 14 children without

diabetes who were similar in age and gender to the

experimental group. They were given the same

neuropsychological testing as the children with diabetes;

this testing was repeated 4-6 hours later. Overall, the

results showed that performance on the tests of motor speed,

attention, and memory were poorer both immediately following

a hypoglycemic episode and 4-6 hours later, even though

physical symptoms were not present and blood glucose was

assumed to be at normal levels. These findings were

especially potent for tasks involving the use of the

dominant hand. Twenty-four hours later, the diabetic

children were all performing at "similarly high levels"

(Reich et al., 1990, p. 625).

In order to determine the effect of current blood

glucose level on neuropsychological tests involving mental

efficiency, attention and concentration, and rapid

responding, Ryan et al. (1990) utilized an insulin glucose

clamp. This apparatus consisted of two intravenous

catheters, one for blood sampling and one for infusion of

glucose and insulin. This allowed precise control of blood

glucose levels. Subjects in this study included eleven

children with IDDM between the ages 11 and 18 who were

hospitalized for education and metabolic stabilization. The

measures included a test of simple and choice visual

reaction time (20 trials each) requiring a response to a

blue light, the TMT, and the Stroop Color-Word Test.

Prior to the study, all subjects were given practice

with each measure to avoid possible improvements on the

tasks based on repeated administrations during the actual

study. The night before the study, all subjects were

maintained at euglycemic levels (90-120 mg/dl) The study

began by obtaining baseline testing after a period of stable

euglycemia (100 mg/dl). Plasma glucose concentration was

then maintained at 65 mg/dl and 55 mg/dl before tests were

readministered. Testing was repeated again after blood

glucose level returned to 90 mg/dl and after blood glucose

level was maintained at 100 mg/dl for 15 minutes.

The results documented that reaction time was slowed

and performance on the TMT, Part B and the first two

sections of the Stroop was impaired, relative to baseline,

during the hypoglycemic periods. Performance, in many

cases, did not return to baseline level after a 15 minute

period of euglycemia. Despite the general tendency for

performance to decline during hypoglycemia, there was a

significant amount of inter-subject variability. When the

percent change in scores between baseline and the end of the

hypoglycemic period were correlated with diabetic variables,

such as age at onset, duration of illness, and HA1C values,

no significant associations were found. The authors

concluded that "the complete recovery of mental efficiency

may lag behind restoration to euglycemia" (p. 36) and not

occur simultaneously with the resolution of physical

symptoms associated with hypoglycemic and the

reestablishment of euglycemia. This conclusion is

consistent with the results found by Puczynski et al. (1990)

and Reich et al. (1990).

Although these three studies contained the significant

methodological problems of being poorly controlled and of

having a small sample size with large intersubject

variability, the consistency of the results is compelling.

Moreover, the findings are consistent with the results of

studies involving adults. As early as 1922, Miles and Root,

and later Dashiell in 1930, reported slower reaction times,

decreased strength, and poorer attention and memory during


A series of studies done by Holmes and her colleagues

(cited in Holmes, 1990) examining neuropsychological

performance at various blood glucose levels in adults were

among the best controlled studies as they utilized a

glucose-controlled insulin infusion system similar to the

insulin glucose clamp described above in the study by Ryan

et al. (1990). Their findings demonstrated decrements in

attention, problem solving, and reaction times during

periods of hypoglycemia (bG = 55-62 mg/dl) and hyperglycemia

(300 mg/dl) as compared with performance during periods of

euglycemia (bG = 100-110 mg/dl) (Holmes et al., 1983;

Holmes, 1990). Moreover, Pramming et al. (1986) utilized

measures of attention, memory, and motor functioning and

found impairments in performance during hypoglycemia, even

when the subject was asymptomatic and unaware of the lower

glucose level. Herold et al. (1985) found that reaction

time did not return to baseline after induced hypoglycemia

(50 mg/dl) in normals and subjects with diabetes until 10-40

minutes after restoration to euglycemia (85-110 mg/dl) The

previously described child studies were consistent with this

evidence of a delay in the reestablishment of baseline CNS

functioning after peripheral glucose levels returned to


Summary of Neuropsychological Findings

The results of research examining neuropsychological

functioning in children with IDDM have been inconsistent.

Findings ranged from demonstrating that children with early

onset diabetes were impaired on all tests (Ryan et al.,

1981) to concluding that children with diabetes did not

perform significantly different from normals (Northam et

al., 1992). Significant risk factors have included early

diagnosis age, female gender, long disease duration, a

history of mild or severe hypoglycemia, and a history of

ketoacidosis. Closer examination of the research studies

provides insight into these discrepancies, suggesting that

method variance may account for the mixed results. The

investigators examined children of assorted age ranges, had

different criterion for EOD and LOD, examined dissimilar

cognitive areas, and/or gave different neuropsychological

tests. Furthermore, the control groups employed were

inconsistently matched to the subjects with diabetes,

diverse definitions of mild and severe hypoglycemia were

used, and the accuracy of retrospective metabolic

information could be questioned.

Nevertheless, several common threads have arisen from

this decade of research. The most consistent finding is

that if children with diabetes have difficulties, the

deficits are most likely to appear in the areas of

visuospatial skills, attentional abilities, and motor

functioning. An early onset age was demonstrated to be the

most potent predictor as it was significantly related to

lower performance in all of the studies (9/9) which examined

this variable. Disease duration and history of hypoglycemic

complications were less consistent risk factors for lower

performance. These findings appear to be maintained even

when acute hypoglycemia, which may impair performance, is

controlled. However, it remains to be demonstrated that

these findings persist when recent hypoglycemia (i.e., past

4-6 hours) and acute and recent hyperglycemia is controlled.

Models of Attention

The ability to attend to stimuli is one of the most

important capabilities of the human brain as all other

cognitive functions rely on this (Cooley & Morris, 1990).

Because of this, attention is one of the primary cognitive

functions assessed during recovery from coma or other

conditions affecting the CNS such as metabolic toxicity

(Sohlberg & Mateer, 1989). Without this capacity, normal

conversation, memory, and performance of other cognitive

functions, such as those measured on intellectual testing,

would not be possible. However, before attention can be

measured, it must first be defined. William James (1890)

asserted that "Everyone knows what attention is. It is the

taking possession by the mind, in clear and vivid form, of

one out of what seem several simultaneously possible objects

or trains of thought" (p. 403-404), thus suggesting that

attention could occur as a sensory or as an intellectual

phenomenon (LaBerge, 1990). Since James' time, the

definition of attention had been reworked numerous times

yielding both clinically and experimentally derived

descriptions. The fact that a universal definition of

attention remains elusive and that attention has not been as

well studied as other cognitive abilities (Mirsky et al.,

1991) contributes to the relative lack of cohesiveness of

the results in this area.

The current conceptualization of attention is based on

the belief that the capacity to process information is

finite (Sohlberg & Mateer, 1989), a viewpoint which was not

elucidated until as recently as 1950 (LaBerge, 1990).

Broadbent (1958) is cited as the first to report the concept

of discrimination in attention (LaBerge, 1990; Sohlberg &

Mateer, 1989). His "filter theory" developed from research

demonstrating that stimuli was selectively rearranged, that

not all stimuli was given equal attention depending upon

task requirements, and that there was a bias towards novel

stimuli (Broadbent, 1958).

Since Broadbent, a number of models of attention have

been proposed based on clinical observations and

experimental results. However, few have specifically

focused on attention in children (Cooley & Morris, 1990),

and rarely were normal cognitive development,

neuropsychological functioning, and neuroanatomical

substrates examined in those studies (Cooley & Morris, 1990;

Goodyear & Hynd, 1992).

Pribram and McGuinness' (1975) model was based largely

on research involving animals' responses to tones and, while

now outdated (Mirsky et al., 1991), provided useful insight

into attention. In an effort to consolidate

neuropsychological and psychophysiological information, they

described attention as consisting of three distinct, but

interacting, neural systems, namely arousal, activation, and

effort. They defined arousal in terms of an orienting

response, or more specifically as "phasic physiological

responses to input" (p. 116). They postulated that the

neuronal system involved in this process extended "from the

spinal cord through the brainstem reticular formation,

including hypothalamic sites" (p. 123) and centered on the

amygdala. Activation was the "tonic physiological readiness

to respond" (p. 116) and was regulated by a circuit which

centered on the basal ganglia of the forebrain. Effort

coordinated these processes through a hippocampal circuit.

Currently, studies of attention have typically focused

on impairments in regulation of attentional processes and

resulting behavioral problems. This has resulted in a

clinical description of these disorders defined as the

syndrome of neglect in adults and as the DSM-III-R diagnosis

of Attention-deficit Hyperactivity Disorder (ADHD) in

children. Heilman, Watson, & Valenstein (1985) described

the predominant behavioral symptom of neglect as hemi-

inattention or lack of awareness and response to stimuli

presented contralateral to a hemispheric lesion. Neglect

typically occurs with lesions to the inferior parietal lobe

(especially with right-sided lesions), but may also occur

after damage to the dorsolateral frontal lobe, the cingulate

gyrus, the corpus striatum, and the thalamus (Heilman et

al., 1985; Watson, Valenstein, & Heilman, 1981). Heilman et

al. (1991) postulated that the right hemisphere may be

dominant for mediating attentional functioning.

Children rarely manifest the neglect syndrome (Cooley &

Morris, 1990). However, Heilman et al. (1991) used

information from studies of neglect and other impairments to

guide their investigation of children diagnosed with ADHD.

ADHD is comprised of the symptoms of inattention,

impulsivity, and hyperactivity. Based on these symptoms and

their similarity to the symptoms of patients with known

lesions and cerebral dysfunctions, Heilman et al. (1990)

hypothesized that certain neuroanatomical areas were

dysfunctional in children with ADHD.

Heilman et al. (1990) found that, in addition to

impaired attention, children with ADHD had a propensity to

make left-sided errors on cancellation tasks. These

symptoms are similar to those of patients with known right

hemisphere lesions who have neglect. Children with ADHD

also tend to have motor impersistence and difficulty

inhibiting responses (based on a go-no go paradigm). These

symptoms were described as being similar to those of

patients with frontal lobe impairments. Furthermore,

children with ADHD were found to have decreased blood flow

in the right striatum which governs aspects of inhibition.

Moreover, their motor restlessness was seen as resembling

the akathisia of patients with Parkinson's disease and

patients treated with neuroleptic medications. This

restlessness has been described as inversely correlated with

prefrontal cortical dopamine levels. Based on this

evidence, Heilman et al. (1990) proposed that children with

ADHD had right hemisphere dysfunction with impairments in

the frontal-striatal system.

Additional support for this model comes from

researchers who emphasized the role of the frontal lobes

(Benson, 1991) and dopamine and norepinephrine metabolism

(Roeltgen & Schneider, 1991) as being important in

attention. Decreased dopamine level, which has been

associated with attentional problems (Heilman et al., 1991;

Roeltgen & Schneider, 1991), is one of the neurochemical

changes associated with insulin-dependent diabetes mellitus

(Mooradian, 1988).

Many researchers no longer describe attention as a

unitary concept, but recognize that it has multidimensional

features (Cooley & Morris, 1990; Mirsky, Anthony, Duncan,

Ahearn, Kellam, 1991; Posner & Petersen, 1990; Sohlberg &

Mateer, 1989). Sohlberg and Mateer (1989) outlined a five

dimensional clinical model of attention which they use to

guide their research on head injury. Their model included

(1) focused attention, which is the ability to respond to

stimuli; (2) sustained attention, which is the ability to

maintain a behavioral response over time; (3) selective

attention, which is the ability to maintain a response

despite distractions; (4) alternating attention, which is

the ability to shift the focus of attention between tasks;

and (5) divided attention, which is the ability to attend to

multiple simultaneous demands (p. 120-121).

Utilizing a multidimensional construct of attention,

Cooley and Morris (1990) presented a model of selective

attention in children. They proposed that attention was

comprised of two main components which function

simultaneously. One component identified relevant stimuli;

the other inhibited irrelevant stimuli. Sustained attention

was then defined as performing this function over time.

Divided attention was defined as performing this function

for more than one target at a time.

Posner and Petersen (1990) presented a model of

attention which, like that of Pribram and McGuinness (1975),

included neuroanatomical correlates. They divided attention

into three areas of functioning, namely orienting,

detecting, and alerting. Orienting, which referred to an

automatic process of becoming aware of sensory events and

shifting focus to that stimulus, is similar to Pribram and

McGuinness' (1975) arousal component. Detecting involved

conscious selection of target stimuli and is similar to

Cooley and Morris' (1991) selective component. Alerting was

defined as maintaining a vigilant or alert state and can be

equated with sustained attention.

Posner and Petersen (1990) described the neuroanatomy

of attention as a "unified system" which was "anatomically

separate from the data processing systems" (p. 26), and

consisting of several anatomical areas which regulated the

distinctive aspects of attention. They described the

orienting response as being regulated primarily by a

posterior attention system based in the parietal cortex in

normal humans. The posterior parietal lobe, the lateral

pulvinar nucleus of the posterolateral thalamus, and the

superior colliculus were also included based on information

from studies with primates and human lesions investigations.

Their detecting element of attention was postulated as being

mediated by an anterior attention system. This system

involved midline frontal areas including the anterior

cingulate gyrus which projects into the dorsolateral

prefrontal cortex. According to Posner and Petersen (1990),

the integrity of the right cerebral hemisphere was central

in the functioning of the vigilance component of attention

and cited evidence from studies of neglect to support this

conclusion. They added that vigilance was supported by

those structures involved in the norepinephrine innervation

system originating in the locus coeruleus and including

pathways through the prefrontal areas and the posterior

attention system described above.

Mirsky et al. (1991) presented the most comprehensive

neuropsychological model for attention reported in the

literature. They outlined clinical components of attention,

which were supported and expanded upon by the results of two

principal components analyses of scores from neuro-

psychological tests purported to measure various aspects of

attention. Moreover, they offered neuroanatomical

correlates for each of the components.

Their sample of 203 predominantly neurologically normal

adults (14 had epilepsy, 9 had unspecified head injury) were

given the Stroop, the Talland Letter Cancellation Test, the

Trail Making Test, the Digit Symbol Substitution Test, the

Arithmetic and Digit Span subtests from the WAIS-R, the

Continuous Performance Test (CPT), and the Wisconsin Card

Sorting Test (WCST). Their second sample consisted of 435

children who were given Digit Cancellation, the WCST, the

CPT, and the Coding, Arithmetic, and Digit Span subtests

from the WISC-R.

A principal components analysis and a factor analysis,

which were performed separately on both the adult scores and

the child scores, yielded the same four factors which the

authors referred to as focus-execute, sustain, shift, and

encode. Focus-execute involved the "ability to select

target information from an array" (p. 11, Mirsky et al.,

1991) and to render a behavioral response. This definition

seems to incorporate what has been referred to separately as

attention and intention (Heilman et al., 1991; Watson et

al., 1981). Mirsky et al. (1991) defined sustain, or

vigilance, as the ability to maintain focus over time, and

shift as the ability to change this focus and reengage it

elsewhere. Encode involved mental manipulation of the

stimuli and a working memory component.

Mirsky et al. (1991) cited evidence based on human

clinical cases and lesions studies in animals that the four

factors obtained in their study were governed by specific

neuroanatomical areas. Poor performance on measures of the

focus-execute aspect of attention, Mirsky et al. (1991)

indicated, could result from neglect due to damage in any

neuroanatomical area governing attention. However, they

listed the inferior parietal cortex, the superior temporal

cortex (sulcus), and the corpus striatum as specifically

regulating this aspect of attention. The tectum and

mesopontine regions of the reticular formation as well as

the midline thalamic region and the reticular nuclei of the

thalamus were attributed to the sustain component of


Mirsky et al. (1991) included the dorsolateral

prefrontal cortex, the medial frontal cortex, and the

anterior cingulate gyrus as the neuroanatomical structures

which mediated their shift component of attention. They

implicated the hippocampus and the amygdala in the encode

aspect of attention based on electrophysiological and animal

data. This is largely because the encode factor relied on

short term memory functioning for which the hippocampus and

amygdala play an important role (Kolb & Wishaw, 1990).

As attention is a complex, multi-faceted function, it

is not surprising that there are numerous definitions of

this phenomenon and that many models--behavioral, cognitive,

and neurological--have arisen in attempts to define it.

Furthermore, the large number of neuroanatomical structures

which have been implicated as mediating attention should not

have been unanticipated, for as Colby (1991) asserted, much

of the brain is involved in this complex operation.

The models offered by Pribram and McGuinness (1975),

Heilman and colleagues (Heilman et al., 1985, 1991; Watson

et al., 1981), Cooley and Morris (1990), Posner and Petersen

(1990), and Mirsky et al. (1991) define the functional

aspects of attention in different ways. However, there are

considerable similarities between the different descriptions

and significant overlap in the neurological areas described

as governing attentional processes. The common components

of attention included stimulus detection and selection,


being able to sustain this process over time, the ability to

consciously change the focus of attention, and a readiness

to respond. Certain neuroanatomical regions were repeatedly

implicated as mediating processes of attention. These

included the inferior parietal lobes, the dorsolateral

frontal lobes, the cingulate gyrus, the thalamus, the

hippocampus and amygdala, the corpus striatum, and the

reticular formation. Also, the right hemisphere was

considered as dominant for attention. Additionally, the

neurotransmitters dopamine and norepinephrine were

implicated as mediating attentional functions.

As described above, certain areas of the brain have

been reported as being sensitive to hypoglycemic episodes.

These areas include the frontal lobes, the hippocampus, the

thalamus, and the corpus striatum, areas which were

described as governing aspects of attention. Decreased

blood flow to the frontal regions and basal ganglia have

also been found in children with attentional problems

(Benson, 1990; Heilman et al., 1990). Having examined

various models of attention, it would next be informative to

consider a background in normal brain development to further

elucidate reasons for possible impairments in children with


Brain Development

Brain development is a complex and sequential process

in which maturational changes occur at biologically

determined ages within the organism both prenatally and

postnatally (Kolb & Fantie, 1989; Spreen, Tupper, Risser,

Tuokko, & Edgell, 1984). Growth of the central nervous

system involves cell proliferation, cell migration, the

development and growth of axons, dendritic formation, the

development of synaptic connections, and myelination (Kolb &

Wishaw, 1990; Voeller, 1993). The following is a brief

overview of some of the processes involved in brain


Brain Volume

An increase in brain weight is one of the more obvious

indicators of brain growth, but may also be considered as

one of the indices of other neuro-developmental changes

(Dobbing, 1974). Neurogenesis progresses from the brainstem

to the diencephalic regions (i.e., thalamus, hypothalamus)

to the cortex. Cortical neurons proliferate in the

ventricular proliferation zone of the germinal matrix

(Voeller, 1993). Once "born," neurons migrate from this

region on radial glides (glial fibers) to their final

location in the brain with the deeper layers forming earlier

than the outer layers (Rakic, 1974).

By approximately eight weeks after conception, the head

comprises at least half of the fetus; by birth, the brain

weighs between 300 and 350 grams (Lemire, Loeser, Leech, &

Alvord, 1975). The weight of the brain rapidly increases to

90% of its adult weight of 1250 to 1500 grams by age six


(Majovski, 1989). This expansion in size is largely due to

a rapid "overgrowth" of axons, dendritic arborization and

spine growth, and synaptic connections. Between 12 and 24

months synaptic density is at the peak of 50% more than the

average adult level (Huttenlocher, 1979, 1984). The number

of neurons gradually decreases through a process calling

pruning, or shedding, in which synaptic connections are

modified in terms of locations and type causing neuronal

death until an optimal level is reached (Voeller, 1993).

(Interestingly, abnormally high synaptic density has been

associated with cognitive deficits [Kolb & Fantie, 1989;

Voeller, 1993]). With the increase in brain size comes an

increase in complexity and connectivity, and subsequently,

an increase in responsivity (Goldman & Rakic, 1979).


Part of the increase in responsivity mentioned above

occurs because of myelogenesis, or myelination. In this

process, myelin, a fatty sheath comprised of proteins and

lipids, develops around nerve fibers in order to insulate

them (Kolb & Fantie, 1989). The myelin sheath is thought to

enhance neuronal functioning and efficiency by elevating

conduction velocity, decreasing the action potential

threshold, and expanding neuronal capability to relay

repetitive impulses (Majovski, 1989; Spreen et al., 1984).

Although neurons are functional prior to myelination (Kolb &

Fantie, 1989), myelination is considered to be a significant

determinant of the maturity of a brain region because

neurons attain their full "adult" level of functioning when

this process is completed (Majovski, 1989; Yakovlev &

Lecours, 1967).

First Flechsig (1901), and then Yakovlev and Lecours

(1967) performed extensive studies of the progression of

myelination. Myelogenesis begins prior to birth at the end

of the fourth fetal month with the motor root fibers in the

spinal cord (Yakovlev & Lecours, 1967). This process

progresses to the medulla, pons, and midbrain, then to the

diencephalon and the telencephalon (i.e., basal ganglia,

limbic system) so that myelination has begun in all areas of

the brain by the middle of the first year of life (Spreen et

al., 1984). Areas governing primary motor and sensory

functioning are among the first to have completed

myelination. While myelination is completed in most areas

by age four, some areas, such as the cingulum, striatum, and

certain thalamic projections, are not fully myelinated until

the mid-teens, and some areas, such as the reticular

formation and intracortical association systems continue

with the myelination process throughout adulthood (Kolb &

Fantie, 1989; Yakovlev & Lecours, 1967).

Growth Spurts

The development of the brain does not occur at a

uniform rate, but goes through periods of accelerated growth

which are commonly referred to as "growth spurts" (Kolb &

Fantie, 1989; Spreen et al., 1984). The first main growth

spurt, which involves an increase in glial cells, occurs

around the 30th week of gestation (Dobbing, 1974) and is

followed postnatally by a spurt beginning between 3 and 10

months (Kolb & Fantie, 1989). Epstein's (1978) research

described growth spurts as occurring between the ages of 2

to 4, 6 to 8, 10 to 12, and 14 to 16 years.

Critical Periods

The term "critical period" refers to the fact that at

certain stages of growth, the brain (and body) is more

susceptible to environmental influences. These extrinsic

influences include normal sensory stimulation which promotes

normal neuronal connectivity and function (Voeller, 1993).

Other factors, such as infections, irradiation, chemical

agents, and metabolic problems may interfere with and thwart

normal development (Kolb & Fantie, 1989; Spreen et al.,

1984). The regions of the brain which are the most

sensitive are those which are most rapidly developing. As a

result, the critical periods of the brain are presumed to

correspond to the previously described growth spurts

(Dobbing, 1974). Evidence for qualitative and quantitative

changes occurring during pre- and postnatal critical periods

in humans and animals has been cited (Dobbing, 1974; Spreen

et al., 1984; Voeller, 1993). This is a complex area of

study, however, as there is often reorganization of brain

functioning following damage, especially in young children

(Kolb & Fantie, 1989; Voeller, 1993). This reorganization

which allows for compensation of an impaired ability is

referred to as plasticity (Kolb & Wishaw, 1990). When

compensating for an impaired function, however, another

function may be compromised. This is more likely to occur

in children and may not be recognized until later in their

development (Kolb & Fantie, 1989).

This summary of normal brain development provides an

appreciation of the complexity of the processes involved.

Interestingly, several areas described as regulating aspects

of attention, including the frontal lobes (cingulate gyrus),

thalamic areas, the corpus striatum, and the reticular

formation, appear to take the longest to fully develop. The

first three of these areas have also been described as being

sensitive to hypoglycemia. The theory of critical periods

suggests that when an area of the brain is developing, it is

the most sensitive to injury. However, the concept of

plasticity proposes that younger children are more likely to

exhibit better recovery. Although, with recovery, there are

often functional compromises. Taken together, the evidence

suggests that children who develop diabetes at a earlier age

and who have more episodes of hypoglycemia may exhibit

deficits in attention.


The purpose of this research study was to examine

attentional abilities in children with IDDM and to

investigate possible risk factor(s) for poorer attention

should this finding emerge. Attention underlies all other

cognitive abilities and thus is important to assess before

drawing firm conclusions concerning other abilities.

Several research studies have suggested that attention in

children with IDDM, especially those with early disease

onset, is impaired. However, this conclusion may be

premature and overly simplistic as attention in children

with diabetes has not been adequately measured, especially

with respect to the current neuropsychological theory of

attention as a multidimensional construct.

That children with early onset diabetes have decreased

attention and concentration has been a conclusion of a

number of studies. Two studies made this inference based on

parent report of distractibility in EODs (Rovet et al.,

1983) and certain ADHD symptoms on the Connors Scale (Hagen

et al., 1990). Another study defined attention in terms of

performance on Digit Span which was significantly lower for

EODs than for controls (Ryan, Vega, & Drash, 1985).


Clark (1989) did the most comprehensive examination of

attentional functioning in children with IDDM and found that

children with EOD performed more poorly than LODs or

controls on attentional tests. This was the only study

examining attentional factors which controlled for

hypoglycemia. The tests of attention included the Symbol

Digit Modalities Test, the Trail Making Test, Digit Span,

the Cancellation of Rapidly Recurring Target Figures Test,

the Seashore Rhythm Test, and Knox Cubes. While the latter

two tests were identified as measures of sustained

attention, the Knox Cubes was not included in the analyses,

nor was sustained attention analyzed as a separate factor.

Of the tests analyzed, only Digit Span and the Seashore

Rhythm Test were not dependent on motor speed.

In the present research, attention was treated as a

multidimensional construct, not as a unitary construct as in

the previous studies of children with diabetes. Three

components of attention were measured, namely selective

attention, sustained attention, and alternating attention.

For the purposes of this study, selective attention was

defined as the ability to select target stimuli from an

array of stimuli and is synonymous with Cooley and Morris'

(1990) selective and Posner and Petersen's (1990) detecting

component. Sustained attention referred to the ability to

perform the selective function over a period of time.

Alternating attention referred to the ability to alternate

between task requirements and/or target stimuli.

In addition to assessing the children's performance on

tests of attention, parent measures of child behavioral

symptoms of concentration and distractibility were also

obtained. Furthermore, standardized achievement scores were

requested from schools to examine any relationship between

academic achievement and scores on the tests of attention.

Information on school absences were also obtained as this

may impact on achievement scores.

Moreover, so that the results were not attributable to

the symptoms of poor concentration associated with low blood

glucose levels, children experiencing acute hypoglycemia

(blood glucose < 60 mg/dl) were not tested until later.

They were invited to participate only after their blood

glucose level had been above 80 mg/dl for at least four

hours. This requirement was expected to circumvent any

lingering effects of hypoglycemia which some children

experience (Reich et al., 1990).


It was hypothesized that children with early-onset

diabetes would perform more poorly than children with late-

onset diabetes or controls on the tests of selective,

alternating, and sustained attention. Evidence from

previous studies suggests that EODs have difficulty with

focusing their attention (Clark, 1989; Hagen et al., 1990;

Rovet et al., 1983; Ryan, Vega, & Drash, 1985), thus it was

surmised that they would have difficulty with the three

aspects of attention which were examined in this study. It

was anticipated that this finding would not be related to

gender or current blood glucose level.

Furthermore, it was hypothesized that, in addition to

early onset age, duration of diabetes, and a history of

severe hypoglycemia would predict poorer performance on the

measures of attention. Research describing permanent

neurological impairment as a result of recurrent

hypoglycemic episodes supports the evidence that attentional

abilities may be impaired in children with diabetes (e.g.,

Cirillo et al., 1984; Haumont et al., 1979). A history of

hyperglycemia was not anticipated to significantly relate to

the attention scores, thus replicating the findings of most

studies (e.g., Clark, 1989).

Parent ratings of their child's attention were

predicted to be significantly related to the child's

performance on the measures of attention. Also, children

with EOD were predicted to have more problems with attention

reported by their parents. Past studies have described

parent report of concentration problems in EODs (Hagen et

al., 1990; Rovet et al., 1983). Additionally, a significant

relationship between standardized achievement scores and

child attentional performance was foreseen for all the

children. There is evidence suggesting that children with

diabetes may have academic difficulties (e.g., Holmes et

al., 1992). The suspected attentional deficits in EODs may

contribute to a higher incidence of academic difficulties.



Subjects included 50 children who had insulin-dependent

diabetes mellitus for at least 12 months. They were chosen

from consecutive clinic appointments at the Shands Pediatric

Diabetes Clinic and comprised two groups: 20 who were

diagnosed as having diabetes prior to age six years, 0

months and 30 who were diagnosed after this age. The first

group was referred to as the early-onset diabetes (EOD)

group; the second group was referred to as the late-onset

diabetes (LOD) group. Previous studies of children with

diabetes have defined early-onset as diagnosis prior to

between the ages of 4-7. So that this study could be

compared to the previous ones, the cut-off for EOD was also

within that age range.

Fifty children without diabetes matched on age and sex

to the children with diabetes served as controls. (See

Table 2 for demographic information on the entire sample).

Each control subject was within eleven months (typically

within five months) of age of her/his match. The control

group consisted of siblings of the children with diabetes,

children from the community, and healthy children seen for

routine appointments in the adolescent clinic.


Table 2

Demographics of Sample

Total All All
Sample TDDM Controls
N=100 N=50 N=50

12yr 5mo
2yr 5mo


12yr 5mo
2yr 5mo

12yr 5mo
2yr 5mo

Standard Score*
Mean 105.67
S.D. 15.28

Grade Level

Repeated Grade



School Absences*
(Last Term)

< 1X month
1X month
< 1X week
1X week
2-3X week

IDDM related



*Significant difference between IDDM and Control groups.













Table 2 (Continued)

Total All All
Sample IDDM Controls
N=100 N=50 N=50

Family SES
I (Low) 0% 0% 0%
2 4% 6% 2%
3 23% 32% 14%
4 45% 42% 48%
5 (High) 19% 12% 26%
Unknown 9% 8% 10%

Marital Status
Married 66% 66% 66%
Divorced 26% 22% 30%
Single 4% 6% 2%
Separated 2% 4% 0%
Widowed 2% 2% 2%

< 7th grade 0% 0% 0%
jr HS 0% 0% 0%
partial HS 6% 6% 6%
high school 27% 40% 14%
part. coll. 28% 34% 22%
college grad 18% 12% 24%
grad school 21% 8% 34%

< 7th grade 1% 2% 0%
jr HS 0% 0% 0%
partial HS 1% 2% 0%
high school 32% 44% 20%
part. coll. 20% 22% 18%
college grad 17% 8% 26%
grad school 20% 10% 30%
unknown 9% 12% 6%

*Significant difference between IDDM and Control groups.

Table 2 (Continued)

EOD Controls LOD Controls
N=20 N=20 N=30 N=30

Mean 10yr 9mo 10yr 9mo 13yr 6mo 13yr 6mo
S.D. 2yr imo 2yr imo 2yr 0mo 2yr imo

Girls 9 9 20 20
Boys 11 11 10 10

Standard Score
Mean 104.00 113.55 100.57 106.63
S.D. 13.30 17.18 14.07 14.75

Grade Level
Mean 4.85 4.80 7.67 7.43
S.D. 2.08 1.75 2.02 2.18

Repeated Grade
Never 75% 95% 80% 93%
Once 25% 5% 20% 7%
Twice 0% 0% 0% 0%

School Absences
(Last Term)
none 0% 26% 10% 30%
< 1X month 70% 37% 41% 33%
1X month 10% 16% 24% 3%
< 1X week 15% 21% 10% 23%
1X week 0% 0% 3% 7%
2-3X week 5% 0% 10% 3%

IDDM related
yes 45% 0% 60% 0%
no 55% 100% 40% 100%

*Significant difference between EOD and LOD groups and
between EOD controls and LOD controls.

Table 2 (Continued)

EOD Controls LOD Controls
N=20 N=20 N=30 N=30

Family SES
1 (Low) 0% 0% 0% 0%
2 5% 5% 7% 0%
3 35% 5% 30% 20%
4 40% 40% 43% 53%
5 (High) 15% 25% 10% 27%
Unknown 5% 25% 10% 0%

Marital Status
Married 65% 65% 67% 67%
Divorced 25% 30% 20% 30%
Single 0% 0% 10% 3%
Separated 10% 0% 0% 0%
Widowed 0% 5% 3% 0%

< 7th grade 0% 0% 0% 0%
jr HS 0% 0% 0% 0%
partial HS 10% 5% 3% 7%
high school 50% 15% 33% 13%
part. coll. 25% 15% 40% 27%
college grad 5% 30% 17% 20%
grad school 10% 35% 7% 33%

< 7th grade 5% 0% 0% 0%
jr HS 0% 0% 0% 0%
partial HS 0% 0% 3% 0%
high school 35% 10% 50% 27%
part. coll. 25% 15% 20% 20%
college grad 15% 25% 3% 27%
grad school 15% 35% 7% 27%
unknown 5% 15% 17% 0%

EOD control groups.

*Significant difference between EOD and

To be included, the children had to be 7-16 years old,

have no evidence of psychosis, neurological problems, or

metabolic disorders (other than diabetes). They also had to

have normal, or corrected to normal, vision and motor

functioning. This information was obtained from parents or

determined from the child's ability to complete sample items

on the measures. The children included had not taken any

stimulant or sedating medication recently enough to effect

their attentional functioning. (Ritalin, for example, does

not exert a significant effect after four hours.)

Additionally, for the children with diabetes, current blood

glucose level had to be at least 80 mg/dl immediately prior

to the start of testing. (See Table 3 for diabetes

information). Also, all children had to obtain a standard

score of at least 80 on the PPVT-R.

Assessment Instruments

The measures of attention were chosen based on their

ability to assess aspects of attention without placing a

significant demand on cognitive, verbal, or mathematical

ability or memory functioning; their common usage in

neuropsychological assessment; and their ease of

administration. Research by Mirsky et al. (1991) and the

results of their factor analysis, which statistically

grouped certain measures together based on their ability to

Table 3

Diabetes Characteristics

N=50 N=20 N=30

Onset Age*

Disease Duration*

7yr 5mo
3yr 5mo

4yr limo
2yr 5mo

4yr Imo
lyr 7mo

6yr 9mo
lyr 9mo

9yr 8mo
2yr 3.5mo

3yr 9mo
2yr imo

Blood Glucose
at testing



Loss of


*Significant difference between EOD and LOD groups.



3 .44












3 .95

assess specific aspects of attention, also strongly guided

test selection (see Table 4).

Selective Attention

Cancellation. Cancellation tasks have been used in

numerous studies assessing focused or selective attention in

children (Goodyear & Hynd, 1992). They assess selective

attention, visual tracking, inhibition of impulsive

responding, and motor speed (Lezak, 1983) and may have

letters, digits, shapes, or symbols as stimuli. The

Cancellation of Rapidly Recurring Figures (Rudel, Denkla, &

Broman, 1978 as cited in Gardner, 1979) was included in this

study as it has normative information for children. For

each test, the child was told to draw a line through the

target stimulus as quickly as possible and was timed. A

practice test consisting of 10 rows of 14 digits from 0-9,

with "16"I as the target stimulus was provided. The first

test consists of 14 rows of 10 simple geometric shapes with

a diamond as the target stimulus. The second test consists

of 14 rows of 9 three digit numbers, all starting with 56 or

59, with "592" as the target stimulus. Normative

information for completion time and number of errors was

available for children ages 4-13.

Symbol Digit Modalities Test. The Symbol Digit

Modalities Test (SDMT; Smith, 1968) has been described as

very sensitive to cerebral dysfunction (Lezak, 1983; Smith,

Table 4

Tests of Attention


Scores used

Selective Attention:



Trail Making Test, Part A

Completion time

Number correct

Completion time

Alternating Attention:

Trail Making Test, Part B

Switching Attention

Completion time

Mean reaction time
Number of errors

Sustained Attention:

Mean reaction time
Number of omissions
Number of commissions


1968, 1978) and was originally developed as a screening

instrument for learning disabilities and "cerebral

disorders" (Smith, 1968). It assesses attention, visuomotor

tracking, and motor speed. It is similar to the Coding and

Digit Symbol subtests of the Wechsler Intelligence Scales.

However, the respondent writes the number which corresponds

with the symbol listed in the key, not the reverse as in the

Wechsler subtests. As numbers are more familiar, this task

is easier and allows for an oral version. After practicing

on ten items, the respondent is asked to write/say as many

corresponding numbers as possible within a 90 second time

limit. The test-retest reliability was .80 for the written

version and .76 for the orally administered form (Smith,

1982). In this study, only the written version was used.

Normative information is available for girls and boys ages

8-17, as well as for adults.

Trail Making Test. The Trail Making Test, which

assesses attention, visuomotor tracking, and motor speed,

was originally part of the Army Individual Test Battery

(1944) before becoming part of the Halstead-Reitan battery

(Lezak, 1985). Adult and child versions are available. It

is reported to be highly sensitive to the effects of brain

dysfunction (Spreen & Strauss, 1991; Lezak, 1983). It

consists of two parts: A and B. Part A asks for the

subject to draw lines to consecutively connect randomly

placed numbered circles as quickly as possible. Part B


consists of numbered and lettered circles between which the

subject is asked to alternate. Part B requires the ability

to shift attention between two sequences and will be used as

a measure of alternating attention. Test-retest reliability

has ranged from .46 to .94 for Part A and from .44 to .86

for Part B for adults depending upon subject diagnosis and

age (Spreen & Strauss, 1991). Lezak (1983) stated that the

reliability of Part A is higher because Part B tends to have

more examiner intervention to correct errors. Normative

information is provided for children ages 6-15 and for

adults (Spreen & Strauss, 1991).

Alternating Attention

Switching Attention Test. The Switching Attention Test

was designed to assess alternating, or switching, attention

in adults. This measure was also part of the

Neurobehavioral Evaluation System (NES2) program (Letz &

Baker, 1988). It was adapted from a test designed to assess

attention in patients with neurotoxicity (Eckerman et al.,

1985). This test involves three conditions which are

labelled "Side," "Direction," and "Switching." The "Side"

condition measures choice reaction time by having the

subject respond to a rectangle presented on a computer

screen by pressing the key indicating the side on which the

figure is presented. In the "Direction" condition, the

subject sees an arrow pointing to the left or right and must

respond according to the direction the arrow is pointing.

In the "Switching" condition, the stimuli are again the

arrows which may appear on either side of the screen in

addition to pointing either left or right. Prior to each

trial in this condition, the subject sees either the word

"Side" or "Direction" preceding each trial which indicated

how they were to respond to the stimuli. "Side" means that

the correct answer is the side the arrow is on regardless of

the direction of the arrow. "Direction" indicates that the

correct response is the direction the arrow is pointing

regardless of the side of the screen on which the arrow

appears. Practice trials are at the beginning of each

section. For the purposes of this study, the "Side" and

"Direction" conditions were used only to familiarize the

subjects with the measure so that they understood the

"Switching" subtest. Mean reaction time, total number

correct, and total number of errors were automatically

recorded for each condition. For this study, only the mean

reaction time and the total number of errors for the

"Switching" subtest were used. Psychometrics and normative

data were not available for this measure for children and

adolescent subjects.

Trail Making Test. The Trail Making Test, Part B was

included as a measure of alternating attention. It was

described with Part A under the Selective Attention section.

Sustained Attention

Continuous Performance Test. The Continuous

Performance Test is a commonly used measure of sustained

visual attention (Goodyear & Hynd, 1992) that was first

developed to assess brain damage (Rosvold, Mirsky, Sarason,

Bransome, & Beck, 1956). The CPT used in the present study

was part of the Neurobehavioral Evaluation System (NES2)

program (Letz & Baker, 1988). It consisted of five trials

of the same task. In each trial the subject saw a series of

capital letters on a computer screen, each presented for 60

msec. with 800 msec. between stimuli to allow for response.

Each time the letter "S" appeared, the subject was to press

the left shift key on the key board. For the other letters,

the subject was to refrain from responding. The entire test

took approximately 6 minutes. Reaction times and the number

of omissions and commissions for each trial were

automatically recorded, as was mean reaction time for the

entire test. Psychometrics amd normative data were not

available for children and adolescent subjects.

Receptive Vocabulary and Academic Performance

Peabody Picture Vocabulary Test-Revised. The Peabody

Picture Vocabulary Test-Revised is a measure of receptive

auditory vocabulary. The subject is required to point to

one of four pictures which best corresponds with words read

by an examiner. The words gradually increase in difficulty.

A basal of six correct responses must be met; the test is

discontinued after incorrect responses are given on six out

of eight items. Raw scores may be translated into age

equivalents, standard scores, stanines, and percentiles

(Dunn & Dunn, 1981). Test-retest reliabilities range from

.71-.84 (Spreen & Strauss, 1991). Correlations between the

PPVT-R and the WISC-R range from .16 to .86. The higher

correlations occur with the subtests of the Verbal

Performance Scale (Sattler, 1988). In this study, the PPVT-

R was used as a gross measure of verbal and general

cognitive ability. Normative information for ages 2 1/2 to

40 is available.

Achievement Tests. The California Achievement Test

(CAT) is routinely used in the Alachua County School System,

as well as nation-wide. Percentile scores in language

skills, mathematics, science, and social studies are

obtained for Grades 1-5 and 8. Other grades are tested

depending on requests by individual schools and requirements

of special programs (S. Baker, Alachua County Schools

Testing Department, personal communication, May, 11, 1993).

Other counties and states also use the CAT. However,

because the children in this study were from a number of

different counties, some had scores for other achievement

tests including the California Test of Basic Skills, the

Iowa Test of Basic Skills, the Metropolitan Achievement

Test, and the Stanford Achievement Test. For the purposes

of this study, the different achievement tests were

considered to be equivalent. However, depending on the test

used and the age of the child, different subscores were

available. Thus, the national percentiles for reading,

language, and mathematics, which were available for all

children regardless of age or test, were used in this study.

Parent Report

SNAP-R Rating Scale. The SNAP-R Rating Scale (Swanson,

Nolan, & Pelham, 1982) is a parent report rating scale

consisting of the items from the DSM-III-R diagnoses of ADHD

(Attention-Deficit Hyperactivity Disorder) and ODD

(Oppositional Defiant Disorder); the DSM-III diagnoses of

ADDH (Attention-Deficit Disorder, Hyperactive) and ADDNH

(Attention-Deficit Disorder, Not Hyperactive); four items on

Conduct Disorder (CD) from the ICD-9; and a peer interaction

scale. The parent completing the measure indicated whether

a list of behavioral symptoms described their child not at

all, just a little, pretty much, or very much. These items

are scored 0, 1, 2, 3, respectively. Subscores of

Inattention, Impulsivity, Hyperactivity, ADHD, and ADDNH

were calculated for each child. Norms based on Tallahassee

school children are available for ages 6-11. A cutoff score

of the average of 1.5 for all ages on all subscales is

suggested. A score over this level is considered to be

"high." For the purposes of this study, the items assessing

CD, ODD, and peer interactions were not included.


Demoqraphic questionnaires. Parents were also given a

brief questionnaire on demographics and the child's medical

and academic history. Information requested included parent

education and occupation so that socioeconomic status could

be determined (Hollingshead, 1974?), the child's medication

use, metabolic history, and school absences.


Permission to conduct the study was obtained from the

University of Florida Institutional Review Board, from Janet

Silverstein, M.D., medical director of the Pediatric

Diabetes Clinic, and from the medical director of the

Adolescent Clinic.


Listings of appointments for the Diabetes and

Adolescent Clinics were obtained and attempts were made to

contact parents by telephone prior to their child's

appointment when possible. For control subjects not

involved with the hospital clinics (found largely through

word of mouth in Gainesville and Archer, Florida), parents

were contacted by telephone to schedule an appointment for

testing. When contacted, all parents had the study

explained to them, and participation of their child and a

sibling, if available, was requested.

If the parents initially agreed to the study, screening

questions (e.g., age, duration of diabetes, neurological

history, medication use) were asked to ensure that the child

met the study criteria. If the child met the criteria

except was taking short-acting stimulant medication, it was

requested that the child not be given that medication on the

day of testing. Prior to testing, written informed consent

was obtained from the parent and the child.

Blood Glucose Level

For the children with diabetes, a digital reading of

the child's current blood glucose level was obtained

immediately prior to testing using an Life Scan One Touch

home monitoring device. If a child's blood glucose level

was below 60 mg/dl, the parents and medical staff were

notified and the child was given juice and crackers and

invited to participate on another day provided her/his blood

glucose was above 80 mg/dl. Research has suggested that the

effect of poor concentration due to hypoglycemia may linger

for 4-6 hours after blood glucose levels have increased to

normal levels (Reich et al., 1990). Children whose blood

glucose level was between 60 and 80 mg/dl were asked to eat

a snack (e.g., juice and crackers) and were included in the

study only after their blood glucose level has risen to 80

mg/dl or above. Children without diabetes did not have

their blood glucose level assessed as it was assumed to be


Test Administration

All children were tested in a quiet room without

distractions. For the children with appointments in the

diabetes or adolescent clinic, all testing took place on the

day of the child's clinic appointment in the clinic. Their

siblings, if available, were also tested at the same time.

The other control children were tested either in the

Neuropsychology Laboratory, at their school after school

hours, or at their home.

All children were tested by either the PI or one of her

trained research assistants. All tests were administered in

a standardized manner with the child sitting at a table or

desk in a quiet room. All directions were given from

written scripts to ensure uniformity. To further ensure

standardized administration, each research assistant

completed a "training course" given by the PI which spanned

at least two weeks.

This "training course" consisted of a general

explanation of the study, but did not include specific

information about the hypotheses in an attempt to minimize

examiner bias. For example, none of the research assistants

knew about the distinction between EOD or LOD.

Administration of each assessment instrument was verbally

explained and then demonstrated. Following this, each

assistant practiced the tests on her/his own and in front of

the PI and other assistants. Only after the assistants had

attained proficiency in test administration did they work

with actual subjects.

Test Order

Children participating in the study were first given

the PPVT-R. If they obtained a standard score of > 80, then

they were given the measures of attention in one of three

sequences. The selective attention tests and Trails B were

given together, followed by the CPT, and then the Switching

Attention Test in the first order. The second order started

with the CPT and the third order started with the Switching

Attention Test. This way, the possible effect of sequence

(or fatigue) couldassessed. For each test, the child was

told to work as quickly as s/he could and to inform the

examiner when s/he was done. Testing took approximately 35

minutes per child. Upon completion of testing, each child

was given the opportunity to pick from a box of inexpensive

toys and other items.

Parent and Medical Information

Parents were given a brief questionnaire asking about

demographics, and the child's medical and academic history.

If they endorsed that their child had had diabetes

complications (hospitalizations, seizures, DKAs), specific

information (number and dates of episodes of seizures and

DKAs, blood glucose levels, treatment) was obtained by the

PI. Additionally, they were give the SNAP-R to complete.

These questionnaires typically were completed in

approximately 15 minutes while the child was being tested.

Written consent was also obtained from the parent to request

school records on the child's most recent achievement scores

and history of school absences for the year in which the

achievement test was taken.

Information from the medical charts of the children

with diabetes related to metabolic control was obtained to

augment parent report. This included information on the

number of diabetic ketoacidotic episodes, seizures, and loss

of consciousness. Furthermore, the HAC value, which was

routinely assessed at each clinic visit, was recorded. This

value was used descriptively as it was not expected to be

related to test scores because it is an index of average

metabolic control over a two to four month period rather

than blood glucose levels during the actual testing period

(Johnson, 1990).


All the statistics in this research study were done

using the SPSS-PC statistical program with the aid of the

SPSS User's Guide (1988).

Preliminary Analyses

Standardization of Attention Scores

When preparing raw data for analyses, it is important

to appraise whether direct comparisons can be made between

the groups on the measures. This consideration is

especially important when interpreting the results in this

study which are based on groups of different ages (i.e., EOD

vs. LOD) with scores in different metrics (i.e., reaction

time vs. number of errors). To minimize these differences,

all the raw scores were converted into z-scores based on

normative data. For the scores for which normative data

were not available (e.g., the computerized tests), z-scores

were obtained using the means and standard deviations from

the control group. All the scores, except SDMT, were also

multiplied by (-1) so that higher scores indicate better

performance and lower scores indicate poorer performance.