Differentiating attention deficits in children with fetal alcohol spectrum disorder or attention-deficit–hyperactivity disorder

Authors

  • LIBBE KOOISTRA PHD,

    1.  Department of Pediatrics, University of Calgary, Calgary, Alberta, Canada.
    2.  Behavioural Research Unit, Alberta Children’s Hospital, Calgary, Alberta, Canada.
    Search for more papers by this author
  • SUSAN CRAWFORD MSC,

    1.  Behavioural Research Unit, Alberta Children’s Hospital, Calgary, Alberta, Canada.
    Search for more papers by this author
  • BEN GIBBARD MD MSC,

    1.  Department of Pediatrics, University of Calgary, Calgary, Alberta, Canada.
    2.  Developmental Clinic, Alberta Children’s Hospital, Calgary, Alberta, Canada.
    Search for more papers by this author
  • BARBARA RAMAGE PHD,

    1.  Department of Pediatrics, University of Calgary, Calgary, Alberta, Canada.
    Search for more papers by this author
  • BONNIE J KAPLAN PHD

    1.  Department of Pediatrics, University of Calgary, Calgary, Alberta, Canada.
    2.  Behavioural Research Unit, Alberta Children’s Hospital, Calgary, Alberta, Canada.
    Search for more papers by this author

Dr Libbe Kooistra at Department of Pediatrics, Behavioral Research Unit, Alberta Children’s Hospital, 2888 Shaganappi Trail NW, Calgary, Alberta, Canada T3B 6A8. E-mail: libbe.kooistra@albertahealthservices.ca

Abstract

Aim  The attention and inhibition problems found in children with attention-deficit–hyperactivity disorder (ADHD) are also common in children with fetal alcohol spectrum disorders (FASDs). Attempts to distinguish ADHD from FASDs in terms of these deficits are rare and were pursued in this study.

Method  A total of 116 children (47 with ADHD, 31 males, 16 females; 30 with FASDs, 17 males, 13 females; and 39 comparison children, 20 males, 19 females) participated. The mean age was 9 years 4 months (SD 1y 8mo) in the ADHD groups, 8 years 10 months (SD 1y 2mo) in the FASD group, and 9 years 1 month (SD 1y 1mo) in the comparison group. Sustained attention was tested with a slow event rate continuous performance task (CPT). Inhibitory control was tested with both a slow and fast event rate Go/No-Go task.

Results  On the CPT task, children with ADHD, combined type (ADHD-C), ADHD, primarily inattentive type (ADHD-PI), and FASDs showed greater declines in task performance as a function of time than comparison children, suggesting sustained attention problems in all clinical groups. Children’s Go/No-Go performance was event-rate dependent, with the ADHD-C group being affected in the slow condition and the ADHD-PI and FASD groups having problems with the fast condition.

Interpretation  Children with ADHD-C are typically impaired in handling understimulation, while children with FASDs may have problems with overstimulation. The dissociation in responsivity to event rate between groups may have significant differential diagnostic value.

LIST OF ABBREVIATIONS
CPT

Continuous Performance Task

FASD

Fetal alcohol spectrum disorder

The modern fetal alcohol spectrum disorders (FASDs) concept dates back to 1973, when it was first described as fetal alcohol syndrome (FAS).1 Since then, numerous papers have reported on the neurobehavioural symptoms associated with the disorders, including attention problems, restlessness, and impulsivity.2 Such problems are typically also seen in children with attention-deficit–hyperactivity disorder (ADHD).

The fact that attention and impulsivity problems characterize both FASDs and ADHD has fuelled speculation regarding their overlap, but this possibility has rarely been tested by directly comparing both groups in the same design using neurocognitive tests of attention. Nanson and Hiscock3 compared children with FAS with children with ADHD and a comparison group on a series of attention tasks. While the children with FAS were slower than those in the other two groups, their impulsivity and attention problems were similar to those of the children with ADHD. In addition, Coles et al.4 directly compared the attention profiles of children with FAS and children with ADHD. Children with FAS had difficulties with encoding and shifting attention, while the children with ADHD had problems with focusing and sustaining attention. Interestingly, children with ADHD were found to be impulsive, unlike the children with FAS.

The clinical relevance of the attention and impulsivity problems in ADHD is undisputed. In the search for the etiology of these problems, however, the focus has shifted away from poor attention control. Instead, the leading theories emphasize inhibitory dysfunction5 and inadequate activation6 as the key deficits in ADHD.

With regard to FASDs, there is extensive research on the attention component,7 but there is still ambiguity regarding the nature of the attention deficits, and whether those deficits differ from those found in ADHD.8 As such, the FASD field would benefit from theory-driven research aimed at delineating the inattention symptoms of FASDs in terms of cognitive processes and their brain regions.

The current study investigated children with ADHD and children with FASDs in terms of sustained attention and inhibitory control. Sustained attention was tested with a slow event rate continuous performance task (CPT),9 which is widely used to probe alerting functions associated with right prefrontal brain regions10,11 and brainstem and thalamic structures.12 Slow event rate CPT studies have repeatedly shown greater performance declines as a function of time in children with ADHD than in participants without ADHD, indicating that ADHD is associated with activation and effort allocation problems.13,14

Inhibitory control was tested with a slow and fast event rate Go/No-Go task, which targets both alerting and executive functions supported by frontostriatal and limbic brain areas, and has differentiated ADHD groups from other clinical groups and controls.15,16 The consistent finding has been that Go/No-Go performance deteriorates under slow event rate conditions, supporting the view that a deficiency in effort/activation systems constitutes the key deficit in ADHD.17

Therefore, the present study sought to examine sustained attention and inhibitory control in children with ADHD or FASD from a differential diagnostic perspective. It was expected that children would be differentially affected by time on task as well as event rate manipulation.

Method

Participants

The study was approved by the Conjoint Health Research Ethics Board of the University of Calgary. Consent forms were signed by both parents and children. Participants (n=116) were all Caucasian children, aged 7 to 10 years (47 with ADHD, 30 with FASD, and 39 comparison children; Table I). The children with ADHD were recruited from two private schools and one clinic specializing in learning and attention problems in Calgary, Alberta, Canada. The school principals and clinic director sent letters inviting families with children suspected of having ADHD to participate. Interested parents contacted the researchers for further details. Children had to have been diagnosed with ADHD when aged between 5 and 7 years by a child psychiatrist or developmental pediatrician. A three-step procedure confirmed these diagnoses. First, the Summary ADHD Checklist18 was completed to give an indication regarding the presence or absence of ADHD. Next, the Conners’ Parent Rating Scale-Revised19 was administered to confirm symptomatology. Finally, the Diagnostic Interview for Children and Adolescents-IV20 was used to reconfirm the diagnosis and assign an ADHD subtype. In order to be included children had to meet ADHD criteria on all three measures. Further exclusion criteria were the following: coexisting psychiatric disorders (e.g. oppositional defiant disorder, mood disorder), chronic medical conditions affecting cognitive function (e.g. seizures), and treatment with long-acting psychiatric medication (e.g. risperidone). Of the 47 children with ADHD, 31 were diagnosed as the combined type (ADHD-C), and 16 as having the predominantly inattentive type (ADHD-PI); and 43 were on stimulants. A 24-hour medication washout period was required before testing.

Table I.   Characteristics of study and comparison groups
VariableADHD (n=47)ADHD-C (n=31)ADHD-PI (n=16)FASD (n=30)Comparison group (n=39)
  1. ADHD, attention-deficit–hyperactivity disorder; C, combined; PI, predominantly inattentive; FASD, fetal alcohol spectrum disorders; SES, socio-economic status; FSIQ, Full-scale IQ.

Age y:mo mean (SD)9:4 (1:8)9:0 (1:11)9:8 (0:11)8:10 (1:2)9:1 (1:1)
Males/females 31/16 19/12 12/4 17/1320/19
SES, %
 Low 13.013.312.540.02.8
 Middle 37.036.737.536.727.8
 High 50.050.050.023.369.4
Estimated FSIQ, mean (SD)110.1 (12.8)110.6 (12.9)109.1 (13.1)98.0 (15.5)117.7 (10.0)

All families whose 7- to 10-year-old children attended the FASD clinic at a pediatric hospital were invited to participate by letter from the clinic. Children had to have been identified by a pediatrician as having an FASD. Their classification was based on criteria formulated by the Fetal Alcohol Syndrome Diagnostic and Prevention Network Diagnostic Guide.21 This diagnostic framework provides a four-digit code representing the magnitude of expression (rated 1–4) of the four key FASD features: growth deficiency, FASD facial phenotype, brain dysfunction, and gestational alcohol exposure. Only children who fell in the categories G (sentinel physical findings/neurobehavioural disorder, alcohol exposed; n=5), H (neurobehavioural disorder, alcohol exposed; n=23), M (sentinel physical findings/neurobehavioural disorder, alcohol exposure unknown; n=1) and N (neurobehavioural disorder, alcohol exposure unknown; n=1) were included. Alcohol exposure was deemed to be etiologically significant for all children with alcohol ranks 3 and 4. The records of the two children with unknown alcohol exposure indicated strong evidence of alcohol exposure in utero. A similar three-step screening procedure as described for ADHD was used to verify co-occurring disorders in the children with FASDs. These additional disorders, while recorded, were not exclusion factors. The exclusion criteria for the children with FASDs were a chronic medical condition affecting cognitive function, central nervous system (CNS)-activating medication, and a history of recent child abuse. Twenty-nine of the 30 children with FASDs met criteria for ADHD, all ADHD-C. Twenty-seven of these children were on stimulant treatment and were not given medication for 24 hours preceding testing.

The comparison children were recruited from two elementary schools through posters and parents’ councils. In the initial contact with parents, exclusion criteria were verified, including psychiatric concerns. Next, screening instruments were administered. Only children who scored in the nonclinical range on all screening steps were accepted in the comparison group. They were subject to the same exclusion criteria as their peers with ADHD or FASDs. No data on alcohol use during pregnancy in mothers of children from the comparison or ADHD groups were available because candidate families would have objected to questions on maternal alcohol use during pregnancy.

Screening tools

Summary ADHD checklist

Parents completed the Summary ADHD Checklist18 an instrument based on the Diagnostic and Statistical Manual of Mental Disorders-Fourth edition (DSM-IV) consisting of a 25-item checklist rated on a fourpoint Likert scale. The checklist comprises the following five categories: (1) definitely not ADHD; (2) probably not ADHD; (3) unsure; (4) probably ADHD; and (5) definite ADHD. It has adequate reliability and validity.18

Conners’ parent rating scale – revised (long version)

The Conners’ Parent Rating Scale19 is a standardized DSM-IV-based parent report checklist of a broad range of child and adolescent problems. Its 80 items, scored on a fourpoint scale, represent 14 diagnostic dimensions. A profile based on T-scores permits comparisons with normative age and sex groups. It is among the most prominent ADHD rating scales and has sound psychometric properties.22

Diagnostic Interview for Children and Adolescents-IV, parent version

The Diagnostic Interview for Children and Adolescents-IV,20 a semi-structured computer-assisted diagnostic interview, was administered to parents. It branches to appropriate questions, depending on the respondents’ answers, and includes a DSM-IV-based diagnostic classification module. For this study, the program was configured for the assessment of ADHD, with an additional evaluation of concurrent mood, anxiety, and oppositional symptomatology. It is widely used, has good clinical validity, and moderate to high test–rest reliability.23

Experimental measures

Continuous performance task

Children were asked to fixate on a plus (+) sign in the centre of a computer screen, and to make button-press responses to target stimuli (but not to nontarget stimuli) with their dominant hand. After a variable inter-stimulus interval of 6000, 7000, or 8000ms, the + sign changed into a target (*) or a nontarget stimulus (0). For 25% of trials, a target was displayed, and for 75% of trials a nontarget was displayed. To examine performance decrement over time, the task performance was partitioned into three 9-minute time blocks of 249 trials. The dependent variables were mean response latency, standard deviation of response latency, percentage of errors, subdivided into percentage of omission errors (i.e. misses) and percentage of commission errors (i.e. false alarms), calculated over each 9-minute period.

Go/No-Go task

Children were instructed to make a button-press response to frequent target stimuli (Go trials) with their dominant hand, and to withhold responses to infrequent nontarget stimuli (No-Go trials). Each trial began with a central face with a straight line for the mouth delineating the inter-stimulus interval, which changed to a ‘happy face’ (inline image), indicating a Go, or a ‘frowning face’ (inline image), indicating a No-Go trial. There were two conditions: fast, with inter-stimulus intervals varied between 1000, 1500, and 2000ms; and slow, with variable inter-stimulus intervals of 6000, 7000, or 8000ms. The total session consisted of one fast and one slow condition administered in a counterbalanced order by condition: one participant started with the fast condition, the next participant started with the slow condition. The fast condition consisted of 210 trials (75% Go, 25% No-Go); the slow condition consisted of 60 trials (75% Go, 25% No-Go). Conditions were matched on length: 7 minutes per condition. The dependent variables were response latency, standard deviation of response latency, percentage of errors, subdivided into percentage of omission errors, and percentage of commission errors, calculated for each event rate condition.

Procedure

Testing began with a short form of two subtests (i.e. block design, vocabulary) of the Wechsler Intelligence Scale for Children-III (WISC-III)24 to obtain an estimate of their full-scale intelligence quotient (FSIQ).25 Next, the attention tasks were administered in a counterbalanced order: one participant started with the Go/No-Go task, the next participant with the CPT task. A 5-minute training session preceded each task. Total test time was approximately 105 minutes, with 15-minute breaks between sessions. Socio-economic status (SES) was evaluated by categorizing parental employment using the Blishen index (low SES category 1–2, middle SES category 3–4, high SES category 5–6).26 Assessors were blind to the children’s diagnostic status and to results of other evaluations.

Statistical analyses

Group comparisons on demographics were made using analysis of variance (ANOVA) for continuous variables and χ2 tests for categorical variables. Post-hoc group comparisons were made using the Scheffé test for continuous variables. According to Pedhazur,27 variables that were significantly different among groups were used as covariates for subsequent analyses only if they were significantly and at least moderately correlated (r≥0.30) with the variables in question.

The CPT and Go/No-Go data were analysed using a series of separate repeated-measures ANOVAs followed by contrast analyses using pairwise comparisons. The between-participants factor was group and the within-participants factor was time block for the CPT data; for the Go/No-Go data, group was the between-participants factor and event rate (fast vs slow) was the within-participants factor. Only significant findings (p<0.05) and trends (p<0.10) are reported.

Results

Demographic data

Significant group differences did not emerge for age or sex, but did for estimated Full-scale IQ (F(3,111)=13.61, p<0.001) and SES (χ2(6)=21.88, p=0.001). Full-scale IQ was significantly lower for the FASD group, as was SES (Table I). Full-scale IQ and SES were not moderately correlated with any CPT or Go/No-Go variables (range r=0.07–r=−0.25); thus, in accordance with Pedhazur,27 no covariate was used for subsequent analyses.

CPT data

There was a significant group main effect for response latency (F(3,105)=5.97, p=0.001). Post-hoc group comparisons showed that, overall, the ADHD-C and the FASD groups were slower in terms of response latency than the comparison group (Fig. 1a). A significant group main effect for the standard deviation of response latency (F(3,105)=8.97, p<0.001) followed by post-hoc comparisons indicated that, overall, the ADHD-C group and the FASD group responded more variably than did the comparison participants (Fig. 1b). Significant main effects for group also emerged for the percentage of overall errors (F(3,105)=6.14, p=0.001; Fig. 1c), the percentage of omission errors (F(3,105)=6.43, p<0.001), and the percentage of commission errors (F(3,105)=2.85, p=0.041). Post-hoc comparisons revealed that the ADHD-C and FASD groups were more error prone than were comparison participants, particularly to errors of omission. The ADHD-C group also made significantly more errors of commission (i.e. false alarms) than comparison participants, but the FASD group did not.

Figure 1.

 (a) Response latency (RT) as a function of time for the four groups of children. (b) Standard deviation of RT as a function of time for the four groups of children. (c) Error percentage as a function of time for the four groups of children.

We hypothesized that, if children with ADHD-C, ADHD-PI, or FASDs have sustained attention problems, then the decline in their performance over time should be more pronounced than in comparison children. Group differences in performance decline over time were indeed obtained, as reflected by a significant group by time block interaction for response latency (F(6,208)=4.18, p=0.001). Additional contrast analyses showed that between time block 1 and time block 2, task performance in the ADHD-C, ADHD-PI, and FASD groups declined to a greater extent than in the comparison group in terms of increased response latency (F(1,64)=23.08, p<0.001; F(1,52)=5.94, p=0.018; F(1,63)=7.57, p=0.008 respectively). Similarly, the performance decline in terms of increased standard deviation of response latency was larger in the ADHD-C and ADHD-PI groups than in the comparison group between time block 1 and time block 2 (F(1,64)=4.33, p=0.041; F(1,52)=11.11, p=0.002 respectively). In addition, compared with the comparison group, the increase in the percentage of omission errors between time block 2 and time block 3 was greater in the FASD group (F(1,63)=4.50, p=0.038) and in the ADHD-PI group (F(1,52)=4.24, p=0.045). No difference in the proportion of commission errors emerged from the contrast analyses.

Go/No-Go data

Response latency increased as a function of event rate (F(1,108)=209.32, p<0.001) in all participants regardless of group membership. Three separate repeated-measures ANOVAs showed significant group main effects for standard deviation of response latency (F(3,108)=6.29, p=0.001; Fig. 2a) and total error percentage (F(3,108)=6.17, p=0.001; Fig. 2b), and a trend for response latency (F(3,108)=2.24, p=0.088; Fig. 2c). Post-hoc group comparisons showed that children in the ADHD-C and FASD groups performed more slowly and more variably and made more overall errors than did children in the comparison group. There was no significant group main effect for either omission errors or commission errors.

Figure 2.

 (a) Standard deviation of RT as a function of event rate (fast/slow) for the four groups of children. (b) Error percentage as a function of event rate (fast/slow) for the four groups of children. (c) Response latency (RT) as a function of event rate (fast/slow) for the four groups of children.

Groups were differentially affected by event rate, as evidenced by a significant group by event rate interaction for mean response latency (F(3,108)=2.71, p=0.049). Additional contrasts showed that the ADHD-C group became significantly slower than both the comparison group and the ADHD-PI group in the slow event rate condition (F(1,65)=5.54, p=0.022; F(1,44)=6.80, p=0.012 respectively). By contrast, the ADHD-PI group was slower than the ADHD-C group in the fast event rate condition. The results also showed a significant group by event rate interaction for standard deviation of response latency (F(3,108)=3.14, p=0.028). Contrasts revealed that the ADHD-C group was more variable than the FASD group in the slow event rate condition, and less variable than the FASD group in the fast event rate condition (F(1,57)=6.35, p=0.015). The FASD group became significantly more variable from the slow event rate to the fast event rate condition compared with the comparison group (F(1,64)=6.41, p=0.014). There was no significant group by event rate interaction for overall error percentage or error type.

Discussion

The purpose of this study was to differentiate children with ADHD from children with FASDs in terms of sustained attention and inhibition. The study was set in a theoretical framework which emphasizes difficulty regulating activation through effort allocation as the key deficit in ADHD.28 On the CPT task, the children with ADHD-C, ADHD-PI, and FASDs showed greater declines in task performance over time than did comparison children, as indicated by a significant group by time block interaction for response latency, suggesting sustained attention problems in all three groups. On the Go/No-Go task, groups were differentially affected by event rate. Only the ADHD-C group showed a performance decline in the slow event rate condition, becoming slower and more variable in it. By contrast, the ADHD-PI and FASDs groups both showed a performance decline in the fast event rate condition, with the ADHD-PI group becoming slower and the FASD group more variable. This difference in susceptibility to slow/fast pacing may constitute a distinct differential diagnostic marker that identifies the possibility of different attention/inhibition profiles in children with FASDs compared with children with ADHD.

Many CPT studies have shown that ADHD and sustained attention problems are associated.29 Similarly, FASDs have been linked with sustained attention deficits.3,7,30,31 However, many of these studies are based on only a single overall performance score, thereby disregarding the fact that sustained attention deficits develop as a function of time.17 Moreover, most studies used relatively short task durations in combination with medium speed stimulus presentation times. Long task durations and more extreme presentation rates are recommended to create enough task difficulty to show sustained attention deficits develop as a function of time.17 The current study corrected these limitations. The findings show that the sustained attention deficits in ADHD appear to be associated with an inability to maintain a proper balance between task demands over time and required activation. Further CPT research is necessary to resolve whether the presumed activation modulation problems in ADHD are also involved in FASDs. Findings suggesting that impaired arousal modulation may underlie the poor vigilance performance of children with FAS,3 and that children’s sustained attention problems occur in slow-paced conditions,7 signal the need for incorporating arousal modulation variables (e.g. time on task, reward, event rate) in future FASD research.

It has been shown repeatedly that the Go/No-Go performance in children with ADHD is highly sensitive to stimulus presentation rate. They perform more slowly, and with more variability and errors, especially in slow conditions.17 As such, our finding of response slowness and variability in the slow condition in the ADHD-C group concurs with the literature, suggesting an activation regulation problem.16 Interestingly, the task performance of the ADHD-PI subgroup was affected by fast event rate. Several studies have emphasised that children with ADHD have difficulties regulating their activation levels, which may arise in very slow and/or in very fast conditions.28 As such, the performance decline of the ADHD-PI group in the fast event rate condition compared with the ADHD-C group supports the notion that children with ADHD are overall constrained in their abilities to handle both extreme underactivation and extreme overactivation.

Obviously, the performance inaccuracy we observed in the children with FASDs in the fast event rate condition awaits further confirmation. It is, nevertheless, tempting to speculate that their problems could arise from a similar deficiency in activation regulation, which becomes especially manifest in high event rate situations causing overactivation. Not finding differences in children’s error profiles as a function of event rate was unexpected, and is not in line with the literature showing that children with FASDs or ADHD are prone to making impulsive errors of commission.2,3 A possible reason for this discrepancy may be that the overall number of trials in the slow condition was limited, thereby reducing the opportunity for making impulsive errors.

Neuroimaging data to validate the event rate phenomenon in ADHD or FASDs are lacking. However, given that CPT and Go/No-Go paradigms are established probes of the alerting and executive control networks,10,11 our data converge with several imaging studies32 showing an association between poor inhibitory control in ADHD and frontostriatal dysfunction. Notably, a similar association in children with FASDs was recently obtained, and interpreted as signaling the need for extra effort allocation to mitigate decreased frontostriatal efficiency.33

One may argue that in both our tasks successful performance was dependent on motivation and effort allocation requiring intact frontolimbic circuits including the thalamus. Thalamic activation plays a key role in preserving task performance by modulating arousal in proportion to task demands.34–36 Thalamic dysfunction, therefore, could explain the deficits in effortful attention observed in the current study.

One limitation of this study is that 29 of the 30 children with FASDs also met criteria for ADHD. While some consider the optimal design to include a group of children with FASDs without co-occurring ADHD, the clinical reality is that essentially all children with FASDs struggle with concomitant ADHD symptoms. It is noteworthy that, despite the diagnostic overlap, our Go/No-Go paradigm was sufficiently sensitive to differentiate between the ADHD and FASD groups. Another issue worth mentioning relates to the relatively high IQ level of our FASD sample. By only including children who fell in the G, H, M, or N diagnostic categories,21 a deliberate choice was made for relatively high-functioning children.

Overall, our CPT data demonstrate that both FASDs and ADHD are associated with sustained attention deficits, which may have repercussions for school and work performance. Most importantly, however, children’s Go/No-Go performance was event rate dependent, with the ADHD-C group being affected in slow-paced conditions and the ADHD-PI and FASD groups having problems in fast-paced conditions. This dissociation in responsivity to event rate between groups may have significant differential diagnostic value. It denotes the possibility that FASDs, rather than being a form of ADHD, have a distinct behavioral phenotype.

Acknowledgements

We thank the Alcoholic Beverage Medical Research Foundation (ABMRF) for funding this study and Mahendra Yatawara for his assistance in data collection and processing. We also thank the Alberta Children’s Hospital Foundation for their ongoing support, and the Foothills Academy for the opportunity to recruit children.

Ancillary