Summary of Findings

This study aimed to explore the relationship between pregnancy complications and performance on measures of verbal fluency in children with ADHD. Out of a larger sample of 199 children, 34 children met inclusion criteria. The majority of children were males, Hispanic, school-aged, and in 5th grade (table 2-1). Three separate tasks of verbal fluency were administered in order to determine whether verbal fluency may have been particularly impacted by maternal pregnancy complications, and to determine performance variance. The with and without pregnancy complication groups had comparable demographic characteristics in terms of gender, age, ethnicity and years of education, which were not shown to mediate the effect of pregnancy complications on children’s performance among the three tasks.

To the best of our knowledge, this study is the first retrospective study exploring the relationship between ADHD and verbal fluency with pregnancy complications as an independent variable. As such, our study results are not entirely comparable with current literature without considering the impact of pregnancy complication as an additional performance variable. Our findings support the notion that pregnancy complications impact the performance of children with ADHD on measures of verbal fluency. A significant difference was found between groups in semantic fluency, which is comparable to performance patterns identified by Mahone et al.⁴⁰ and Takács et al.⁴³. Deficits in category fluency were also found to be similar to the findings of Takács et al.⁴³, where significant difference in performance for the category switching task was observed. There could be a number of reasons why no significant performance difference was observed on letter fluency. It is possible that semantic skill development is more significantly impacted by developmental alterations during gestation. The absence of significant performance difference between groups on the letter fluency task can be explained by variability in the level of impact of pregnancy complications during gestation as well as difference in timing during gestation when mothers experienced complications. There is also the possibility that this particular verbal fluency task may be more sensitive to age, resulting in better performance over time. Given that close to 50% of our sample identified as Hispanic, it is possible that bilingualism and differences in language skills development may have played an important role in this ethnic group’s performance on semantic fluency tasks.

Implications

Given the established association between pregnancy complications and neurodevelopmental deficits, significant differences were expected between groups. In terms of overall performance, our findings were commensurate to the current literature where the without pregnancy complication group performance was consistently better than the with pregnancy complication group. However, as mentioned before, none of the existing literature included pregnancy complications as an independent variable, thus limiting our ability to compare our results to previous findings. Our results found common trends with Cohen et al.³⁹ in that, although performance differences between both groups were identified, verbal fluency skills were not found to be impaired. With the exception of category switching, where the pregnancy complication group mean performance was very low, performance across groups ranged between low average to average.

In terms of ADHD performance patterns, the with pregnancy complication groups suggested clinically relevant implications for letter fluency (low average) and category switching (very low). That being said, our findings showed variability when compared to the present literature between ADHD and verbal fluency performance. For example, our outcomes differ from those observed by McGee et al.⁵⁷ who did not identify an association between verbal fluency deficits and attention deficit disorder (ADD). Differences between outcomes can be explained by heterogeneity in sample population, where this research group was made up of only male participants, whereas our group, although mainly comprised by males, also included 20% females. Additionally, the age range for this study only included boys of 13 years of age, whereas our study had a range between 8 and 18 years of age. Interestingly, and different from current literature, pregnancy complication did not influence the performance of children on letter fluency tasks, suggesting that pregnancy complications were not a predictor of letter fluency dysfunction for this group. This particular finding is of great significance given the neurodevelopmental association of letter fluency with executive control as a frontal lobe function, which is consistent with the etiology of ADHD. However, it is important to highlight that the average performance for both groups on the letter fluency task was relatively lower when compared to category fluency task, but better than the category switching task. Letter fluency performance findings differ from deficits identified by Mahone et al.⁴⁰. This difference could be attributed to the use of a different verbal fluency measure (NEPSY-I) whereas this study utilized the DKEFS’s Verbal Fluency subtests.

The current study obtained different outcomes when compared to findings by Hurks et al.⁴¹, who identified less developed letter fluency ability, particularly in the first 15 seconds of the task in the ADHD group. Because verbal fluency performance was coded as a scale score in our database, individual performance on each letter was not tracked. Therefore, we are unable to compare performance results on a letter by letter basis. Our letter fluency findings were also different to those obtained by Abreu et al.⁴² in that our groups did not evidence less developed letter fluency ability. Nevertheless, we did not assess an in-between letter comparison to identify difference in performance between letters.

Our results were also diverse in the category fluency task and were different from Hurks et al.⁴¹, who found no interaction between SCF performance and ADHD when compared to controls. Differences in results can be attributed to the absence of pregnancy complication as a variable from the analysis between groups. In contrast to Mahone et al.⁴⁰, who found weaker verbal fluency performance in an ADHD sample, our results did not indicate less developed semantic fluency ability in any of the two groups when compared to the letter fluency and category switching tasks. In fact, both groups evidenced stronger performance in the category fluency task when compared to the other verbal fluency tasks. Our outcomes are also partially different from those found by Takács et al.⁴³ who identified lower semantic performance on an ADHD sample, whereas we did not. These differences in results can be better explained by variability in their sample’s age range (8 to 12 years old); while our sample ranged from 8 to 18, suggesting that performance in this task may improve over time for this population. Our results were similar to McGee et al.⁵⁷ who identified no deficits in verbal fluency in children with ADHD. Cross-sectional studies indicate that word fluency improves with increasing age to at least 13 years, with written word fluency increasing up to 18 years⁵⁸. It is possible that the response organization and processing efficiency of children in the study who were younger than 9 years of age, the age at which frontal lobe changes related to executive function are believed to emerge, had not fully developed at the time the assessment was completed. The frontal lobe (FL) dysfunction model has been used to conceptualized ADHD given that the performance of children with ADHD differed significantly from normal controls on measures of frontal lobe function⁵⁹. However, our results did not support this theory, and yielded opposite results to Moscovitch’s ⁶⁰ hypothesis that patients with frontal lobe lesions are impaired on tests of letter but not category fluency.

In summary, our results evidenced important similarities as well as differences to findings in the current literature, specifically in the category fluency and category switching tasks. This variability in outcome is mainly attributed to differences in verbal fluency measures utilized to assess verbal fluency performance in previous studies as well as differences in age and gender among sample groups.

Limitations

The most important limitation of the present study is its retrospective nature. Having limited control over adherence to standardized test administration, proper documentation of maternal pregnancy complications, participants’ bilingual skills, and language functioning poses as significant drawback to maintaining uniformity among participants. Another limitation is the relative small size (N=34) of the ADHD sample meeting inclusion criteria. We believe the current study’s small sample size limits the generalization power of our results. Because the reporting of maternal pregnancy complications was conducted in a non-standardized fashion, there was significant variability in the way maternal pregnancy complication history was documented. This contributed to significant inconsistency among participants in terms of pregnancy complication history and specific medical conditions experienced by mothers during pregnancy. Participants were therefore unable to be grouped by medical condition or pregnancy complication type, and thus performance patterns within specific conditions could not be evaluated. The sample was not matched in terms of age and gender for the purpose of obtaining adequate sample size. As such, there is a rather large age-range variance between groups.

Although almost all of the average scores from the present sample are within a standard deviation from the standardization sample and clinically significant impairment was not observed, except in category switching, the study did not control for participants’ level of language functioning. Therefore, it is possible that variable language performance exists across both groups, thus impacting the results of the study, particularly in category switching. Given the diverse demographical characteristics of the sample and the geographic location where the assessment was conducted, it is no surprise that 40% of the sample was identified as Hispanic. Also a result of the socio-cultural characteristics of the geographic location is the expected bilingual nature of children’s language abilities. Because verbal fluency has also been identified as a function of verbal ability, language skills will play an important role in performance on these tasks. Although all children were assessed in English, following identification of language dominance through standard language proficiency testing, information on children’s bilingual skills and number of languages spoken at home was not available, and thus the extent by which bilingualism influenced performance on verbal fluency is unknown.

One of the biggest limitations involved inconsistencies found in the recording of maternal pregnancy history. For future research and clinical practice, the intake process should ensure consistency in how pregnancy history is documented, including specific questions added to existing and future parent history questionnaires to be filled out at initial contact. Future research on verbal fluency and pregnancy complications ought to be conducted within groups with specific pregnancy complications to identify patterns of performance within each pathology. Given our knowledge on language functioning and its identified relationship with category fluency performance, it will be crucial to control for language functioning to explore correlations and differences in performance in the presence of language dysfunction.

Conclusion

This study was founded on the understanding of atypical developmental expression as the result of changes experienced in utero that manifest throughout the lifespan. Much of what we know about pregnancy complications and neurodevelopmental deficits is grounded on neurobiological changes occurring at the cellular level during gestation. Because such changes occur at different times across the developmental spectrum, it is challenging to isolate with certainty which neurobiological mechanisms are responsible for atypical development in specific areas of functioning, resulting in a much more complex neuro-behavioral developmental assessment process. One of the important areas of human function that is believed to be impacted during gestation is executive control, guiding the focus of our study on defining the relationship between pregnancy complications and verbal fluency performance in a group of children with ADHD. The goal of this study was to identify the impact of maternal pregnancy problems on verbal fluency measures in children with ADHD. Age, gender, education, and ethnicity were not found to impact performance. A comparative analysis of covariance and subsequent test of between subject effect was conducted to identify specific variance was conducted. Broadly, children with ADHD without maternal history of pregnancy complications demonstrated better performance across all tasks. Our findings suggest a significant difference in performance on verbal fluency tasks when pregnancy complications are present, specifically on tasks of semantic fluency and set-shifting.

Although category fluency is one of the skills evidenced to be significantly different, this could be the result of the category fluency task having more resemblance to how language is stored in the brain, thus being more sensitive to language development variations during gestation. Because semantic fluency is associated with language function, future research must include language performance on ADHD cohorts to rule out other contributing factors impacting semantic fluency. Age, gender, education, and ethnicity did not seem to influence the effect of pregnancy complications on any of the verbal fluency tasks. Further investigation of verbal fluency performance within specific pregnancy complications sub-groups remains necessary

As observed in the existing literature, studies exploring the relationship between pregnancy complications and ADHD evidence clear differences in geographical location, timing, age ranges, and prenatal medical conditions. However, a consistent factor among all studies was the assessment method selected by researchers to assess for ADHD, which only included self-report measures from parents and teachers across all ages. Given the limitations previously discussed, future research should look at conducting similar studies with a larger sample size, as this will permit for results to be generalized to the rest of population.