Executive functioning and verbal fluency performance in youth with Down syndrome

Abstract

Background:

Executive functioning (EF) is an area of challenge for individuals with Down syndrome (DS) associated with a variety of downstream difficulties. Verbal fluency performance is one potential downstream effect that is commonly assessed in individuals with DS due to the measure’s utility as a predictor of dementia. Verbal fluency requires individuals to inhibit irrelevant responses, shift between groupings of related words, and monitor to prevent repetition, all skills related to EF.

Aims:

This study aimed to determine the association between semantic verbal fluency performance and three EF subdomains (inhibition, shifting, and working memory) in youth with DS after taking into account vocabulary and cognitive ability.

Methods and procedures:

Neuropsychological assessments (verbal and visuospatial), and parent reports of EF, were completed at one time point by 69 youth with DS 6 to 17 years old and their caregivers. Expressive and receptive vocabulary skills and cognitive ability were also assessed.

Outcomes and results:

The results revealed that verbal fluency performance was significantly associated with neuropsychological assessments of EF and parent report of inhibition even after controlling for the effects of vocabulary and cognitive ability.

Conclusions and implications:

The findings highlight the underlying importance of EF in verbal fluency tasks in youth with DS.

1. Introduction

Executive functioning (EF) is an area of challenge for individuals with Down syndrome (DS) (Daunhauer & Fidler, 2013). EF refers to a set of top-down cognitive processes integral to adaptive and goal-directed actions. The ability to control one’s attention and behavior (inhibition), the flexibility of thought and action (shifting), and simultaneous storage and processing of information (working memory) are three central EF subdomains (Miyake et al., 2000). Relative to mental age, children and adults with DS show substantial challenges in most EF processes (Daunhauer et al., 2017; Daunhauer et al., 2020; de Weger et al., 2021; Iralde et al., 2020; Tungate & Conners, 2021). EF abilities in individuals with DS are associated with a variety of downstream difficulties in social behavior (Amadó et al., 2016), communication (Kristensen et al., 2022), academic achievement (Will et al., 2021), and adaptive behaviors (Esbensen et al., 2021). There is a need to understand other potential downstream challenges for individuals with DS to guide intervention targets. Verbal fluency is one potential area of EF’s downstream effects, which is commonly assessed in adults with DS due to the measure’s utility as a predictor of dementia (Chen & Ringenbach, 2019; del Hoyo et al., 2015).

Verbal fluency refers to the ability to generate as many spoken words as possible in a given time. Verbal fluency performance is typically assessed using tasks in which participants produce words quickly according to either semantic (words in a specific category such as foods or animals) or phonemic (words beginning with a specific letter) prompts (Korkman et al., 2007). Although both semantic and phonological verbal fluency tasks have been used in previous studies for participants with DS (Nash & Snowling, 2008; Stavroussi et al., 2016), it has been recently reported that the phonological task is not feasible and reliable for use in children with DS (Smeyne et al., 2022). However, the semantic task is appropriate for children with DS who are 10 years and older (Smeyne et al., 2022).

Lower scores on verbal fluency tasks have been observed in children with DS compared to children without disabilities matched for receptive language abilities (Nash & Snowling, 2008) and children with unspecified intellectual disabilities matched for receptive vocabulary and chronological age (Stavroussi et al., 2016). Verbal fluency performance also declines with age in adults with DS (Ghezzo et al., 2014). Challenges with verbal fluency performance in individuals with DS has received considerable attention given its relevance to the early onset of Alzheimer’s disease in this special population (Chen & Ringenbach, 2019; del Hoyo et al., 2015). Specifically, it has been found that semantic verbal fluency performance is associated with a biomarker of Alzheimer’s disease (plasma levels of Aβ₄₂) in adults with DS (del Hoyo et al., 2015).

Verbal fluency performance is thought to rely on the higher-order cognitive processes involved in EF (Filippi et al., 2022). Generating a high frequency of words on verbal fluency tasks typically depends on the ability to inhibit irrelevant responses, shift between groupings of related words, and monitor responses to prevent repetition. Generally, the association between EF and verbal fluency performance has been supported by previous studies conducted among typically developing participants (Aita et al., 2018; Amunts et al., 2020; Amunts et al., 2021; Filippi et al., 2022; Gligorović & Buha, 2014; Gonzalez-Burgos et al., 2019; Gustavson et al., 2019; Henry et al., 2015; Shao et al., 2014). A set of inhibition, shifting, and working memory measures were significantly correlated with semantic verbal fluency tasks in a sample of healthy adults (Amunts et al., 2020), and also latent variables of shifting and working memory were significantly associated with a general fluency factor (comprising both phonemic and semantic fluency measures) in samples of adolescents and adults (Gustavson et al., 2019). Similarly, shifting and working memory factors were significantly associated with both semantic and phonemic verbal fluency measures in a college-aged sample of healthy participants (Aita et al., 2018). Inhibition and working memory had significant relations with semantic and phonemic verbal fluency tasks in older adults (Shao et al., 2014). In addition to the behavioral studies, neuroimaging studies have reported similar brain regions, specifically the prefrontal cortex, involved in both verbal fluency and EF-related tasks supporting the association between EF and verbal fluency (Baldo et al., 2001; Ghanavati et al., 2019; Paek et al., 2020; Riello et al., 2021; Robinson et al., 2012; Sun et al., 2017; Yuan & Raz, 2014).

The link between EF and verbal fluency has also been described among participants with neurodevelopmental conditions, although some mixed results have been reported regarding the associations of three EF subdomains (inhibition, shifting, and working memory) and verbal fluency tasks. Both semantic and phonemic verbal fluency tasks were significantly correlated with inhibition but not with working memory and shifting in a combined sample of children with and without language difficulties (Henry et al., 2015). Among 10–13-year-old children with unspecified mild intellectual disabilities, shifting, but not working memory or inhibitory control was associated with a semantic verbal fluency task (Gligorović & Buha, 2014). In a sample of adults with DS, the total words produced in the semantic verbal fluency task was significantly associated with a component process of working memory, verbal short-term memory (Stavroussi et al., 2016). Verbal fluency performance was also associated with attention deficit hyperactivity disorder (ADHD) symptoms (Tang et al., 2022) and children, adolescents, and adults with ADHD demonstrated lower verbal fluency performance than their typically developing peers (Abreu et al., 2013; Hurks et al., 2004; Koziol & Stout, 1992; Takács et al., 2014).

Besides relying on executive control of lexical retrieval, verbal fluency is also a key component of human language processing and verbal fluency tasks have a substantial verbal component (Henry & Crawford, 2004; Isacoff & Stromswold, 2014). The underlying importance of language abilities in word generation tasks like verbal fluency has been emphasized in previous studies (Delgado-Álvarez et al., 2021; Henry & Crawford, 2004; Isacoff & Stromswold, 2014; Whiteside et al., 2016). Specifically, it has been reported that children with language difficulties have lower scores in verbal fluency tasks compared to typically developing peers (Henry & Crawford, 2004).

Given the difficulties in verbal fluency and EF performance reported in individuals with DS, a critical gap exists in understanding the associations between EF and verbal fluency in this population. This study aimed to investigate the associations between semantic verbal fluency and verbal, visuospatial, and parent reports of EF subdomains (inhibition, shifting, and working memory) in youth with DS. We investigated these associations taking into account the effects of vocabulary and cognitive ability. We hypothesized that receptive vocabulary, expressive vocabulary, cognition, and EF subdomains would be significantly related to verbal fluency performance. We expected that verbal, visuospatial, and parent reports of EF subdomains would have independent significant contributions to semantic verbal fluency performance after partialing out the variance explained by vocabulary and cognitive ability. Comparing the cognitive correlates of verbal fluency with traditional EF measures may better characterize the role of EF in verbal fluency performance.

2. Method

2.1. Study Design

Study procedures were approved by the Institutional Review Board (IRB) and the Streamlined, Multisite, Accelerated Resources for Trials (SMART) IRB platform at the supervising medical center. Participants were recruited through a pediatric medical center, university, and local Down syndrome associations. To participate, youth were required to have a diagnosis of DS and an approximate developmental age of at least three years. Families were required to use English as their primary language. The participants were directly assessed on measures of verbal fluency, cognition, EF, and vocabulary at one time point, and one parent provided information on the participant’s demographics and EF abilities.

2.2. Participants

The sample included 69 children and adolescents with DS, between the ages of 6–17 years old (M = 13.9 years, SD = 2.9). Only participants able to complete the verbal fluency task were included in the current study. The 25 children who participated in the larger study and who were unable to complete the verbal fluency task did not differ in demographic characteristics from the 69 children included in the current sample. There was an approximately equal distribution by gender (39 [56.5%] males, 30 [43.5%] females). Participants identified as predominantly White (92.8%), or Black (7.2%). Abbreviated IQ composite scores of the participants measured by Stanford-Binet Intelligence Scales, Fifth Edition (SB-5, Roid, 2003) ranged from 47 to 76 (M = 49.4, SD = 6.0).

2.3. Measures

2.3.1. Verbal Fluency

The Word Generation semantic trial from the Neuropsychological Assessment of Children, 2nd Edition (NEPSY-II), was used to measure verbal fluency performance (Korkman et al., 2007). NEPSY-II is a comprehensive battery purported to measure several neurocognitive functions in children between 3 and 16 years of age (Korkman et al., 2007). In the NEPSY-II semantic word generation task, the child is given a semantic category (animal and food/drink) and asked to produce as many words as possible in 60 seconds. Before beginning the trials, two example responses are provided (Korkman et al., 2007). The feasibility and reliability of word generation semantic trials were high and the task has been recommended for use with children with DS 10 years and older (Smeyne et al., 2022). The current analyses included children younger than 10 years that were able to do the verbal fluency task. In the current study, the number of correct responses across the semantic animal and food/drink trials was summed for use in the analyses.

2.3.2. Executive Functioning

2.3.2.1. Verbal Measures.

Three verbal measures of inhibition, shifting, and working memory were selected, including the Cat/dog Stroop task, Rule Shift task, and SB-5 Verbal Working Memory, respectively. The Cat/dog Stroop task requires the examinee to verbally label cats as dogs and dogs as cats (Ball et al., 2008). There were 8 cats and 8 dogs to label on the 12 × 12-inch assessment card. The number of correct responses and the time on the task are recorded by the examiner. To produce a more reliable measure through speed-accuracy integration (Bakun Emesh et al., 2021), the number of correct responses divided by the time spent on the tasks was calculated for each participant and used for the analyses. The Rule Shift task requires the participant to learn a task requiring a verbal response (name color) and to then shift to a new rule (identify if color is the same as previous) also requiring a verbal response (Lanfranchi et al., 2010). Colored cards for this assessment were approximately 2 × 3 inches. The number of correct and incorrect responses and the time to complete the task are recorded. Again, to produce a more reliable measure, the number of correct responses divided by the time spent on the tasks was calculated for each participant and used for the analyses. The Verbal Working Memory is a subdomain of the SB-5 (Roid, 2003). The participant is required to repeat a simple sentence or listen to a series of sentences and then sorting out the last word in each sentence for recall. Given the lack of variance in the SB-5 scaled scores, the SB-5 Verbal Working Memory raw score was used in the statistical analyses. All three tasks (Cat/dog Stroop, Rule Shift, and SB-5 Verbal Working Memory) have been used or recommended for use among individuals with DS (Cody et al., 2020; Esbensen & Hoffman, 2018; Esbensen et al., 2017; Kristensen et al., 2022; Schworer, Esbensen, et al., 2021). In all verbal EF measures, higher scores reflect better performance.

2.3.2.2. Visuospatial measures.

Three visuospatial measures of inhibition, shifting, and working memory were selected, including the NEPSY-II Statue task, Weigl task, and the SB-5 Nonverbal Working Memory, respectively. The NEPSY-II Statue task was originally developed for typically developing children ages 3–6 years to require the participants to remain still when presented with distracting stimuli (Korkman et al., 2007). All children were scored as 6-year-olds, the oldest age range for generating T-scores. T-scores were used in analyses. The Weigl task requires the examinee to sort cards and then shift to sorting cards in different formats. A scoring algorithm accounts for the ability of the child to sort by shape or color and to shift between these two methods of sorting, with scores ranging from 0 to 5 and higher scores reflecting greater ability with shifting between sorting formats (Grant & Berg, 1948). In the SB-5 Nonverbal Working Memory, the participant is asked to find an object hidden under a cup, and then, at higher levels, repeat block tapping sequences (Roid, 2003). Given the lack of variance in the SB-5 scaled scores, the SB-5 Nonverbal Working Memory raw score was used in the statistical analyses. Higher scores indicate better performance in all three visuospatial EF measures.

2.3.2.3. Parent Report.

Inhibition, shifting, and working memory subscales of the Behavior Rating Inventory of Executive Function, second edition (BRIEF-2), were used as the parent-reported measure of EF (Gioia et al., 2015). The BRIEF-2 was designed to assess EF skills in 5–18-year-old children and adolescents (Gioia et al., 2015). Parents are asked to rate their child’s behaviors on a three-point scale (i.e., never, sometimes, often). Scores are standardized based on age and gender, with a mean T-score of 50 and a standard deviation of 10. Higher scores reflect more challenges with EF. BRIEF-2 measures nine subdomains of EF including Inhibition, Self-Monitor, Shift, Emotional Control, Initiating, Working Memory, Plan/Organize, Task-Monitor, and Organization of Materials. Test-retest reliability estimates of the nine scales ranged from .67 to .92 with a mean of .79 (Gioia et al., 2015). Specifically, the BRIEF-2 is recommended as an appropriate and reliable tool for use in studies involving participants with DS (Esbensen et al., 2019; Esbensen et al., 2017).

2.3.3. Vocabulary

2.3.3.1. Receptive Vocabulary.

The Peabody Picture Vocabulary Test, Fourth and Fifth Edition (PPVT-4 and PPVT-5), were used to assess receptive vocabulary (Dunn, 2019; Dunn & Dunn, 2007). The PPVT-4 and PPVT-5 are untimed individually administered tests developed to measure receptive vocabulary of 2.5 through greater than 90-year-old English-speaking individuals. The examinee is shown four colored pictures and asked to choose the one that best matches the word orally presented by the examiner. Excellent overall reliability of the normative sample and test-retest reliability (all ages) have been reported by the test developer (Dunn, 2019; Dunn & Dunn, 2007). Both PPVT-4 and PPVT-5 have been widely used or recommended for use in studies involving individuals with DS (Esbensen et al., 2017; Hartley et al., 2017; Kristensen et al., 2022; Lao et al., 2017; Schworer, Belizaire, et al., 2021; Schworer, Esbensen, et al., 2021; Schworer, Hoffman, et al., 2021). In the current study, PPVT-4 standard scores were used for the first 16 participants, and when published, the PPVT-5 standard score was used for the remaining participants. The standard scores were used for the analyses and higher scores indicate better performance.

2.3.3.2. Expressive Vocabulary.

The Expressive Vocabulary Test, Second and Third Edition (EVT-2 and EVT-3) were used to assess expressive vocabulary (Williams, 2007, 2019). The EVT-2 and EVT-3 are individually administered tests developed to measure the expressive vocabulary of 2.5 through greater than 90-year-old English-speaking individuals (Williams, 2007, 2019). The examinee is given pictures and asked to label them or provide a one-word synonym for them. Previous studies have used EVT-3 as valid and reliable tools to assess expressive vocabulary in children and adults with DS (Kristensen et al., 2022; Schworer, Belizaire, et al., 2021; Schworer, Esbensen, et al., 2021; Schworer, Hoffman, et al., 2021), and the EVT-2 has been recommended for use among individuals with DS (Esbensen et al., 2017). In the present study, the first 16 participants were assessed by EVT-2, and, when published, the EVT-3, was used for the remaining participants. The standard scores were used for the analyses and higher scores reflect better performance.

2.3.4. Cognitive Ability (IQ)

Stanford-Binet Intelligence Scales, Fifth Edition, Abbreviated Battery Intelligence Quotient (SB-5 ABIQ) was used to assess participants’ cognitive abilities (Roid, 2003). The SB-5 ABIQ is a strongly reliable instrument measuring two major areas of intelligence including Nonverbal Fluid Reasoning and Verbal Knowledge in children and adults from 2 to 85 years (Roid, 2003). Regarding the significant floor effect of the SB-5 ABIQ observed among individuals with intellectual disabilities, Sansone et al (2014) developed the SB-5 ABIQ deviation scores using the method of raw Z score transformation (Sansone et al., 2014). The modified scores have been used in recent studies in different samples of individuals with intellectual disabilities specifically those with DS (Schworer et al., 2022; Schworer, Hoffman, et al., 2021; Thurman et al., 2021). In the current study, the original SB-5 ABIQ scores were used to describe the range of the participants’ IQ. The SB-5 ABIQ deviation scores were included in the analyses to examine associations with other variables.

3. Results

Data were screened to detect any univariate and multivariate outliers. No data point was identified as outliers (Z Scores <|3.29), and no participants were identified as multivariate outliers (Field, 2013). Table 1 provides the mean, standard deviation, range, skewness, and kurtosis of all variables. The values of skewness and kurtosis were between +1 and −1 for most variables except for expressive vocabulary and visuospatial shifting (Weigl) which were approximately between +2 and −2.

Table 1.

Descriptive statistics for verbal fluency, executive function, vocabulary, and cognition

	Mean	SD	Minimum	Maximum	Skewness	Kurtosis
Verbal Fluency
NEPSY-II Verbal fluency	20.2	8.2	1	41	.38	−.16
Verbal assessment
Cat/dog Stroopa	0.6	0.3	0	1.4	.50	.45
Rule shiftb	0.4	0.2	0.1	0.9	.24	.21
SB5 Verbal Working Memory	7.7	4.2	0	21	.56	.39
Visuospatial assessment
NEPSY-II Statuea	6.5	3.9	1	14	.25	−.91
Weiglb	2.5	1.6	0	5	.40	−1.26
SB5 Nonverbal Working Memory	11.8	4.6	1	23	−.26	−.25
Parent report
BRIEF-2 Inhibit	56.3	10.8	39	86	.77	.54
BRIEF-2 Shift	59.9	11.5	39	86	.10	−.97
BRIEF-2 Working memory	61.1	7.9	44	85	.42	.57
Vocabulary
PPVT	52.0	14.4	20	87	.14	−.15
EVT	62.2	12.5	20	85	−1.01	2.43
Cognition
SB-5 ABIQ Deviation scores	39.4	15.2	12.8	76.9	.48	−.10
SB-5 ABIQ Standard scores	49.4	5.0	47	76	3.25	11.09

SD: Standard Deviation; NEPSY-II: Neuropsychological Assessment of Children, 2nd Edition; EVT; BRIEF-2: Behavior Rating Inventory of Executive Function, second edition; EVT: Expressive Vocabulary Test; PPVT: Peabody Picture Vocabulary Test; SB5-ABIQ: Stanford-Binet Intelligence Scales, Fifth Edition, Abbreviated Battery Intelligence Quotient;

^aInhibition measure;

^bShifting measure.

Table 2 presents correlation coefficients between each pair of the variables. Verbal fluency was significantly correlated with all neuropsychological assessments of EF, expressive vocabulary, receptive vocabulary, cognitive abilities, and parent-report of inhibition. There were no significant associations between verbal fluency and parent reports of shifting and working memory.

Table 2.

Correlations among verbal fluency, executive function, vocabulary, and cognition

		1	2	3	4	5	6	7	8	9	10	11	12
1	NEPSY-II Verbal fluency	-
Verbal assessment
2	Cat/dog Stroopa	.51**	-
3	Rule Shiftb	.50**	.45**	-
4	SB-5 Verbal Working Memory	.48**	.49*	.28*	--
Visuospatial assessment
5	NEPSY-II Statuea	.29*	.14	.12	.35**	-
6	Weiglb	.46**	.36**	.36**	.49**	.19	-
7	SB-5 Nonverbal Working Memory	.51**	.47**	.40**	.64**	.41**	.68**	-
Parent report
8	BRIEF-2 Inhibit	−.37**	−.19	−.03	−.38	−.51**	−.30*	−.32**	-
9	BRIEF-2 Shift	−.19	−.09	−.10	−.23	−.26*	−.14	−.22	.54**	-
10	BRIEF-2 Working Memory	−.20	−.07	−.07	−.12	−.19	−.18	−.17	.58**	.56**	-
Vocabulary & Cognition
11	PPVT	.32**	.34**	−.01	.52**	.12	.37**	.32**	−.20	−.17	−.16	-
12	EVT	.35**	.30*	−.03	.50**	.22	.30*	.31**	−.30*	−.17	−.17	.84**	-
13	SB-5 ABIQ deviation scores	.33**	.35**	.08	.55**	.07	.51**	.37**	−.29*	−.25*	−.13	.70**	.64**

^*p<.05,

^**p<.01;

NEPSY-II: Neuropsychological Assessment of Children, 2nd Edition; BRIEF-2: Behavior Rating Inventory of Executive Function, second edition; EVT: Expressive Vocabulary Test; PPVT: Peabody Picture Vocabulary Test; SB5-ABIQ: Stanford-Binet Intelligence Scales, Fifth Edition, Abbreviated Battery Intelligence Quotient;

^aInhibition measure;

^bShifting measure.

Seven series of regression analyses were conducted to determine the independent contribution of each EF subdomain to verbal fluency after controlling the effect of vocabulary and cognitive ability. Due to the multicollinearity between receptive and expressive vocabulary (r = .84), we included only expressive vocabulary in all models. Cognitive ability was also included given its significant correlations with verbal fluency and the EF subdomains. Parent- reports of shifting and working memory were not included, given their non-significant correlations with verbal fluency (see Table 2). The assumption of no multicollinearity was met (VIF < 5), and the plot of standardized predicted values against standardized residuals indicated that the assumptions of linearity and homoscedasticity had been met in all models (Field, 2013). In addition to the standardized beta values, 95% bootstrap confidence intervals of unstandardized values (B values) were also reported given nonnormality distributions of a few variables including expressive vocabulary and nonverbal shifting (skewness and kurtosis were approximately between −2 and +2).

The results of regression analyses (Table 3) show that, in all seven models, the three predictors (expressive vocabulary, cognitive ability, and one of the EF measures) significantly explained the variance of verbal fluency performance (p < .05). All neuropsychological assessments of inhibition, shifting, working memory, and parent-report of inhibition had significant independent contributions to verbal fluency performance (95% bootstrap confidence intervals exclude zero). In other words, verbal fluency performance was significantly associated with both verbal and visuospatial measures of inhibition, shifting, and working memory, and parent report of inhibition after controlling the effects of expressive vocabulary and cognitive ability. As Table 3 shows, expressive vocabulary, and cognitive ability had no independent contributions to verbal fluency after taking into account the effects of EF measures (95% bootstrap confidence intervals include zero).

Table 3.

Results of regression analyses predicting NEPSY-II Verbal Fluency

Model	Predictors	β	p	B	95% Bootstrap CI		R 2	F	p
					Lower	Upper
Verbal assessment
1	EVT	.27	.05	.19	−.02	.39	.37	10.91	.00
	SB-5 ABIQ	.06	.66	.03	−.15	.21
	Cat/dog Stroopa	.43	.00	12.48	5.48	19.19
2	EVT	.28	.05	.23	−.06	.45	.38	9.64	.00
	SB-5 ABIQ	.08	.58	.04	−.12	.21
	Rule Shiftb	.52	.00	27.28	12.47	38.49
3	EVT	.11	.47	.07	−.08	.25	.21	5.55	.02
	SB-5 ABIQ	.04	.80	.02	−.16	.20
	SB-5 Verbal WM	.37	.01	.70	.17	1.17
Visuospatial assessment
4	EVT	.15	.33	.10	−.06	.32	.15	3.30	.03
	SB-5 ABIQ	.12	.42	.06	−.12	.24
	NEPSY-II Statuea	.24	.05	.50	.04	.92
5	EVT	.24	.09	.16	−.07	.35	.26	7.38	.00
	SB5-ABIQ	.03	.84	.01	−.20	.17
	Weiglb	.41	.00	2.16	.67	3.56
6	EVT	.14	.30	.09	−.07	.27	.29	8.78	.00
	SB-5 ABIQ	.05	.71	.03	−.14	.19
	SB-5 Nonverbal WM	.45	.00	.77	.40	1.13
Parent report
7	EVT	.19	.21	.12	−.05	.33	.20	5.40	.00
	SB-5 ABIQ	.13	.38	.07	−.12	.24
	BRIEF-2 Inhibit	−.26	.02	−.20	−0.42	−0.08

EVT: Expressive Vocabulary Test; SB5: Stanford-Binet Intelligence Scales, Fifth Edition; ABIQ: Abbreviated Battery Intelligence Quotient; WM: Working Memory;

^aInhibition measure;

^bShifting measure.

4. Discussion

In the current study, we examined the associations between a semantic verbal fluency task with verbal, visuospatial, and parent-report of inhibition, shifting, and working memory in a sample of youth with DS. Vocabulary and cognitive abilities were included in the analyses as control variables. Generally, the results were consistent with those reported in other study populations conducted among various clinical and non-clinical samples (Aita et al., 2018; Amunts et al., 2020; Amunts et al., 2021; Filippi et al., 2022; Gligorović & Buha, 2014; Gonzalez-Burgos et al., 2019; Gustavson et al., 2019; Henry et al., 2015; Kavé & Sapir-Yogev, 2020; Shao et al., 2014) and provide novel empirical evidence of the association between these two cognitive domains in youth with DS.

The findings specifically point to the crucial influences of inhibitory control (i.e., inhibition) on verbal fluency performance in youth with DS. Verbal, visuospatial, and parent reports measures within the subdomain of inhibition were all shown to be significantly correlated with verbal fluency performance and even had significant independent contributions to verbal fluency performance after controlling for the effects of vocabulary and cognitive ability. Youth with higher abilities in inhibition may show better performance in verbal fluency than those with lower inhibitory control. The crucial role of inhibitory control on verbal fluency has been specifically emphasized by previous studies conducted among typically developing children and adults, and among children with and without language difficulties (Amunts et al., 2020; Henry et al., 2015). Individuals with DS who are better able to control impulses in generating erroneous responses are also better able to stay on task and generate new words in a verbal fluency task. This skill likely translates to the classroom where children with more difficulties with inhibitory control likely have more difficulty generating longer sentences or novel ideas and has implications for how these children are supported. Treatment options to support inhibitory control might be considered. An alternative interpretation is also possible given the cross-sectional nature of the current study, and the results may indicate that individuals with high verbal fluency also simultaneously have high inhibition skills.

The results also indicate the relevance of neuropsychological assessments of shifting to performance on verbal fluency in youth with DS. Both verbal and visuospatial shifting were significantly correlated with verbal fluency performance and had significant independent contributions to verbal fluency performance after controlling the effects of vocabulary, and cognitive ability. The results are in line with a previous study that reported a significant positive correlation between shifting and verbal fluency in a sample of children with unspecified mild intellectual disabilities (Gligorović & Buha, 2014). Clustering and switching are two components of verbal fluency performance that may rely on higher-order cognitive control of EF including cognitive flexibility or shifting (Troyer, 2000; Troyer et al., 1997). Clustering refers to the production of words within semantic subcategories, and switching is defined as the ability to shift between clusters. Performance on verbal fluency tasks requires producing words within an identified category (clustering) and, once a subcategory is exhausted, quickly move to another subcategory or cluster (switching). Such a process requires the ability of cognitive flexibility involved in EF shifting tasks (Troyer, 2000) and justifies the association between verbal fluency and shifting in youth with DS in the current study. Because data were collected at one time point, a casual conclusion cannot be made, and it is possible that there may be a bidirectional relation between shifting and verbal fluency.

Additionally, working memory abilities support verbal fluency performance. Both verbal and visuospatial working memory tasks had significant correlations with the verbal fluency performance and had significant independent contributions to verbal fluency even after taking into account the effects of vocabulary and cognitive ability. This association corroborates previous work in adults with DS which found a significant association between verbal fluency and verbal short-term memory, a component process of working memory (Stavroussi et al., 2016). To perform verbal fluency tasks, participants need to hold the instructions and the earlier response in working memory (Filippi et al., 2022). Holding responses in working memory is necessary to prevent perseverations or repeated responses (Azuma, 2004) which is critical for better performance on the verbal fluency task. These findings have implications for academic instruction as well as conversational skills. Interventions targeting working memory could demonstrate downstream impacts on a child being able to use a greater variety of words in conversation or academic work, as reflected by performance on verbal fluency tasks. As was expressed with other EF subdomains, verbal fluency and working memory were measured simultaneously, the results may also be interpreted as a bidirectional correlation without any causal direction between the two cognitive constructs.

Contrary to our hypotheses, we did not find any significant correlations between verbal fluency performance and parent-report measures of shifting and working memory. Some explanations may justify these nonsignificant relations. From an operationalization viewpoint, direct and rating measures of EF are different in terms of how they are administered and scored. Neuropsychological assessments of EF involve highly standardized procedures that are administered by an examiner and usually assess accuracy, response time, and/or speeded responding under a time constraint. Rating measures of EF involve an informant reporting on difficulties with carrying out everyday tasks (Toplak et al., 2013). It is assumed that direct and rating measures assess different aspects of EF which may differently contribute to other cognitive or behavioral functions (Ten Eycke & Dewey, 2016; Toplak et al., 2013). Empirical literature also indicates that the two types of measure are minimally correlated (Toplak et al., 2013) specifically in children with DS (Daunhauer et al., 2017). Thus, given its performance-based nature, the verbal fluency task may be more associated with the direct performance-based measures of EF rather than the EF parent-report measures.

The results of the correlation analyses characterize the role of vocabulary abilities in verbal fluency performance in youth with DS. Both receptive and expressive vocabulary had significant positive correlations with verbal fluency performance. The results are generally in line with those reported by previous studies conducted among children with language difficulties (Henry & Crawford, 2004), and mixed clinical disorders (Whiteside et al., 2016), and provide novel empirical evidence of the link between vocabulary ability and verbal fluency performance in youth with DS. The findings indicate that high variability in vocabulary skills specifically observed among individuals with DS (Martin et al., 2009) may give rise to the variability in verbal fluency performance and support the underlying importance of language abilities in generation tasks like verbal fluency (Delgado-Álvarez et al., 2021; Henry & Crawford, 2004; Isacoff & Stromswold, 2014; Whiteside et al., 2016). However, the results of regression analyses revealed that, after taking into account the effects of EF subdomains and cognitive ability, expressive vocabulary does not reflect a significant independent contribution to verbal fluency. The findings do not underscore the impact of language on verbal fluency, but they do suggest that inhibition, shifting, and working memory are critical components associated with verbal fluency in youth with DS independent of vocabulary. EF subdomains may have mediatory roles in the relation between vocabulary and verbal fluency in youth with DS, although more empirical evidence, as well as longitudinal data, are required to support this assumption.

4.1. Clinical Implications

Uncovering the associations among EF, language, and verbal fluency in youth with DS may benefit intervention and clinical practice to support outcomes for these children. Verbal fluency performance in individuals with DS has garnered considerable attention, given its relevance to the early onset of Alzheimer’s disease in this special population (Chen & Ringenbach, 2019; del Hoyo et al., 2015). Specific interventions such as acute exercise have been suggested to improve verbal fluency specifically in individuals with DS (Chen & Ringenbach, 2019). Understanding the association between verbal fluency and EF provides insight for how the intervention may be impacting overall EF and detecting that change in the verbal fluency outcome measure. Thus, verbal fluency may serve as downstream outcome measure for interventions that support EF in children with DS. For instance, home-based rehearsal training to improve verbal working memory in children with DS (Conners et al., 2008), and computerized cognitive training aimed to improve EF inhibitory control, shifting, and working memory in children and adults with DS (Bennett et al., 2013; McGlinchey et al., 2019) may cumulatively demonstrate their effectiveness in a verbal fluency outcome measure. Further, regarding the correlations observed between vocabulary and verbal fluency, youth with DS may also benefit from language intervention to improve verbal fluency. Language training co-delivered by parents and clinicians within a naturalistic setting may have the potential to provide positive outcomes for younger children with DS (O’Toole et al., 2018; Seager et al., 2021). Children and youth with DS will continue to benefit from professional speech-language therapy services (McDaniel & Yoder, 2016) or specific evidence-based speech and language interventions such as Broad Target Speech Recasts (Yoder et al., 2016), and palatal plate therapy (Carlstedt et al., 2003).

4.2. Limitations and Recommendations

While the present study provided evidence regarding the associations between the three EF subdomains and verbal fluency performance in youth with DS, some limitations need to be acknowledged. Only a semantic verbal fluency measure was used and the phonemic verbal fluency scores was not included, given its unreliability for use in participants with DS (Smeyne et al., 2022). Further, perseverations (i.e., the number of repeated responses), and intrusions (i.e., the number of incorrect responses) were excluded from the analyses, given their observed unreliability, and non-normal distributions (Schworer, Belizaire, et al., 2021). Finally, given the correlational nature of the current study, and the relatively small sample size (n = 69), any causal conclusion and generalizability should be made with caution. It is recommended that further studies replicate the analyses using a larger sample size to confirm the results and increase their generalizability. Inclusion of comparison groups of typically developing participants, or children with other neurodevelopmental conditions would also increase the clinical application of the study. Given various language difficulties experienced by youth with DS, future studies are needed to focus more on language ability as a prominent factor associated with verbal fluency performance and included more specific language features such as phonological awareness, rapid automated naming, and lexical access speed (Marques et al., 2022). In addition, processing speed, episodic memory, and delayed retrieval, as more general cognitive abilities associated with verbal fluency in the non-DS population (Delgado-Álvarez et al., 2021; Kavé & Sapir-Yogev, 2020; Stolwyk et al., 2015), are recommended to be included when studying the cognitive correlates of verbal fluency in youth with DS.

4.3. Conclusion

The current study empirically supports the association between semantic verbal fluency performance and three EF subdomains including inhibition, shifting, and working memory in a sample of youth with DS. Specifically, the study revealed that associations between verbal fluency and all neuropsychological assessments and parent-report of inhibition remained significant after controlling the effects of vocabulary and cognitive ability. However, the associations between verbal fluency performance and parent-report measures of shifting, and working memory were not supported. The study emphasizes the underlying importance of EF in generation tasks like verbal fluency performance in youth with DS.

What this paper adds?

This study adds to the emerging literature on the link between executive functioning and verbal fluency performance among individuals with Down syndrome and supports the underlying importance of executive functioning processes in word generation tasks like verbal fluency. Specifically, it provides novel empirical evidence of the association between semantic verbal fluency performance and three areas of executive functioning (inhibition, shifting, and working memory) in youth with Down syndrome who experience challenges in both executive functioning and verbal fluency performance. Further, the relations between both receptive and expressive vocabulary skills and verbal fluency performance were supported. The results also highlight the more prominent role of all three areas of executive functioning over expressive vocabulary ability in predicting verbal fluency in youth with Down syndrome.