Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Mar 1.
Published in final edited form as: Brain Cogn. 2021 Jan 24;148:105694. doi: 10.1016/j.bandc.2021.105694

Word Retrieval Difficulty in Adult Females with the FMR1 Premutation: Changes Over Time and Across Contexts

Shelley L Bredin-Oja a, Steven F Warren a,b, Rebecca E Swinburne Romine a, Kandace K Fleming a, Nancy Brady a,b, Elizbeth Berry-Kravis c
PMCID: PMC7928209  NIHMSID: NIHMS1662833  PMID: 33503544

Abstract

Individuals with a premutation of the fragile X mental retardation (FMR1) gene are at risk for a variety of psychological, physical, and cognitive issues, including difficulty with word retrieval. The present study examined three indicators of word retrieval difficulty; reduced productivity, reduced lexical diversity, and increased errors in word retrieval in a group of 38 female premutation carriers during standard-length speech samples collected over a period of eight years. Our results revealed that as women aged, they produced fewer words, produced fewer different words, and had greater word retrieval errors. In addition, the rate of word retrieval errors was highly correlated between two speaking contexts, indicating that this difficulty was pervasive and not solely the result of speaking in monologue. Our results suggest that subtle areas of cognitive decline emerge at a much earlier age among female premutation carriers than would be expected during healthy aging.

Keywords: fragile X premutation carriers, language dysfluencies; word retrieval difficulty

1. Introduction

In the general population of the US a premutation of the fragile X mental retardation gene (FMR1 gene) is relatively common with estimates of over one million carriers and is more prevalent among women than men with estimates ranging from 1 in 113–207 women and 1 in 194–530 men (Maenner et al., 2013). The FMR1 gene is responsible for producing the fragile X mental retardation protein (FMRP), which is essential for normal cognitive development. The premutation is defined as the presence of a CGG trinucleotide repeat expansion of 55–200 repeats on the long arm of the X chromosome (at position Xq23.11) in the 5’ untranslated region of FMR1 gene. The normal gene has between 5 and 44 CGG repeats. Alleles with 45 and 54 repeats are considered in the “gray zone”, and more than 200 CGG repeats constitute a full mutation of the FMR1 gene.

A consequence of the premutation allele is an increase in FMR1 mRNA levels and a decrease in FMRP levels most notably at higher repeat sizes over >120 (Garcia-Arocena & Hagerman, 2010). Two known conditions are associated with the premutation, fragile X-associated primary ovarian insufficiency [FXPOI; (A. K. Sullivan et al., 2005)] which occurs in approximately 16% to 24% of carriers (S. D. Sullivan, Welt, & Sherman, 2011), and fragile X-associated tremor/ataxia syndrome [FXTAS; (R. Hagerman & Hagerman, 2013)] which affects approximately 45.5% of male premutation carriers and 16.5% of female carriers over the age of 50 (Rodriguez-Revenga et al., 2009). However, there is a growing body of evidence that premutation carriers may suffer from a variety of psychological, physical, and cognitive issues that are not encompassed by these two conditions. (Hagerman et al., 2018; Wheeler et al., 2014). Indeed, the full significance of being a fragile X premutation carrier is unclear and there is a need for further research regarding conditions associated with the premutation (Johnson et al., 2020).

For example, researchers have found an increased difficulty with word retrieval, as measured by verbal dysfluency, in premutation carrier men with FXTAS compared to normative data (Grigsby et al., 2006). Increased dysfluencies were also noted in premutation carrier women without a diagnosis of FXTAS who were compared to a group of women matched for age and stress level (Klusek et al., 2018; Sterling, Mailick, Greenberg, Warren, & Brady, 2013). These latter findings suggest that there may be subtle specific areas of deficit in cognitive functioning in a substantial portion of premutation carriers, even without a diagnosis of FXTAS. That is, word retrieval difficulty may be an early indicator of fragile X-associated conditions.

In addition to finding group differences between women with and without the premutation, Sterling et al. (Sterling et al., 2013) found a significant positive correlation between the number of verbal dysfluencies during connected speech and age for the premutation carrier group but this relationship did not hold for the comparison group of mothers of children with autism spectrum disorder. That is, older women with the premutation had a greater number of dysfluencies than younger female premutation carriers in this cross-sectional study.

Difficulty with word retrieval, of which verbal dysfluency is a part, is prevalent among adults with aphasia but older adults who are cognitively healthy commonly complain of this difficulty as well (Condret-Santi et al., 2013). Presumably, difficulty with word retrieval arises because the connection in the lexicon between the semantic representation of a word and its phonological representation deteriorates with increasing age. This weak connection causes a representation to not be fully activated and production of the word fails (Burke & Shafto, 2004). An important question is, at what age do these word finding problems begin to manifest?

Kavé and Goral (2017) conducted a narrative review of studies comparing word retrieval during connected speech in healthy aging to word retrieval in adults with aphasia. They reasoned that if difficulty with word retrieval was evident in healthy aging there would be (1) reduced overall productivity in connected speech (i.e., fewer spoken words); (2) more limited lexical variety (i.e., fewer different words), based on the assumption that adults who have difficulty with word retrieval are likely to use words that are more easily accessed and to use these same words repeatedly, and (3) that they would have an increase in retrieval errors (i.e., greater dysfluency), just as they found in studies of adults with aphasia that they reviewed. Unfortunately, none of the studies they reviewed were longitudinal; rather, they were cross-sectional across a range of ages.

In contrast to their assumption, Kavé and Goral (2017) found that many studies reported an increase in number of produced words with age and, therefore, concluded that there was no reliable evidence that difficulty with productivity was common in connected speech by older cognitively healthy adults. Furthermore, studies they reviewed did not support the assumption that older cognitively healthy adults produced fewer different words (i.e., had restricted lexical variety) during connected speech. That is, reduced productivity and limited lexical variety during connected speech are not features of healthy aging.

Regarding retrieval errors, specifically, dysfluencies such as revisions, repeated words, and filled pauses (e.g., uh, um), the results from Kavé and Goral’s review (2017) as well as other studies are mixed. Two studies found an effect for age (i.e., (Bortfeld, Leon, Bloom, Schober, & Brennan, 2001; Manning & Monte, 1981). Other studies found an effect for age only under some conditions, such as describing pictures with negative or unpleasant content compared to neutral pictures (i.e., (Castro & James, 2014), or when producing infrequent words compared to more common words (i.e., (Dennis & Hess, 2016). Schmitter-Edgecombe and colleagues (Schmitter-Edgecombe, Vesneski, & Jones, 2000) found an age effect for only one type of dysfluency—revisions. In contrast, two other studies found no effect for age across participants in their 20s to 80s (i.e., (Cooper, 1990; Duchin & Mysak, 1987).

When an effect for age was found in the above noted studies, it was in age groups that were 55 years or greater. Only three studies included a middle age group that more closely resembles the participants of the current study, that is, women in their mid-forties. Bortfeld and colleagues (2001) found that while their oldest group (M = 67;2 years) produced a higher rate of dysfluencies at 6.65 per 100 words, the middle age group (M = 47;11 years) did not differ from the youngest group (M = 28;10 years) with rates of 5.69 and 5.55 per 100 words, respectively. Duchin and Mysak (1987) found no difference in rates of dysfluencies among their five age groups of young adults (M = 25 years), middle age adults I (M = 49 years), middle age adults II (M = 60 years), elder adults I (M = 68 years), and elder adults II (M = 80 years). Similarly, Cooper (1990) found no effect for age when they compared six age groups ranging in age from their 20s through 70s and concluded that aspects of expressive language, including dysfluencies do not change as a function of age among healthy adults. Cooper (1990) asserts that if language differences are found, some process other than normal aging (i.e., disease) may be present.

Cross-sectional studies of different age groups can suggest a relationship between age and various outcomes, but a more definitive approach is a longitudinal study that follows the same cohort of participants over an extended period of years. This allows each subject to be compared to their own baseline instead of some group statistic. In this study, we present data from a cohort of premutation carrier mothers who have a child with fragile X syndrome (FXS). Data were collected over a span of eight years. Following the findings of Sterling and colleagues (2013) that older premutation carrier mothers may be at greater risk for difficulty with word retrieval as measured by verbal dysfluency, we asked if fragile X premutation carrier mothers showed impaired word retrieval abilities over time.

Evidence of word retrieval difficulty at an earlier age than would be expected in healthy adults may lead to earlier detection of a dysexecutive syndrome, a syndrome marked by impairments in executive functioning evident in neurologic disease, that can progress to early dementia (Berry-Kravis & Hall, 2011). Furthermore, clearer delineation of the phenotype of female premutation carriers is key to possible treatment options for this relatively large population.

We used the same set of assumptions that Kavé and Goral (2017) employed regarding difficulty with word retrieval. Namely, we determined if the total number of spoken fluent words (i.e., productivity) and the total number of different words (i.e., lexical variety) declined over time, and if word retrieval errors (i.e., dysfluencies and abandoned utterances) increased over time. In addition, we examined several biomarkers (i.e., CGG repeat length, activation ratio, and FMRP levels) to determine whether there was a relationship between these biomarkers and difficulty with word retrieval.

Both the study by Sterling and colleagues (2013) and the study be Klusek and colleagues (2018) described above, included a comparison group of mothers of children with autism spectrum disorder (ASD) to control for the stress of raising a difficult child and found higher rates of dysfluencies in the premutation carrier groups, indicating that stress was not a contributing factor. However, a study by Mahl (1987) found that higher rates of dysfluency were correlated with higher rates of anxiety in women (r = .59, p < .01) but not men (r = −.47, p < .16).

Because our speech samples focused on a personal reflection of the mothers’ relationship with their child, we compared two measures of participants’ mental health to the rates of word retrieval difficulty in our sample of 38 premutation mothers. Specifically, we hypothesized that anxiety may be related to difficulty with word retrieval. We also tested whether the ASD status and challenging behaviors of their child with FXS were related to word retrieval errors. Previous research has found a relationship between negative or unpleasant content and increased difficulty with word retrieval (Castro & James, 2014); however, an examination of the content of the mothers’ speech samples regarding the relationship with their child did not reveal any negativity toward their child.

Finally, we examined the proportion of word retrieval errors (i.e., dysfluencies) in two different speaking contexts—a five-minute monologue and a dyadic semi-structured interview—to determine if word retrieval difficulty was dependent on context. We hypothesized that speaking extemporaneously in monologue for five minutes without feedback from the listener may be a more challenging language production task than engaging in a back-and-forth conversation with an active listener and, therefore, more word retrieval errors may be evident during the five-minute speech sample.

1.1. Research Questions

Our specific research questions were:

  1. Does the total number of fluent words produced in a standard speech sample (i.e., overall productivity) change over time?

  2. Does the number of different words produced in a standard speech sample (i.e., lexical variety) change over time?

  3. Does the proportion of utterances produced in a standard speech sample that contains a word retrieval error (i.e., dysfluencies) change over time?

  4. Does context (i.e., monologue versus dialogue) affect the amount of word retrieval errors?

  5. Are specific biomarkers related to difficulty with word retrieval?

  6. Is anxiety related to difficulty with word retrieval?

  7. Are child characteristics that may cause anxiety, such as ASD status, related to difficulty with word retrieval?

2. Method

2.1. Participants

Thirty-eight premutation carrier women participated in this longitudinal study over the span of eight years. All participants gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Human Research Protection Program at the University of Kansas. All women had at least one child with FXS and were part of a larger longitudinal study (Warren, Brady, Sterling, Fleming, & Marquis, 2010). The mean ages of the participants at the first observation was 38.6 (SD= 4.5) and at the last observation was 46.8 (SD= 4.3) years. The participants’ level of education ranged from high school graduate to multiple post graduate degrees. A categorical variable for participants’ education was created, grouping participants into (1) less than a bachelor’s degree (N=11), (2) a bachelor’s degree (N=15) and (3) more than a bachelor’s degree (N=12). All participants were native English speakers, 34 of 38 (90%) were non-Latina White, one participant was Latina White, one participant was Black, one participant was Pacific Islander, and one participant was biracial Black and White. See Table 1 for further descriptors.

Table 1.

Participant Characteristics

Measure Range Mean (SD)
Mothers’ age at last data point 36.7 – 55.03 47 (4.26)
CGG repeat length 75 – 127 97.29 (13.32)
Activation Ratio .10 – .90 .47 (.224)
POMS Tension-Anxiety Subscale 1 – 20 7.30 (5.4)
CES-D 0 – 29 7.32 (7.4)
Open in a new tab

Note: POMS = Profile of Mood States; CES-D = Center for Epidemiologic Studies Depression Scale

2.2. Word Retrieval Measures

2.2.1. Contexts

Uninterrupted five-minute speech samples [FMSS; (Magaña et al., 1986)] were collected longitudinally at three or four points in time over a period of eight years and analyzed for number of non-dysfluent words, number of different words, and number of word retrieval errors. Participants were instructed to speak about their child with FXS for five minutes without interruption. Specifically, they were given the following instructions:

I’d like to hear your thoughts and feelings about (child’s name) in your own words. For the next five minutes I’d like you to tell me what kind of person (child’s name) is and how the two of you get along together. I can’t answer any questions or interrupt you until the five minutes are up. Do you have any questions before you begin?

A semi-structured interview regarding the participant’s perceptions, experiences, and parenting strategies of their child with FXS was also collected once, at the final time point, and analyzed for word retrieval errors (i.e., dysfluencies) to address research question 4.

2.2.2. Transcript Analyses

The samples were audio-recorded digitally and then transcribed and coded for word retrieval errors (i.e., dysfluencies and abandoned utterances) using the Systematic Analysis for Language Transcripts [SALT: (Miller & Iglesias, 2012)]. Utterances were separated by communication units (c-units), SALT coding conventions for bound morphemes were applied, and all instances of language dysfluencies and abandoned utterances were coded according to the SALT manual. Abandoned utterances occurred when the participant voluntarily stopped mid-utterance and did not finish the utterance.

2.2.3. Word retrieval errors

Word retrieval errors (i.e., dysfluencies) comprised three types—repetitions, revisions, and filled pauses. Repetitions and revisions were coded at the part word, word, and phrase level. Part word repetitions and revisions had to contain at least one full syllable. Repetitions included an immediate and verbatim repeat of a word or phrase. Revisions included words or phrases that differed from what had previously been produced. Filled pauses are lexical items that did not carry meaning (e.g., um, uh, hm) or did not contribute to the meaning of the utterance (e.g. “you know”). The proportion of c-units that either contained a dysfluency or were abandoned served as the variable of interest for research questions 3 and 4. We chose to examine rates of dysfluency at the level of C-units rather than at the rate per 100 words so that abandoned utterances, a more conspicuous type of revision, could be included in the overall word retrieval error variable. Definitions and examples of word retrieval errors are provided in Table 2.

Table 2.

Definitions and Examples of Word Retrieval Errors

Type of Retrieval Error Definition Example
Abandoned utterance The speaker voluntarily stops speaking without being interrupted. I think it’s just>
Filled Pause Words or vocalizations that fill a pause. He’s (um) usually pretty happy. So I will (you know) read it to him.
Repetition The exact duplication of a word or phrase. (He’s) he’s always helpful. (We don’t) We don’t really do much on weekends.
Revision False starts or reformulations such that what follows is distinctly different. (We) he loves to listen to music. (He’s always) he and his brother spend time together.
Proportion of total word retrieval errors Number of Independent clauses plus any modifiers (c-units) that contain a dysfluency or are abandoned over the total number of c-units.
Open in a new tab

Analysis of the FMSS and semi-structured interviews in this manner is the same as language sample analysis and is considered to be the gold standard for assessing spoken language production (Miller et al., 2016). Furthermore, unlike many language measures, language sample analysis can be used frequently (even daily) to measure language production in discourse (Miller et al., 2016). Therefore, the scores derived from the FMSS and semi-structured interviews can be considered reliable repeated measures of word retrieval difficulty.

Audio files collected at the final time point were transcribed by either the first author or a research assistant who was trained in SALT coding. Every file was checked by a second transcriber. Differences in coding were resolved by consensus. Reliability for language transcripts from the earlier time points are described by Sterling and colleagues (2013).

2.3. Biomarkers

Thirty-four of the participants supplied a blood sample that was analyzed for CGG repeat length, activation ratio, and level of FMRP at the Rush University Medical Center Genetics Laboratory. Polymerase chain reaction (PCR) analysis of the CGG repeat region in the fragile X gene was conducted using primers flanking and within the repeat sequence according to the manufacturer’s instructions (Asuragen, Inc). Whole cellular DNA was extracted from blood, digested with restriction endonucleases Eco R1 and Eag 1, resolved by electrophoresis through agarose and blotted onto nylon membranes. Southern blots were hybridized to radiolabeled DNA probe StB12.3 and subjected to autoradiography, and band intensities measured by densitometry for activation ratios. For FMRP determinations, lymphocytes were isolated from cell separation tubes and FMRP quantified as described by LaFauci et al. (2013) using Luminex technology.

Length of CGG repeat expansion served as a continuous variable. In addition, a categorical variable of low (55–89), mid-size (90–110) and high (111–199) repeats was also tested, with 12, 15, and 6 participants in each category, respectively. Previous studies have found a curvilinear CGG repeat length relationship with language dysfluencies and other aspects of the premutation carrier phenotype (e.g., age of menopause, major depression disorder) with the mid-size range showing the highest clinical involvement (Ennis, Ward, & Murray, 2006; Klusek et al., 2018; Mailick, Hong, Greenberg, Smith, & Sherman, 2014; Roberts et al., 2016).

2.4. Measures of Child Behavior

ASD symptomology was measured at the final time point using two assessments. The Childhood Autism Rating Scale [CARS2-ST; (Schopler, Van Bourgondien, Wellman, & Love, 2010)] and the Autism Diagnostic Observation Schedule-2 [ADOS-2; (Lord et al., 2012)], which was administered by an examiner who was trained to research reliability. A total score of 28 to 33.5 on the CARS2-ST classifies an individual as having mild to moderate symptoms of autism; scores of 34 or greater indicate severe symptoms. The total score of the ADOS-2 is used to classify an individual as non-spectrum, autism spectrum, or autism. Consensus between the two measures (i.e., 28 or higher on the CARS2-ST and classification of autism spectrum or autism on the ADOS-2) served as a dichotomous categorical variable of ASD or non-ASD. Thirteen children (34%) were classified as having comorbid ASD.

2.5. Maternal Mental Health

Two self-report measures of mental health were collected at the final time point; the Center for Epidemiologic Studies Depression Scale [CES-D; (Radloff, 1977)], which is designed to measure depressive symptomatology in the general population and the tension/anxiety subscale of the Profile of Mood States [POMS;(McNair, Lorr, & Droppleman, 1992)], which measures transient, fluctuating affective mood states.

2.6. Data Analysis

The three primary outcomes of interest comprised: (1) word retrieval errors (i.e., dysfluencies); (2) total number of words spoken (i.e., overall productivity); and (3) number of different words used (i.e., lexical variety) during a standard-length speech sample collected at three or four time-points over eight years. In addition, word retrieval errors (i.e., dysfluencies) were measured during a semi-structured interview at the final data collection point.

Each outcome of interest from the FMSS was examined separately using a series of two-level models in which observations (level 1) were nested within individuals (level 2). Maximum Likelihood estimation was used for all models. In the initial model, maternal age, centered at age 40, was added to each model at level one. Next, each additional variable of interest was added as a level 2 predictor, as well as its interaction term with maternal age. The only exception to this was for the biological variables of activation ratio and length of CGG repeats. The CGG repeat length interacts with the activation ratio; that is, CGG repeat length will only impact symptoms in cells in which it is expressed; therefore, an interaction term was created. The interaction between CGG repeat length and the activation ratio is important; therefore, each variable and the interaction term were added to the maternal age model together, rather than one at a time as was done for the other models. Finally, to address question 4, a correlational analysis comparing the rate of word retrieval errors (i.e., dysfluencies) between the two speaking contexts from the last data collection point was conducted.

3. Results

Growth curve models were conducted for the three measures of word retrieval difficulty with age centered at 40 years. The education categorical variable and the interaction term for education and maternal age were added to each model as fixed effects. Neither variable was significant (all p values were greater than .20); therefore, they were dropped from all models.

3.1. Word Retrieval Difficulty

3.1.1. Verbal productivity

There was a significant effect of age on productivity (F (1,32.5) = 19.06, p < .0001) based on a two-level random intercept model with observations nested in participants. Specifically, there was a significant negative trend for productivity over time, with a slope of −7.02, such that a typical participant would be expected to produce approximately 7 fewer fluent words per year. A typical mother in this sample is expected to have a total of 675.23 words at age 35, and 604.04 words at age 45. See Figure 1.

Figure 1.

Figure 1.

Open in a new tab

Total number of words across maternal age.

3.1.2. Lexical variety

There was a significant effect of age on lexical diversity (F (1,34.6) = 15.89, p= .0003) based on a two-level random intercept model with observations nested in participants. Specifically, there was a significant negative trend for lexical diversity over time, with a slope of −1.76, such that a typical participant would be expected to use approximately 2 fewer different words per year. A typical mother in this sample is expected to use a total of 241.34 different words at age 35, and a total of 223.75 different words at age 45. See Figure 2.

Figure 2.

Figure 2.

Open in a new tab

Number of different words across maternal age.

3.1.3. Word retrieval errors

There was a significant effect of age on the proportion of c-units that contained a dysfluency or were abandoned (F (1,114) = 5.47, p = .021), based on a two-level random intercept model with observations nested in participants. Specifically, there was a significant positive trend over time, with a slope of 0.433, such that a typical participant in this sample is expected to have a word retrieval error rate per c-unit of 50.61% at age 35, and 55% at age 45. See Figure 3.

Figure 3.

Figure 3.

Open in a new tab

Word retrieval errors across maternal age.

3.2. Word Retrieval Errors by Context

Semi-structured interviews at the last data point were transcribed and coded in the same manner as the FMSSs that were collected over a period of eight years. A correlational analysis of the proportion of c-units that either contained a dysfluency or were abandoned revealed that the two contexts were highly correlated (r = .74, p < .001) for the rate of word retrieval errors. See Figure 4.

Figure 4.

Figure 4.

Open in a new tab

Word retrieval errors across speaking contexts.

3.3. Biomarkers

Activation ratio and CGG repeat length were available for 34 of the 38 participants. Both variables and the interaction term were added to each of the three growth curve models. None were significant; p levels ranged from .3 to .93. CGG repeat length ranged from 75 to 127, with a mean of 97.29 and a standard deviation of 13.3. Approximately half were in the mid-size category while smaller numbers were in the high and low categories. Activation ratios ranged from .10 to .90, with a mean of .47 and a standard deviation of. 22.

3.4. Maternal Mental Health

Measures of depression and anxiety were added to each of the three growth curve models as a fixed effect, as well as the interaction term of each with maternal age. None were significant; p levels ranged from .54 to .84.

3.5. ASD status

ASD status was added to each of the three models as a fixed effect, as well as the interaction term of ASD status and maternal age. Neither were significant; p levels ranged from .08 to .8.

4. Discussion

This study examined word retrieval difficulties in fragile X premutation carrier women over a span of eight years. An effect for age was found on all three aspects of word retrieval. As women aged, they produced fewer words, produced fewer different words, and had greater word retrieval errors during a standard five-minute speech sample that was administered three or four times in eight years. Reduced productivity and limited lexical variety are not typical in healthy aging (Kavé & Goral, 2017). Therefore, even a downward trend of seven fewer words per year and two fewer different words per year, which may be barely perceptible to others, could be a clinical indicator of cognitive decline. Findings for word retrieval errors in healthy aging is mixed. Some studies (e.g., Bortfeld et al., 2001; Manning & Monte, 1981) found word retrieval errors increased as healthy adults age, but this occurred for participants in their 60s, a much later age than the participants of this study.

The proportion of utterances that contained a dysfluency or were abandoned during the five-minute extemporaneous speech sample was highly correlated with this same measure during a semi-structured interview, a dialogue between the participant and a researcher. This finding indicates that these word retrieval errors were pervasive and not solely the result of speaking in monologue.

The women in this study were, on average, 46 years old at the last data collection. Evidence of a decline in word retrieval relative to their own previous performance decades before such a decline can be expected in healthy adults is alarming. To our knowledge, this study is the first longitudinal study to document declines within participants rather than across age groups. It is, of course, unknown if this decline will continue as these women enter their 50s and 60s or if it will stabilize. The effect of aging on word retrieval difficulties at a much earlier age than would be expected in healthy adults adds to the growing body of evidence of age-related cognitive decline in female premutation carriers not only in language abilities but other areas of cognition such as visuospatial cognition, numerical enumeration tasks, and arithmetic (Naomi J Goodrich-Hunsaker et al., 2011; Naomi Jean Goodrich-Hunsaker et al., 2011; Lachiewicz, Dawson, Spiridigliozzi, & McConkie‐Rosell, 2006; Sterling et al., 2013).

Similar to findings by Sterling and colleagues (2013) but in contrast with Klusek and colleagues (2018), we did not find significant associations between CGG repeat length and dysfluencies. In addition, we did not find an association between difficulty with word retrieval and measures of anxiety or child challenging behaviors that are likely to induce anxiety such as ASD symptomology. The lack of association between anxiety and word retrieval errors complements the results of Sterling and colleagues (2013) who found an effect for language dysfluencies in fragile X premutation carrier mothers of children with FXS, but not in a comparison group of mothers of children with autism spectrum disorders who served as a control for the documented effects of stress and anxiety of having a child with a severe disability (Smith, Seltzer, & Greenberg, 2012).

4.1. Strengths and Limitations

An obvious strength of this study is the longitudinal nature of the data. This study has provided clear evidence of increasing difficulty with three aspects of word retrieval, an indicator of cognitive decline, within premutation carrier mothers that more strongly shows an effect for age than cross-sectional studies across age groups. The comparison of two different speaking contexts provides strong evidence that the word retrieval errors exhibited by the participants were not due to any presumed pressure of speaking extemporaneously for a full five minutes.

This study also has limitations. Word retrieval errors in the semi-structured interview were only examined at one timepoint. In addition, we did not include other measures of executive functioning which, presumably, would have shown a relationship with our measures of word retrieval difficulty. Lastly, we were unable to associate the decline in word retrieval with the putative combined biomarkers of CGG length repeat and activation ratio. With biomarkers from only 34 of the participants, it may be that our study did not have adequate power to detect a relationship, or this relationship may not exist in our sample given the distribution of CGG repeat lengths observed.

4.2. Conclusions

This study has important clinical implications for the health and well-being of a relatively large portion of the population who are premutation carrier women. Subtle areas of cognitive decline emerging at such an early age and when the responsibilities of caring for a child with FXS are still extant may have serious consequences for the overall well-being of the mother and, consequently, for the family as a whole. Further studies, and in particular, studies that follow the same families over an extended period of time, substantially longer than the eight years of observations in the current study, are essential to develop a clearly defined phenotype of the female fragile X premutation carrier and for the data needed to explore effects of targeted treatments for this population.

Highlights.

We examined word retrieval difficulty in female premutation carriers over eight years

As women aged, they produced fewer words, fewer different words, and had more errors

Word retrieval difficulty occurred much earlier than can be expected in healthy aging

Subtle cognitive decline may be a factor of premutation carrier status for women

Acknowledgments

We wish to express our profound gratitude to the women and their families who have been a part of our larger longitudinal study, from which this data came, for the past 15 years.

Funding

This research was supported by NICHD grants P30-HD003110, P30-HD002528, R01-HD084563 and U54-HD090216.

Footnotes

Data Statement

The data analyzed in this study are not publicly available. Participants were assured data would be presented in aggregate form only and language transcripts and other raw data would not be shared.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Berry-Kravis E, & Hall DA (2011). Executive dysfunction in young FMR1 premutation carriers: forme fruste of FXTAS or new phenotype?, 77, 612–613. [DOI] [PubMed] [Google Scholar]
  2. Bortfeld H, Leon SD, Bloom JE, Schober MF, & Brennan SE (2001). Disfluency rates in conversation: Effects of age, relationship, topic, role, and gender. Language and Speech, 44(2), 123–147. [DOI] [PubMed] [Google Scholar]
  3. Burke DM, & Shafto MA (2004). Aging and language production. Current Directions in Psychological Science, 13(1), 21–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Castro N, & James LE (2014). Differences between young and older adults’ spoken language production in descriptions of negative versus neutral pictures. Aging, Neuropsychology, and Cognition, 21(2), 222–238. [DOI] [PubMed] [Google Scholar]
  5. Condret-Santi V, Barbeau E, Matharan F, Le Goff M, Dartigues J-F, & Amieva H (2013). Prevalence of word retrieval complaint and prediction of dementia in a population-based study of elderly subjects. Dementia and Geriatric Cognitive Disorders, 35(5–6), 313–324. [DOI] [PubMed] [Google Scholar]
  6. Cooper PV (1990). Discourse production and normal aging: Performance on oral picture description tasks. Journal of Gerontology, 45(5), P210–P214. [DOI] [PubMed] [Google Scholar]
  7. Dennis PA, & Hess TM (2016). Aging-related gains and losses associated with word production in connected speech. Aging, Neuropsychology, and Cognition, 23(6), 638–650. [DOI] [PubMed] [Google Scholar]
  8. Duchin SW, & Mysak ED (1987). Disfluency and rate characteristics of young adult, middle-aged, and older males. Journal of Communication Disorders, 20(3), 245–257. [DOI] [PubMed] [Google Scholar]
  9. Ennis S, Ward D, & Murray A (2006). Nonlinear association between CGG repeat number and age of menopause in FMR1 premutation carriers. European Journal of Human Genetics, 14(2), 253–255. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Garcia-Arocena D, & Hagerman PJ (2010). Advances in understanding the molecular basis of FXTAS. Human Molecular Genetics, 19(R1), R83–R89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Goodrich-Hunsaker NJ, Wong LM, McLennan Y, Srivastava S, Tassone F, Harvey D, et al. (2011). Young adult female fragile X premutation carriers show age-and genetically-modulated cognitive impairments. Brain and cognition, 75(3), 255–260. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Goodrich-Hunsaker NJ, Wong LM, McLennan Y, Tassone F, Harvey D, Rivera SM, et al. (2011). Adult female fragile X premutation carriers exhibit age-and CGG repeat length-related impairments on an attentionally based enumeration task. Frontiers in Human Neuroscience, 5, 1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Grigsby J, Brega AG, Jacquemont S, Loesch DZ, Leehey MA, Goodrich GK, et al. (2006). Impairment in the cognitive functioning of men with fragile X-associated tremor/ataxia syndrome (FXTAS). Journal of the Neurological Sciences, 248(1–2), 227–233. [DOI] [PubMed] [Google Scholar]
  14. Hagerman R, & Hagerman P (2013). Advances in clinical and molecular understanding of the FMR1 premutation and fragile X-associated tremor/ataxia syndrome. The Lancet Neurology, 12(8), 786–798. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hagerman RJ, Protic DD, Rajaratnam A, Salcedo-Arellano MJ, Aydin EY, & Schneider A (2018). Fragile X-associated neuropsychiatric disorders (FXAND). Frontiers in Psychiatry, 9, 564. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Johnson K, Herring J, & Richstein J (2020). Fragile x premutation associated conditions (FXPAC). Frontiers in Pediatrics, 8, 266. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kavé G, & Goral M (2017). Do age-related word retrieval difficulties appear (or disappear) in connected speech? Aging, Neuropsychology, and Cognition, 24(5), 508–527. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Klusek J, Porter A, Abbeduto L, Adayev T, Tassone F, Mailick M, et al. (2018). Curvilinear Association Between Language Disfluency and FMR1 CGG Repeat Size Across the Normal, Intermediate, and Premutation Range. Frontiers in Genetics, 9, 344. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lachiewicz AM, Dawson DV, Spiridigliozzi GA, & McConkie‐Rosell A (2006). Arithmetic difficulties in females with the fragile X premutation. American Journal of Medical Genetics Part A, 140(7), 665–672. [DOI] [PubMed] [Google Scholar]
  20. LaFauci G, Adayev T, Kascsak R, Kascsak R, Nolin S, Mehta P, et al. (2013). Fragile X screening by quantification of FMRP in dried blood spots by a Luminex immunoassay. The Journal of Molecular Diagnostics, 15(4), 508–517. [DOI] [PubMed] [Google Scholar]
  21. Lord C, Rutter M, DiLavore P, Risi S, Gotham K, & Bishop S (2012). Autism diagnostic observation schedule–2nd edition (ADOS-2). Los Angeles, CA: Western Psychological Corporation. [Google Scholar]
  22. Maenner MJ, Baker MW, Broman KW, Tian J, Barnes JK, Atkins A, et al. (2013). FMR1 CGG expansions: prevalence and sex ratios. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 162(5), 466–473. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Magaña AB, Goldstein MJ, Karno M, Miklowitz DJ, Jenkins J, & Falloon IR (1986). A brief method for assessing expressed emotion in relatives of psychiatric patients. Psychiatry research, 17(3), 203–212. [DOI] [PubMed] [Google Scholar]
  24. Mahl GF (1987). Explorations in nonverbal and vocal behavior. Hillsdale, NJ: Erlbaum. [Google Scholar]
  25. Mailick MR, Hong J, Greenberg J, Smith L, & Sherman S (2014). Curvilinear association of CGG repeats and age at menopause in women with FMR1 premutation expansions. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 165(8), 705–711. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Manning WH, & Monte KL (1981). Fluency breaks in older speakers: Implications for a model of stuttering throughout the life cycle. Journal of Fluency Disorders, 6(1), 35–48. [Google Scholar]
  27. McNair D, Lorr M, & Droppleman L (1992). Revised manual for the Profile of Mood States. San Diego, CA: Educational and Industrial Testing Services, 731, 732–733. [Google Scholar]
  28. Miller J, & Iglesias A (2012). Systematic analysis of language transcripts (SALT), Research Version 2012. Middleton, WI: SALT Software LLC. [Google Scholar]
  29. Miller J, Andriacchi K, & Nockerts A (2016). Using language sample analysis to asses spoken language production in adolescents. Language, Speech, and Hearing Services in Schools, 47(2), 99–112. [DOI] [PubMed] [Google Scholar]
  30. Radloff LS (1977). The CES-D scale: A self report depression scale for resarch in the general population. Applied Psychological Measurements, 1(3), 385–401. [Google Scholar]
  31. Roberts JE, Tonnsen BL, McCary LM, Ford AL, Golden RN, & Bailey DB Jr (2016). Trajectory and predictors of depression and anxiety disorders in mothers with the FMR1 premutation. Biological psychiatry, 79(10), 850–857. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Rodriguez-Revenga L, Madrigal I, Pagonabarraga J, Xuncla M, Badenas C, Kulisevsky J, et al. (2009). Penetrance of FMR1 premutation associated pathologies in fragile X syndrome families. European Journal of Human Genetics, 17(10), 1359. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Schmitter-Edgecombe M, Vesneski M, & Jones D (2000). Aging and word-finding: A comparison of spontaneous and constrained naming tests. Archives of Clinical Neuropsychology, 15(6), 479–493. [PubMed] [Google Scholar]
  34. Schopler E, Van Bourgondien ME, Wellman GJ, & Love SR (2010). CARS-2: Childhood Autism Rating Scale–Second Edition. Los Angeles, CA: Western Psychological Services. [Google Scholar]
  35. Smith L, Seltzer MM, & Greenberg JS (2012). Daily health symptoms of mothers of adolescents and adults with fragile X syndrome and mothers of adolescents and adults with autism spectrum disorder. Journal of Autism and Developmental Disorders, 42, 1836–1846. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Sterling AM, Mailick M, Greenberg J, Warren SF, & Brady N (2013). Language dysfluencies in females with the FMR1 premutation. Brain and cognition, 82(1), 84–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Sullivan AK, Marcus M, Epstein MP, allen EG, Anido AE, Paquin JJ, et al. (2005). Association of FMR1 repeat size with ovarian dysfunction. Human Reproduction, 20(2), 402–412. [DOI] [PubMed] [Google Scholar]
  38. Sullivan SD, Welt C, & Sherman S (2011). FMR1 and the continuum of primary ovarian insufficiency. Paper presented at the Seminars in reproductive medicine. [DOI] [PubMed] [Google Scholar]
  39. Warren SF, Brady N, Sterling A, Fleming K, & Marquis J (2010). Maternal responsivity predicts language development in young children with fragile X syndrome. American journal on intellectual and developmental disabilities, 115(1), 54–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Wheeler AC, Bailey DB Jr, Berry-Kravis E, Greenberg J, Losh M, Mailick M, et al. (2014). Associated features in females with an FMR1 premutation. Journal of Neurodevelopmental Disorders, 6(1), 30. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES