Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Network analysis of narrative discourse and attention-deficit hyperactivity symptoms in adults

  • Rafael Martins Coelho ,

    Contributed equally to this work with: Rafael Martins Coelho, Cláudia Drummond, Natália Bezerra Mota, Felipe Kenji Sudo, Rosemary Tannock, Paulo Mattos

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing

    rafaelmartinsmd@gmail.com

    Affiliation Institute D’Or for Research and Education (IDOR), Rio de Janeiro, RJ, Brazil

  • Cláudia Drummond ,

    Contributed equally to this work with: Rafael Martins Coelho, Cláudia Drummond, Natália Bezerra Mota, Felipe Kenji Sudo, Rosemary Tannock, Paulo Mattos

    Roles Conceptualization, Data curation, Funding acquisition, Methodology, Supervision, Writing – original draft, Writing – review & editing

    Affiliations Institute D’Or for Research and Education (IDOR), Rio de Janeiro, RJ, Brazil, Department of speech and hearing pathology, Federal University of Rio de Janeiro (UFRJ), Rio de Janeiro, RJ, Brazil

  • Natália Bezerra Mota ,

    Contributed equally to this work with: Rafael Martins Coelho, Cláudia Drummond, Natália Bezerra Mota, Felipe Kenji Sudo, Rosemary Tannock, Paulo Mattos

    Roles Formal analysis, Software, Supervision, Writing – original draft, Writing – review & editing

    Affiliation Physics Department, Federal University of Pernambuco (UFPE), Recife, PE, Brazil

  • Pilar Erthal ,

    Roles Data curation, Formal analysis, Investigation, Project administration

    ‡ These authors also contributed equally to this work.

    Affiliation Institute D’Or for Research and Education (IDOR), Rio de Janeiro, RJ, Brazil

  • Gabriel Bernardes ,

    Roles Data curation, Investigation

    ‡ These authors also contributed equally to this work.

    Affiliation Institute D’Or for Research and Education (IDOR), Rio de Janeiro, RJ, Brazil

  • Gabriel Lima ,

    Roles Data curation, Investigation, Project administration

    ‡ These authors also contributed equally to this work.

    Affiliation Institute D’Or for Research and Education (IDOR), Rio de Janeiro, RJ, Brazil

  • Raquel Molina ,

    Roles Data curation, Investigation, Project administration

    ‡ These authors also contributed equally to this work.

    Affiliation Institute D’Or for Research and Education (IDOR), Rio de Janeiro, RJ, Brazil

  • Felipe Kenji Sudo ,

    Contributed equally to this work with: Rafael Martins Coelho, Cláudia Drummond, Natália Bezerra Mota, Felipe Kenji Sudo, Rosemary Tannock, Paulo Mattos

    Roles Formal analysis, Methodology, Supervision, Writing – original draft, Writing – review & editing

    Affiliation Institute D’Or for Research and Education (IDOR), Rio de Janeiro, RJ, Brazil

  • Rosemary Tannock ,

    Contributed equally to this work with: Rafael Martins Coelho, Cláudia Drummond, Natália Bezerra Mota, Felipe Kenji Sudo, Rosemary Tannock, Paulo Mattos

    Roles Supervision, Writing – original draft, Writing – review & editing

    Affiliation Research Institute, The Hospital for Sick Children, Toronto, Canada

  • Paulo Mattos

    Contributed equally to this work with: Rafael Martins Coelho, Cláudia Drummond, Natália Bezerra Mota, Felipe Kenji Sudo, Rosemary Tannock, Paulo Mattos

    Roles Conceptualization, Data curation, Funding acquisition, Methodology, Project administration, Supervision, Writing – original draft, Writing – review & editing

    Affiliations Institute D’Or for Research and Education (IDOR), Rio de Janeiro, RJ, Brazil, Institute of Psychiatry, Federal University of Rio de Janeiro (UFRJ), Rio de Janeiro, RJ, Brazil

Network analysis of narrative discourse and attention-deficit hyperactivity symptoms in adults

  • Rafael Martins Coelho, 
  • Cláudia Drummond, 
  • Natália Bezerra Mota, 
  • Pilar Erthal, 
  • Gabriel Bernardes, 
  • Gabriel Lima, 
  • Raquel Molina, 
  • Felipe Kenji Sudo, 
  • Rosemary Tannock, 
  • Paulo Mattos
PLOS
x

Abstract

Previous research investigating language in attention-deficit hyperactivity disorder (ADHD) has demonstrated several deficits in many aspects. However, no previous study employed quantitative methodology providing objective measures that could be compared among different studies with diverse samples. To fill this gap, we used network analysis to investigate how ADHD symptomatology impacts narrative discourse, a complex linguistic task considered to be an ecological measure of language. Fifty-eight adults (34 females and 24 males) with a mean age of 26 years old and a mean of 17 years of educational level were administered the Adult Self-Rating Scale for ADHD symptomatology. They also completed the State-Trait Anxiety Inventory, the Beck Depression Inventory and the Urgency, Premeditation, Perseverance, Sensation Seeking Behavior Scale. Intelligence quotient was calculated. Individuals were asked to tell a story based on a wordless picture book. Speech was recorded and transcribed as an input to SpeechGraphs software. Parameters were total number of words (TNW), number of loops of one node (L1), repeated edges (RE), largest strongly connected component (LSC) and average shortest path (ASP). Verbosity was controlled. Statistical analysis was corrected for multiples comparisons and partial correlations were performed for confounding variables. After controlling for anxiety, depression, IQ, and impulsiveness ADHD symptomatology was positively correlated with L1 and negatively correlated with LSC. TNW was positively correlated with ADHD symptoms. In a subdomain analysis, both inattention and hyperactivity-impulsivity were negatively correlated with LSC. Only hyperactivity-impulsivity positively correlated with TNW and L1. Results indicated a correlation between ADHD symptoms and lower connectedness in narrative discourse (as indicated by higher L1 and lower LSC), as well as higher total number of words (TNW). Our results suggest that the higher the number of ADHD symptoms, the less connectivity among words, and a higher number of words in narrative discourse.

Introduction

Attention-Deficit Hyperactivity Disorder is defined by the core symptoms of inattention, hyperactivity and impulsivity [1], but there is large body of evidence for deficits beyond those diagnostic features [2]. Adults with the disorder may present clinically significant impairments in tasks demanding working memory, memory span, processing speed, decision making, delay aversion to rewards, time perception, executive function, and general communication abilities [3]. As a result, an array of occupational, academic, social and domestic difficulties may arise from different sets of those deficiencies [4].

Among the phenotypic dimensions of ADHD, language problems may encompass a large set of discourse comprehension and production difficulties. A myriad of ADHD-related deficits has already been documented: poor comprehension of main plotline, ambiguous reference, event sequencing errors, incomplete clauses, discourse interruptions and embellishment. In addition—or possibly as a consequence of the aforementioned difficulties—their discourse often displays pragmatic problems like excessive speech, poor turn-taking and fail in maintaining appropriate topics in conversations. Finally, deficits were demonstrated in comprehension and elaboration of the main plot, monitorization and autocorrection of speech, online monitorization and organization of discourse and story resolution; such deficits could be a linguistic expression of executive deficits often seen in ADHD [512].

Narrative discourse (ND) refers to the ability of verbally reporting real or imaginary events by translating them into comprehensive structured sequences of logically-linked ideas [13,14]. It is a complex linguistic skill, which requires integration of primary language components (phonological, lexical, semantic, morphosyntax and pragmatic) with several other cognitive functions (memory, attention, planning, mental model generation and inferential production) [1517]. Because ND assessment usually involves the reproduction of a story, it requires that the individual establishes temporal, spatial, and causal relationships among events. ND is an ecological measure of language used in daily conversations, requiring the ability to plan and organize thoughts into an expected structure easily comprehensible by the interlocutor [1823]. Tasks addressing ND require the speaker to verbally recount an episode experienced in the present (for example, the perception of visual stimuli portrayed in scenes) or past (memory recall of events) while respecting the temporal, causal and spatial relationships among events that unfold in particular scenarios [8,24].

Previous studies investigating language in children with ADHD, reported a wide range of language problems such as: 1) less overall recall of story units, more production of ambiguous references, semantically inappropriate word substitutions, more inaccurate information and more sequence errors [6,7]; 2) misinterpretations [7]; 3) fewer verbal production, fewer utterances, fewer autocorrections then controls, and use of more words then controls to correct an error; fewer utterances related to the main plot and fewer utterances of resolution, resulting in incomplete narratives, as well with less emphasis in the main story plot then controls [8]; 4) less sustained use of the goal plan throughout their narratives, higher rates of coherence errors then control, and production of narratives resembling those produced by younger children which has less consistent use of the story goal plan [9]; 5) more difficulty producing a grammatical and fluent utterances when speakers have less syntactic flexibility, suggesting more problems with syntactic planning then controls; those deficits were seen even in adult participants who had recovered symptomatically from ADHD [11]; 6) more difficulty then controls to detect sequence errors in narrative production instructions [12]; 7) difficulties with discourse management, presupposition and narrative discourse production [8]; 8) more difficulties with global coherence in story production [8]; 9) higher rates of embellishment errors in narrative production [8] and 10) abrupt unannounced changes in topic conversations, inappropriate responses, inappropriate use of intonation, inappropriate use of pronouns and infrequent use of information by ellipsis, unintelligible rate of speech (cluttering), inappropriate loud and amount of speech, and overlapping speech [8]. All those characteristics will have consequences in discourse organization, story resolution and expected coherence in narrative production [59,11,12,25,26]. While functional consequences associated with linguistic abnormalities in ADHD have been increasingly recognized, instruments for assessing ND in ADHD are scarce and mostly qualitative (and therefore more prone to examiner’s biases) in children and scarce in adults [11,27,28].

The two main methods to assess language production are story retelling and conversational samples. Both methods may employ either qualitative or quantitative analyses and have given heterogeneous and sometimes discordant results according to the literature. Most studies investigating ADHD and language included samples with children and adolescents. Although there are very few previous studies investigating adult ADHD, they have either used instruments with uncertain ecological validity or qualitative analyses, which are prone to interpretation biases. The main contribution of this study is that we have used a novel quantitative approach (Speech Graph Analyses, previously used in a few psychiatric disorders) to investigate Narrative discourse, a complex linguistic skill considered to be an ecological measure [2932].

Network analysis of discourse (also called graph analysis of speech) has been proposed as an useful method to investigate ND, providing quantitative indices of many elements, such as long and short-range recurrences of nodes (i.e., words from the story); the former being a proxy of connectedness [2933]. Such analysis has proven to provide insights beyond language. For example, in schizophrenia graph analysis of speech demonstrated that long-range recurrence was inversely correlated with negative symptoms and impaired performance on several cognitive tests.

The present study aimed to investigate the structure of ND in adults considering their ADHD symptomatology using network analysis. For this, individuals were asked to tell a story based on a wordless picture book. We hypothesized based on previous findings that network analysis would reveal differences in ND attributes in individuals with ADHD symptoms [24]. Specifically, we predicted that higher ADHD symptomatology would be associated to a more poorly connected report (with fewer long-range and more short-range recurrences). To our knowledge, this is the first study addressing ND using this methodology.

Materials and methods

Sample

The study was conducted in Brazil, with Brazilian young adults subjects. All participants volunteered and signed an informed consent approved by the Ethics Committee of Copa D’Or Hospital (Submission Certification for Ethics Appreciations (CAAE) number: 38000614.1.0000.5249). Participants (n = 71) were primarily volunteers recruited among university students (graduate and undergraduate); a small number (10%) was referred by their mental health professionals aware of the study. Inclusion criteria were a) age between 18 and 40 years old, and b) current or past diagnosis of ADHD. Exclusion criteria were: a) IQ lower than 80 (WAIS-III–Wechsler Adult Intelligence Scale) [34,35]; b) current or past diagnosis of language/communication disorders, bipolar disorder or psychotic disorders; c) presence of severe anxious-depressive state (1); d) altered consciousness due to substance abuse (1); e) presence of severe sensory deficits or severe motor difficulties that precluded neuropsychological assessment; f) history of stroke, traumatic brain injury (TBI) or any known brain lesions; g) current diagnosis of uncontrolled epilepsy or delirium; h) not being a native Brazilian Portuguese speaker (1); and i) not completing the entire protocol (10). Of 71 interviewed subjects, 58 met the eligibility criteria.

Neuropsychological measure of IQ was used as screening measure for eligible criteria attendance. It was obtained using the Vocabulary and Blocks subtests of Wechsler Adult Intelligence Scale (WAIS-III) [36,37].

Individuals with known ADHD (n = 6), who were already under pharmacological treatment, were asked to undergo a 48-hour washout of psychostimulants before the assessment.

Our final sample comprised 34 females and 24 males with a mean age of 26 years old and a mean educational level of 17 years of scholarship. Their mean IQ was 119 (Table 1). Table 2 details clinical characteristics of sample.

Symptomatology measures

ASRS.

Participants were given the Adult Self Rating Scale (ASRS) in Portuguese [38] to access the current 18 ADHD symptoms pertaining to two symptom-domains: inattention and hyperactive-impulsivity; each item is scored from 0 to 3. Because there is no solid normative data for ASRS in Brazil, scores 2 (“often”) and 3 (“very often”) were considered to be positive. ADHD symptoms were used as predictors variables, and higher scores on ASRS reflect greater ADHD symptoms. ASRS has high internal consistency, with a Cronbach’s alpha coefficient of 0.84 [34].

State-Trait Anxiety Inventory (STAI).

The Brazilian-Portuguese version of State-Trait Anxiety Inventory is composed of two 20-item scales; with each question scored from 1 to 4; measures situational anxiety (State Subscale) and anxious trait (Trait Subscale). Shows high internal consistency estimated as Cronbach’s alpha coefficient of 0.89 [35].

Beck Depression Inventory (BDI).

The Brazilian-Portuguese version of Beck Depression Inventory composed of 21 questions; each item being scored on a scale value of 0 to 3; final score is a measure of depression symptom severity. BDI also shows high internal consistency estimated as Cronbach’s alpha coefficient of 0.89 [38,39].

UPPS.

The Brazilian-Portuguese version of Urgency, Premeditation, Perseverance, Sensation Seeking Behavior Scale composed of 45 items that address four personality factors related to impulsive behavior; each item is scored on a scale from 1 to 4. UPPS provides a total score of impulsivity, and besides that, also provides subscale scores of each impulsivity subtype: lack of premeditation, urgency, sensation seeking and lack of perseverance; each one with the respective Cronbach’s alpha coefficient of 0.87 (lack of premeditation), 0.85 (urgency), 0.84 (sensation seeking), and 0.76 (lack of perseverance) [40,41].

Narrative discourse assessment and measures

Narrative discourse task.

The book “Frog, where are you?” [42] previously employed for assessment of healthy subjects [4345], ADHD children [8,9,46,47]; and older adults with ADHD [24], was used in this study. It is a wordless book composed by twenty-four sequential black and white drawings telling the story of a boy who loses his frog and engages in a journey to recover it. The story unfolds in twenty-four frames evenly distributed throughout the book pages, showing events comprising a main plot (the search for the pet frog and events directly related to the aim of getting it back) and a secondary plot (peripheral series of events not necessary for comprehending the story).

The book was presented to the individual with the following instructions: “Here is a picture book. It is about a boy and his pets, a frog, and a dog. You should look at each page, and then tell me the story. I will record the story you produce. You can look through the whole book as many times as you want before we start. It is not necessary to memorize the book; you will keep it with you while you tell me the story. You should try telling the best possible story, pretending that I don’t know it”.

Narrative discourse evaluation and network analyses.

Network analysis of discourse is the attempt to apply small-world network theory in assessment and objective description of linguistic properties. This method analyzes the networks (graphs) of transcripted discourse (language sample, in this paper, narrative discourse). As words relation in discourse is a complex system that could be represented as a network (graph), those network structures provide intuitive and useful representations for language modeling knowledge and inference [29,31,48,49]. The networks are represented by graphs composed of nodes and edges, whereas each node is a word, and each edge represents the temporal sequence between those words (links between successive words). Those graphs have attributes (speech graph attributes or SGA) that permits useful insights about language characteristics.

The analysis automatization generated the SpeechGraphs software, created by the same group that proposed the analysis [50]. It automatically generates network graphs representative of a transcripted language sample in a txt file. It also provides/calculates attributes from each graph. Those attributes concerns to general characteristics (total of Nodes (N) and Edges (E), and total of words–Word Count (WC)); recurrence characteristics (Repetitive Edges (RE), Parallel Edges (PE), and Loops of one, two and three nodes (L1, L2, and L3); connectivity characteristics (number of nodes on the Largest Connected Component (LCC), number of nodes on the Largest Strongly connected Component (LSC), and Average Total Degree (ATD)); four global measures (Density (D) or the amount of edges in a given graph divided by the potential number of edges given the number of nodes, Diameter (DI) or the shortest distance measured by the number of edges between the pair of nodes with the highest distance in the graph, Average Shortest Path (ASP) or the average shortest distance measured by the average number of edges between all pair of nodes in the graph, and the clustering coefficient (CC) or, considering a node, CC of this node is the estimation of how the direct linked nodes are also linked to each other, and CC of a graph is the average number of CC considering all nodes in a graph [32].

Narratives were represented as word graphs using the SpeechGraphs software [32] (Fig 1). The results of SGA analysis are networks, in which each node corresponds to a word and the temporal link between two words are represented by an edge, a reliable method for speech structure investigation by non-semantic graph analysis of discourse transcripts. Preprocessing procedures text normalization like backbone speech elements examination (corresponding to subject, verb and object) and conversion to canonical elements (lexemes) [29] was not performed, since our data derived from ND elicited from a picture wordless book, which usually generates limited variations in word choice. Stop-words, comprising liaison terms with no specific meaning in oral speech (i.e.: “a”, “an”, “and”, “of”, numerals, etc.), were automatically removed from the original text. Identification of stop-words was based on the Portuguese language repository of the Stop-Words Project from Google Code Archives [51].

thumbnail
Fig 1. Word graphs generated by Speech Graph Analysis (SGA) software.

N = node; E = edge; LSC = largest strongly connected component; L1 = number of loops of one node; RE = repeated edges; ASP = general measure of graph size.

https://doi.org/10.1371/journal.pone.0245113.g001

In this present paper we chose to apply a widely used method for network analysis, where each node represents a word, different from our previous paper where each node corresponded to a different nucleus of the narrative plot. In order to control SGAs for verbosity, which has a strong influence on them according to previous findings [30,32,33,52], we used moving windows (length = 30 words, 50% overlap) (Fig 2).

thumbnail
Fig 2. Split text method for verbosity control.

Example of moving windows with 18 words length, and 50% of overlap; used to calculate mean values per graph for the different attributes.

https://doi.org/10.1371/journal.pone.0245113.g002

We investigated elements already shown to be relevant in ND using network analysis [32]; they are depicted in Fig 1. SGAs were used as predicted variables. Total Number of Words (TNW) refers to the total number of words used in the ND. The number of loops of one node (L1) and repeated edges (RE) are measures of short-range recurrences (repetition of words, or words pairs, e.g.: short repetitions). The largest strongly connected component (LSC) is a measure of long-range recurrences (repetition of sentences or phrases with many words). The average shortest path (ASP) is a general measure of graph size.

Statistical analysis

All the statistical analysis used the SPSS software. Normality and variance homogeneity were assessed by the Kolmogorov-Smirnov test. Spearman correlations were calculated for SG attributes (RE, L1, LSC and ASP) and ADHD symptoms assessed by ASRS scale. Partial correlation to control for confounding factors were performed with SPSS software. Statistical analysis was corrected for multiples comparisons (Bonferroni’s Correction; p = .0125).

Results

Table 3 shows the correlation between current ADHD symptoms and SGA parameters. ADHD current symptoms was positively correlated with loops of one node (L1) and negatively correlated with the largest strongly connected component (LSC); a visual depiction of these findings is seen on Fig 3. In a subdomain analysis of ADHD symptoms, both current inattention and hyperactivity-impulsivity symptoms positively correlated with L1, but only Hyperactivity-impulsivity symptoms had a negative correlation with LSC. Total number of words (TNW) was positively correlated with Hyperactivity-Impulsivity. We found no correlations with ADHD and repeated edges (RE) or graph size (ASP).

thumbnail
Fig 3. Examples of networks graphs.

Examples of individual graphs from subjects with low (top pair; no ADHD symptoms) and high (bottom pair; 18 ADHD symptoms) ADHD symptoms.

https://doi.org/10.1371/journal.pone.0245113.g003

When the correlation between ADHD and SGA parameters were controlled for IQ, educational level, UPPS, BDI and STAI, the total number of words (TNW), loops of one node (L1) and largest strongly connected component (LSC) persisted in positive correlation with ADHD symptoms (Table 4). L1 positively correlated with hyperactivity-impulsivity symptoms only. LSC had negative correlation with ADHD symptoms, as well as inattention and hyperactivity-impulsivity symptoms. Repeated edges (RE) and graph size (ASP) persisted without significative correlations.

thumbnail
Table 4. Correlations between TNW, RE, L1, LSC, ASP and ADHD symptoms controlled for Age, IQ, educational level, BDI, STAI and UPPS.

https://doi.org/10.1371/journal.pone.0245113.t004

Discussion

In this study, the first to address ADHD and narrative discourse in young adults using network analysis, we provided a different framework to investigate ADHD-related language problems previously demonstrated in different languages across the world. Our findings from this quantitative methodology provide strong support for previous research in the field, however using mainly qualitative methodologies which strongly rely on the examiners’ expertise and may vary depending on the tasks employed for language assessment. Our study may provide the basis for more studies using this strategy in ADHD.

The small sample size should be considered a main limitation to this study, as well the IQ and educational level of the sample. Further studies should investigate narrative discourse in larger samples of subjects. Besides that would try to correlate with neuropsychological variables (as working memory and executive function). Notwithstanding, we used a task addressing an ecological language measure—narrative discourse (ND). The use of network analysis of narrative discourse (speech graph analysis) allowed structural quantification of language characteristics of young adults with objective and unbiased measures. We chose parameters theoretically and functionally important for narrative discourse analysis, such as attributes of short-range and long-range recurrences from SGA graph networks [3033,52].

As we hypothesized, ADHD symptoms positively correlated with short-range recurrence (loops of one node, L1) and the opposite was found for long-range recurrence (largest strongly connected component, LSC), where a negative correlation with ADHD symptoms was found. ADHD symptomatology was associated to narrative discourses with less connectivity among words (lower LSC); our results are in accordance to a large study with children where poor language skills have been associated higher inattention or hyperactivity symptoms in primary school [52]. A genome-wide study on ADHD identified risk loci located in FOXP2 gene, involved in neural mechanisms mediating the development of speech and learning [53].

Those results suggest that the more ADHD symptoms the individual has, the more he presents short-range recurrences and fewer long-range recurrences. In typical children, less long-range recurrences (i.e., connectedness) correlated with lower IQ and lower verbal memory as well as worse performance on tasks addressing theory of mind [52,54]; increased connectedness has been associated with higher reading abilities and higher educational levels [29,52]. To date, there are no studies addressing the clinical correlates of lower connectedness in ADHD.

Because ADHD may be associated with comorbid anxious and depressive disorders [55] we controlled for those parameters; correlations persisted. In addition, because impulsivity, a hallmark of ADHD, could potentially contribute to a worse performance in narrative discourse, we have also controlled for this aspect, however obtaining similar results.

Total number of words was positively correlated with Hyperactivity-Impulsivity. Of note, this result is in accordance to DSM-5 criterion “talks excessively”. Our finding may add further understanding of previous descriptions of verbosity in adult ADHD discourse, however using a different methodology [8] ADHD associated deficits in oral narrative could be due to deficits in executive functioning (including working memory), commonly seen in ADHD. Working memory impacts the ability to organize the narrative production, maintaining the principal story plot. Executive function deficits could also lead to excessive digressions from main plot; attempts to correct or adjust an utterance already in course may lead to production of more speech, but without incremental quality or relevant information. Although our study design does not allow to draw conclusions on this, higher values for L1 (an attribute of short-range recurrence) and lower LSC (an attribute of long-range recurrence) could potentially be secondary to this behavior. Another way to interpret the higher number of short-range recurrences would be a tendency to produce short utterances (like false starts, hesitations, pauses), while the individual gains time to organize the next language structure of the topic, in order to overcome the working memory deficit. Again, our study design did not allow us to investigate those aspects; further studies will be necessary to clarify this association.

In summary, our study has demonstrated that ADHD symptoms are associated with narrative discourse problems, in particular the degree of connectedness and number of words used to convey the story. From a clinical perspective, our results suggest that whenever individuals with ADHD symptomatology are requested to narrate a story based on visual stimuli, they are more verbose, repeat more words or words pairs and produce a discourse with less cohesion among the words they have chosen to tell the plot.

Of note, in a study employing different tasks addressing ND in adults with mild dementia, it was shown that the same profile of impairment was seen across all modalities (with and without visual stimuli) [56]. There are no such studies on ADHD, but it seems reasonable to presume that the same language impairment would be present, at various levels, when ADHD patients are asked to verbally report current or past facts. One could expect ADHD individuals to portray a less cohesive and more verbose speech in clinical interviews.

Acknowledgments

We thank Instituto D’Or de Pesquisa e Ensino (IDOR) for project support.

References

  1. 1. APA. Diagnostic and Statistical Manual of Mental Disorders - 5th ed. 5th ed. Arlington, VA, US: American Psychiatric Association; 2013.
  2. 2. Sjöwall D, Roth L, Lindqvist S, Thorell LB. Multiple deficits in ADHD: Executive dysfunction, delay aversion, reaction time variability, and emotional deficits. J Child Psychol Psychiatry Allied Discip. 2013;54: 619–627. pmid:23061803
  3. 3. Gordon CT, Fabiano GA. The Transition of Youth with ADHD into the Workforce: Review and Future Directions. Clinical Child and Family Psychology Review. Springer US; 2019. pmid:30725305
  4. 4. Merrill BM, Molina BSG, Coxe S, Gnagy EM, Altszuler AR, Macphee FL, et al. Functional Outcomes of Young Adults with Childhood ADHD: A Latent Profile Analysis. J Clin Child Adolesc Psychol. 2020;49: 215–228. pmid:30689405
  5. 5. Beitchman JHH, Nair R, Clegg M, Ferguson B, Patel PGG. Prevalence of Psychiatric Disorders in Children with Speech and Language Disorders. J Am Acad Child Psychiatry. 1986;25: 528–535. pmid:3489024
  6. 6. Tannock R, Purvis KL, Schachar RJ. Narrative abilities in children with attention deficit hyperactivity disorder and normal peers. J Abnorm Child Psychol. 1993;21: 103–117. pmid:8463500
  7. 7. Purvis KL, Tannock R. Language abilities in children with attention deficit hyperactivity disorder, reading disabilities, and normal controls. J Abnorm Child Psychol. 1997;25: 133–44. pmid:9109030
  8. 8. Peets K, Tannock R, Kathleen Peetsa RT, Peets K, Tannock R, Kathleen Peetsa RT. Los errores y autocorrecciones en la narración distinguen el TDAH del TDAH con trastornos del lenguaje. Rev Logop Foniatr y Audiol. 2011;31: 228–236.
  9. 9. Renz K, Lorch EP, Milich R, Lemberger C, Bodner A, Welsh R. On-line story representation in boys with attention deficit hyperactivity disorder. J Abnorm Child Psychol. 2003;31: 93–104. pmid:12597702
  10. 10. Harris KM, Halpern CT, Haberstick BC, Smolen A. The National Longitudinal Study of Adolescent Health (Add Health) sibling pairs data. Twin Res Hum Genet. 2013;16: 391–8. pmid:23231780
  11. 11. Engelhardt PE, Veld SN, Nigg JT, Ferreira F. Are language production problems apparent in adults who no longer meet diagnostic criteria for attention-deficit/hyperactivity disorder? Cogn Neuropsychol. 2012;29: 275–299. pmid:23005917
  12. 12. McInnes A, Humphries T, Hogg-Johnson S, Tannock R. Listening comprehension and working memory are impaired in attention deficit hyperactivity disorder irrespective of language impairment. J Abnorm Child Psychol. 2003;31: 427. pmid:12831231
  13. 13. Van Dijk T a. Narrative macro-structures. A J Descr Poet Theory Lit. 1976;1: 547–568.
  14. 14. Menezes de Oliveira e Paiva VL. Narrative research: an introduction. Rev Bras Linguística Apl. 2008;8: 261–266.
  15. 15. Drummond C, Coutinho G, Fonseca RP, Assunção N, Teldeschi A, de Oliveira-Souza R, et al. Deficits in narrative discourse elicited by visual stimuli are already present in patients with mild cognitive impairment. Front Aging Neurosci. 2015;7: 1–11. pmid:25653617
  16. 16. Cannizzaro MS, Coelho CA. Analysis of Narrative Discourse Structure as an Ecologically Relevant Measure of Executive Function in Adults. J Psycholinguist Res. 2013;42: 527–549. pmid:23192423
  17. 17. Finestack LH, Fey ME, Catts HW. Pronominal reference skills of second and fourth grade children with language impairment. J Commun Disord. 2006;39: 232–248. pmid:16464464
  18. 18. Justice LM, Bowles RP, Kaderavek JN, Ukrainetz TA, Eisenberg SL, Gillam RB. The Index of Narrative Microstructure: A Clinical Tool for Analyzing School-Age Children’s Narrative Performances. Am J Speech-Language Pathol. 2006;15: 177. pmid:16782689
  19. 19. Hayes PA, Norris J, Flaitz JR. A Comparison of the Oral Narrative Abilities of Underachieving and High-Achieving Gifted Adolescents. Lang Speech Hear Serv Sch. 1998;29: 158–171. pmid:27764422
  20. 20. Liles BZ, Duffy RRJ, Merritt DD, Purcell SL. Measurement of narrative discourse ability in children with language disorders. J Speech Hear Res. 1995;38: 415–425. pmid:7596107
  21. 21. Scott CM, Windsor J. General language performance measures in spoken and written narrative and expository discourse of school-age children with language learning disabilities. J Speech Lang Hear Res. 2000;43: 324–39. pmid:10757687
  22. 22. Kaderavek JN, Sulzby E. Narrative Production by Children With and Without Specific Language Impaiment: Oral Narratives and Emergent Readings. J Speech, Lang Hear Res. 2000;43: 34–49.
  23. 23. Drummond C. Discurso narrativo e integridade de substância branca em indivíduos com Comprometimento Cognitivo Leve e Demência de Alzheimer Tese de Doutorado apresentada ao Programa de Pós-Graduação em Ciências Morfológicas, I. UFRJ. 2016.
  24. 24. Coelho R, Mattos P, Tannock R. Attention-Deficit Hyperactivity Disorder (ADHD) and narrative discourse in older adults. Dement Neuropsychol. 2018;12: 374–379. pmid:30546847
  25. 25. Staikova E, Gomes H, Tartter V, McCabe A, Halperin JM. Pragmatic deficits and social impairment in children with ADHD. J Child Psychol Psychiatry Allied Discip. 2013;54: 1275–1283. pmid:23682627
  26. 26. Lorch EP, Milich R, Flake RA, Ohlendorf J, Little S. A developmental examination of story recall and coherence among children with ADHD. J Abnorm Child Psychol. 2010;38: 291–301. pmid:20024672
  27. 27. Lorch EP, Diener MB, Sanchez RP, Milich R, Welsh R, van den Broek P. The effects of story structure on the recall of stories in children with attention deficit hyperactivity disorder. J Educ Psychol. 1999;91: 273–283. pmid:10503647
  28. 28. Fine J. Language in Psychiatry: A Handbook of Clinical Practice. Equinox Publishing Ltd. London, United Kingdom: Equinox Publishing Ltd; 2006.
  29. 29. Palaniyappan L, Mota NB, Oowise S, Balain V, Copelli M, Ribeiro S, et al. Speech structure links the neural and socio-behavioural correlates of psychotic disorders. Prog Neuro-Psychopharmacology Biol Psychiatry. 2019;88: 112–120. pmid:30017778
  30. 30. Mota NB, Vasconcelos NAP, Lemos N, Pieretti AC, Kinouchi O, Cecchi GA, et al. Speech graphs provide a quantitative measure of thought disorder in psychosis. PLoS One. 2012;7: 1–9. pmid:22506057
  31. 31. Bedi G, Carrillo F, Cecchi GA, Slezak DF, Sigman M, Mota NB, et al. Automated analysis of free speech predicts psychosis onset in high-risk youths. npj Schizophr. 2015;1: 15030. pmid:27336038
  32. 32. Mota NB, Furtado R, Maia PPCC, Copelli M, Ribeiro S. Graph analysis of dream reports is especially informative about psychosis. Sci Rep. 2015;4: 1–7. pmid:24424108
  33. 33. Mota NB, Sigman M, Cecchi G, Copelli M, Ribeiro S. The maturation of speech structure in psychosis is resistant to formal education. npj Schizophr. 2018;4. pmid:30531913
  34. 34. Mattos P, Segenreich D, Saboya E, Dias G, Louzã M, Romano M, et al. Transcultural adaptation of the Adult Self-Report Scale into portuguese for evaluation of adult attention-deficit/hyperactivity disorder (ADHD). Rev Psiquiatr Clin. 2006;33: 188–194.
  35. 35. Andrade L, Gorenstein C, Vieira Filho AH, Tung TC, Artes R. Psychometric properties of the Portuguese version of the State-Trait Anxiety Inventory applied to college students: factor analysis and relation to the Beck Depression Inventory. Brazilian J Med Biol Res. 2001;34: 367–374. pmid:11262588
  36. 36. Wagner F, Pawlowski J, Yates DB, Camey SA, Trentini CM. Viabilidade da estimativa de QI a partir dos subtestes Vocabulário e Cubos da WAIS-III. Psico-USF. 2010;15: 215–224.
  37. 37. Nascimento E. Escala de Inteligência de Wechsler para Adultos- Terceira Edição-WAIS-III- Adaptação e Padronização de uma amostra Brasileira. 1a ed. Casa do Psicólogo, editor. Casa do Psicólogo. São Paulo: Casa do Psicólogo; 2004.
  38. 38. Gomes-Oliveira MH, Gorenstein C, Neto FL, Andrade LH, Wang YP. Validation of the Brazilian Portuguese Version of the Beck Depression Inventory-II in a community sample. Rev Bras Psiquiatr. 2012;34: 389–394. pmid:23429809
  39. 39. Beck AT, Beamesderfer A. Assessment of depression: the depression inventory. Mod Probl Pharmacopsychiatry. 1974;7: 151–169. pmid:4412100
  40. 40. Garcia MS. Adaptação da escala UPPS-P e sua aplicabilidade na população brasileira. Universidade Federal de Minas Gerais. 2018.
  41. 41. Sediyama CYN, Moura R, Garcia MS, da Silva AG, Soraggi C, Neves FS, et al. Factor analysis of the Brazilian version of UPPS impulsive behavior scale. Front Psychol. 2017;8: 1–5. pmid:28197108
  42. 42. Mayer M. Frog, Where Are You? New York: Dial Books; 2003.
  43. 43. Berman Ruth & Dan Isaac Slobin. Relating events in narrative: a crosslinguistic developmental study. NJ: Erlbaum, 1994; 1994.
  44. 44. Berman RA. On the Ability to Relate Events in Narrative. Discourse Process. 1988;11: 469–497.
  45. 45. John SF, Lui M, Tannock R. Children’s Story Retelling and Comprehension Using a New Narrative Resource. Can J Sch Psychol. 2003;18: 91–113.
  46. 46. Van Lambalgen M, Van Kruistum C, Parigger E. Discourse processing in Attention-Deficit Hyperactivity Disorder (ADHD). J Logic, Lang Inf. 2008;17: 467–487.
  47. 47. Francis S, Fine J, Tannock R. Methylphenidate Selectively Improves Story Retelling in Children with Attention Deficit Hyperactivity Disorder. J Child Adolesc Psychopharmacol. 2001;11: 217–228. pmid:11642472
  48. 48. Steyvers M, Tenenbaum JB. The large-scale structure of semantic networks: Statistical analyses and a model of semantic growth. Cogn Sci. 2005;29: 41–78. pmid:21702767
  49. 49. Markošová M. Network model of human language. Phys A Stat Mech its Appl. 2008;387: 661–666.
  50. 50. Mota N. Análise de grafos aplicada a relatos de sonhos: ferramenta diagnóstica objetiva e diferencial para psicose esquizofrênica e bipolar. 2013; 123.
  51. 51. Google Code Archive. Stop-words—Google Code Archive. In: GNU GPL V3 [Internet]. 2014 [cited 10 Mar 2020]. 10.6084/m9.figshare.13176080.
  52. 52. Mota NB, Weissheimer J, Madruga B, Adamy N, Bunge SA, Copelli M, et al. A Naturalistic Assessment of the Organization of Children’s Memories Predicts Cognitive Functioning and Reading Ability. Mind, Brain, Educ. 2016;10: 184–195.
  53. 53. Demontis D, Walters RK, Martin J, Mattheisen M, Als TD, Agerbo E, et al. Discovery of the first genome-wide significant risk loci for attention deficit/hyperactivity disorder. Nat Genet. 2019;51: 63–75. pmid:30478444
  54. 54. Mota NB, Callipo R, Leite L, Torres AR, Weissheimer J, Bunge SA, et al. Verbal Short-Term Memory Underlies Typical Development of “Thought Organization” Measured as Speech Connectedness. Mind, Brain, Educ. 2020;14: 51–60.
  55. 55. Rösler M, Retz W, Thome J, Schneider M, Stieglitz RD, Falkai P. Psychopathological rating scales for diagnostic use in adults with attention-deficit/hyperactivity disorder (ADHD). Eur Arch Psychiatry Clin Neurosci. 2006;256: i3–i11. pmid:16977549
  56. 56. Bertola L, Mota NBNB, Copelli M, Rivero T, Diniz BS, Romano-Silva M a., et al. Graph analysis of verbal fluency test discriminate between patients with Alzheimer’s disease, mild cognitive impairment and normal elderly controls. Front Aging Neurosci. 2014;6: 185. pmid:25120480