Four Decades of ADHD: A Systematic AI-assisted Analysis of Conceptual Shifts Across Six DSM Editions

Four Decades of ADHD: A Systematic AI-assisted Analysis of Conceptual Shifts Across Six DSM Editions | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Four Decades of ADHD: A Systematic AI-assisted Analysis of Conceptual Shifts Across Six DSM Editions Yaakov Ophir, Yaffa Shir-Raz, Refael Tikochinski This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6405218/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Considering the central role of the Diagnostic and Statistical Manual of Mental Disorders (DSM) in psychiatric classification, multiple studies have examined how it describes Attention-Deficit/Hyperactivity Disorder (ADHD)—one of the most common psychiatric diagnoses. However, despite analyzing the same DSM texts, these studies yielded conflicting conclusions, likely influenced by the subjectivity of qualitative research and the challenge of systematically tracking subtle changes in large textual corpora. This study addresses these limitations by providing the first systematic, Artificial Intelligence (AI)-assisted analysis of all ADHD-related texts across six DSM editions ( DSM-III to DSM-5-TR ). Methods The analysis employed two AI models ( GPT-4o and Claude 3.5 Sonnet ) and followed five steps: (A) preliminary human review, (B) AI-assisted comparison, (C) refinement through self-prompting to detect subtle linguistic changes, (D) thematic synthesis by each model, and (E) cross-model validation. Strict adherence to DSM texts ensured that all findings were grounded in verifiable evidence. To reduce interpretive bias, human interpretations were intentionally minimized. Results The complete analysis is available in the Supplementary Materials. Overall, the findings revealed six overarching shifts toward: (1) a neurodevelopmental framework, (2) a lifespan condition across genders, (3) a broader concept of impairment, (4) greater diagnostic flexibility, (5) expanded comorbidities and differential diagnoses, and (6) cultural and contextual influences. Conclusions These trends, together with the full study materials, systematically map how ADHD has been described and classified in the DSM over four decades, offering a structured and transparent foundation for future discussions. ADHD DSM psychiatric classification AI-assisted text analysis Introduction Attention-Deficit/Hyperactivity Disorder (ADHD)—one of the most common psychiatric diagnoses in children 1 —was introduced in the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III). 2 Widely regarded as a turning point in psychiatry, DSM-III adopted a symptom-based, atheoretical approach that prioritized diagnostic reliability and standardization while aligning psychiatry with the medical model. 3,4 As part of this shift, it consolidated several earlier childhood classifications (e.g., Hyperkinetic Reaction of Childhood) into a single diagnosis—Attention Deficit Disorder—stating that " attentional difficulties are prominent and virtually always present among children with these [earlier] diagnoses ". 2, p. 41 Four decades and five revisions later, the current edition (DSM-5-TR) 5 retains a closely related label (Attention-Deficit/Hyperactivity Disorder) but reflects substantial conceptual shifts in how ADHD is defined, classified, and understood. 6–8 These changes have been the focus of extensive research, not only because of the DSM’s authority as the leading psychiatric classification system but also due to its broader influence on clinical practice and societal discourse. 9,10 On the one hand, studies supporting the DSM’s revisions typically interpret them as scientific advancements that enhance diagnostic accuracy. These studies often argue that the rising prevalence of ADHD (e.g., 11,12 ) reflects improved recognition of previously underdiagnosed cases, particularly among adults and females. 13–16 On the other hand, critics question the empirical basis of these changes and suggest that the increasing prevalence may reflect diagnostic expansion, raising concerns about potential overdiagnosis. 8,17,18 This divergence between supporters and critics of the DSM’s descriptions—despite analyzing the same texts—may stem from the inherent subjectivity and complexity of traditional text analysis. Human reviewers may struggle to systematically track subtle textual changes across multiple DSM editions, and their interpretations may be shaped by preexisting perspectives. Furthermore, to our knowledge, no prior study has systematically analyzed all DSM editions that have described ADHD. The Current Study This study presents the first systematic analysis of all available descriptions of ADHD across the six editions of the psychiatric manual (DSM-III to DSM-5-TR). To address the limitations of prior research, we developed a structured analytical pipeline that leveraged two state-of-the-art AI language models (Fig. 1). This approach enabled the identification of subtle textual changes across DSM editions, incorporated cross-validation between the models, and required all findings to be grounded in verbatim quotations from the DSM texts. Importantly, we intentionally minimized potentially biased human interpretations throughout the process. All methodological steps—including the wording of prompts, model outputs, and validation procedures—are fully documented in the following sections and in the Supplementary Materials. In this way, this study provides a structured, transparent, and replicable reference point for clinical discussions about ADHD and its evolving conceptualization. Methods The dataset included all relevant ADHD-related texts from DSM-III (1980) through DSM-5-TR (2022), totaling 17,580 words. We analyzed this corpus using a structured five-step pipeline developed to systematically track conceptual and linguistic shifts in ADHD descriptions (Fig. 1). Figure 1 Overview of the Five-Stage AI-Assisted Analytical Pipeline Step 1: Preliminary Human Review Two researchers manually examined the compiled DSM texts to familiarize themselves with key conceptual shifts and establish the first prompts for the subsequent AI-assisted analysis. This initial phase also provided a reference point for evaluating the accuracy, relevance, and consistency of the AI-generated results throughout the entire pipeline. Step 2: Initial AI-Assisted Comparative Analysis The full ADHD corpus was analyzed using two leading AI language models, GPT-4o and Claude 3.5 Sonnet. 19,20 Each model was instructed to detect key differences across DSM editions while being strictly bound to the original textual corpus. All AI findings were required to be supported by direct verbatim quotations, which were then validated against the original DSM texts ( Supplementary Materials 1 and 2 ). Step 3: Refinement of the Analysis Through AI Self-Prompting To enhance the depth and precision of the analysis, each AI model was tasked with generating its own set of analytical prompts, tailored to its linguistic strengths. This step addressed limitations of traditional qualitative analysis, particularly in systematically detecting subtle textual changes across large corpora. In response, GPT-4o and Claude 3.5 Sonnet generated 10 and 7 targeted prompts, respectively, focusing on aspects such as shifts in language and tone, as well as diagnostic uncertainties (Table 1 ). Table 1 Thematic Focus of AI Self-Prompting in GPT-4o and Claude 3.5 Sonnet GPT-4o Claude 3.5 Conceptual evolution of ADHD over time Changes in diagnostic criteria Subtle shifts in descriptive language and tone Evolution of ADHD-related risk factors Changes in ADHD’s relationship with comorbid conditions Differences in functional impairment and prognosis descriptions Cultural and gender bias in ADHD diagnosis Neurobiological and genetic explanations over time Changes in ADHD prevalence estimates ADHD in adults: Emerging recognition Language pattern analysis Contextual framework evolution Diagnostic boundary analysis Gender and cultural considerations Temporal language analysis Impairment conceptualization Diagnostic uncertainty All model-generated prompts were manually reviewed to ensure their relevance and accuracy before being reintroduced into their respective models. Importantly, all prompts explicitly required the models to substantiate their findings with verbatim quotations from the DSM texts, which were attached as a document file each time ( Supplementary Materials 1 and 2 ). Step 4: Thematic Synthesis by Each Model The complete AI-generated findings were then categorized into structured thematic groups by the respective AI models. Each model was instructed to: (a) summarize the most significant changes in each DSM edition, (b) identify overarching trends in how ADHD has been defined, framed, and classified over time, (c) analyze the implications of these changes, and (d) support all responses with verbatim quotations. Altogether, this process (Steps 2–4) produced two comprehensive files: one containing the full analysis from GPT-4o (16,512 words) and another from Claude 3.5 Sonnet (8,631 words). These files, which represent the complete AI-assisted analyses of ADHD in the DSM, are available in Supplementary Material 1 and Supplementary Material 2 . Step 5: Cross-Model Validation In the final step, each model was provided with the two comprehensive files containing the full AI-assisted analyses and tasked with identifying key insights that both models independently recognized and agreed upon. The models were instructed to rely solely on the AI-generated analyses without introducing new interpretations or external knowledge ( Supplementary Material 3 ). Next, the two sets of agreed-upon insights were compiled and fed into GPT-4o, which was tasked with synthesizing a final summary of the changes that both models determined to be consensual. This final comparison represents the culmination of the methodological process, systematically consolidating the findings and highlighting the major conceptual shifts in ADHD across six DSM editions ( Supplementary Material 4 ). Results The final results of the AI-assisted analytical pipeline ( Supplementary Material 4 ) were manually reviewed, systematically organized, and synthesized into six overarching shifts. 1. Neurodevelopmental Framework In its initial formulation (DSM-III), ADHD was conceptualized primarily as a behavioral disorder. Over time, its classification evolved into a neurodevelopmental framework, incorporating genetic, neurobiological, and cognitive components (DSM-5, DSM-5-TR). DSM-5 also marked the first introduction of the distinct diagnostic category Neurodevelopmental Disorders , under which ADHD was placed. In all previous editions, ADHD had been classified under Disorders Usually First Diagnosed in Infancy, Childhood, or Adolescence . Notably, despite this shift, recent editions (DSM-5 and DSM-5-TR) also mention that " no biological marker is diagnostic for ADHD ." DSM-5-TR further clarifies that " meta-analyses of all neuroimaging studies do not show differences between individuals with ADHD and control subjects " and concludes that " until these issues are resolved, no form of neuroimaging can be used for diagnosis of ADHD ." 2. Lifespan Condition Across Genders In its early editions, the DSM conceptualized ADHD as a childhood-specific disorder, with no formal recognition of persistence into adulthood (DSM-III, DSM-III-R). DSM-III estimated its prevalence at 3% among prepubertal children in the U.S, stating it was " ten times more common in boys than in girls ." Over time, the reported prevalence among children steadily increased, reaching 7.2% in DSM-5-TR. Newer editions also acknowledged gender differences in ADHD presentation, with DSM-5-TR stating that " females are more likely than males to present primarily with inattentive features ." Correspondingly, the male-to-female prevalence ratio has gradually declined from 10:1 (DSM-III) to 6–9:1 (DSM-III-R), 4–9:1 (DSM-IV), and approximately 2:1 (DSM-5 and DSM-5-TR). Among adults the current reported ratio is 1.6:1. In parallel, DSM-5 formally recognized adult ADHD, lowered the diagnostic threshold for individuals aged 17 and older, and acknowledged its persistence across the lifespan. The required age of onset was also extended from before age 7 (DSM-III through DSM-IV-TR) to before age 12 (DSM-5). DSM-5 also introduced the first specific prevalence estimate for adult ADHD (2.5%). 3. Impairment Early DSM editions primarily defined ADHD based on core symptoms of inattention and hyperactivity, with impairment framed mostly in terms of academic underperformance (DSM-III). Over time, later editions (DSM-5, DSM-5-TR) placed increasing emphasis on functional impairment across multiple domains, including occupational, social, and emotional regulation difficulties. A designated section on " Functional Consequences of Attention-Deficit/Hyperactivity Disorder " was introduced in DSM-5 and DSM-5-TR. DSM-5 listed severe negative consequences such as antisocial personality disorder, substance use disorders, incarceration, traffic accidents, obesity, and peer rejection. DSM-5-TR further extended this list, adding poor job stability, lower self-esteem, increased risk of trauma and subsequent PTSD, as well as a higher overall mortality rate—primarily due to accidents and injuries. In addition to functional impairment, the relationship between ADHD and suicide risk was first introduced in DSM-5 and further elaborated in DSM-5-TR. DSM-5 mentioned an " increased risk of suicide attempt " under the section " Associated Features Supporting Diagnosis. " DSM-5-TR expanded on this by introducing a dedicated section titled " Association With Suicidal Thoughts or Behavior ," explicitly stating that " ADHD is a risk factor for suicidal ideation and behavior in children. " Notably, however, the direct texts on the impairment criterion have fluctuated over the years in a non-linear trend: Early editions (DSM-III, DSM-III-R) set relatively mild dysfunction thresholds for diagnosis. DSM-III stated that " academic difficulties are common; and although impairment may be limited to academic functioning, social functioning may be impaired as well ." DSM-III-R similarly noted that " some impairment in social and school functioning is common " and introduced severity criteria, including a " mild " ADHD category, where symptoms caused " only minimal or no impairment in school and social functioning ." DSM-IV and DSM-IV-TR appear to have made the impairment criterion more stringent. The severity criteria were omitted, and a stricter " Criterion D " was introduced, requiring " clear evidence of clinically significant impairment in social, academic, or occupational functioning ." Finally, DSM-5 and DSM-5-TR reintroduced severity levels, once again allowing for a " mild " ADHD diagnosis. DSM-5-TR further introduced the category " in partial remission " to describe cases where individuals no longer meet the full diagnostic criteria but continue to experience some symptoms. Correspondingly, Criterion D was revised to require " clear evidence that the symptoms interfere with, or reduce the quality of, social, academic, or occupational functioning ," reflecting a more flexible approach to defining impairment. 4. Diagnostic Flexibility The current flexibility in the impairment criterion (Finding 3) appears to be part of a broader trend toward a more adaptable diagnostic framework. ADHD was initially categorized into subtypes, with DSM-III distinguishing between ADD with and without hyperactivity. DSM-IV restructured these classifications into three subtypes: Predominantly Inattentive, Predominantly Hyperactive-Impulsive, and Combined Type. DSM-5 later replaced subtypes with presentations , allowing for greater diagnostic flexibility and acknowledging that symptom profiles may change over time. This shift is also reflected in the DSM’s evolving approach to subthreshold cases of ADHD. DSM-III (1980) introduced Residual Type of the disorder, which referred to individuals who had previously met full diagnostic criteria but continued to experience some symptoms in a reduced form. DSM-III-R (1987) removed Residual Type without a direct replacement but introduced a severity scale, including a " mild " ADHD category as mentioned above (Finding 3) – a category, which may have compensated for the removal of the Residual Type. DSM-IV and DSM-IV-TR replaced Residual Type with ADHD Not Otherwise Specified (NOS) , broadening the framework to include clinically significant cases that did not fully meet ADHD criteria, even if they had never met the full criteria before. DSM-5 and DSM-5-TR refined this structure further, replacing NOS with two distinct diagnostic alternatives: Other Specified ADHD (used when clinicians specify why a presentation does not fully meet ADHD criteria) and Unspecified ADHD (used when clinicians either do not specify the reason or lack sufficient information). 5. Comorbidities and Differential Diagnoses In DSM-III, ADHD was primarily linked to conduct disorder and learning disabilities. Over time, DSM-5 and DSM-5-TR expanded its comorbidity profile to include autism spectrum disorder (ASD), mood disorders, substance use disorders, and anxiety disorders, reflecting a broader and more complex clinical framework (see also Finding 3). Simultaneously, the differential diagnosis section expanded considerably. DSM-III and DSM-III-R listed 5 and 6 conditions, respectively, including “ age-appropriate overactivity ” as a diagnostic consideration. DSM-IV and DSM-IV-TR increased this to 10, while DSM-5 and DSM-5-TR further expanded it to 16 and 17. Notably, the updated lists excluded “ age-appropriate overactivity ” and did not include “ normal variations ,” a consideration found in the differential diagnosis of some other neurodevelopmental disorders. 6. Contextual and Cultural Considerations Finally, despite the shift towards a neurodevelopmental framework (Finding 1), contemporary DSM editions have expanded their acknowledgment of environmental and contextual factors in shaping symptom expression. While earlier DSM editions made little to no reference to cultural considerations, DSM-5-TR explicitly recognizes racial and ethnic disparities, highlighting patterns of underdiagnosis in minority populations and the potential influence of clinician bias on diagnostic practices. Additionally, recent editions have incorporated environmental influences on symptom manifestation, acknowledging the roles of family dynamics, digital environments, and workplace settings in modulating the presentation and severity of ADHD symptoms. Discussion This study provides the first systematic, AI-assisted analysis of how ADHD has been described and classified in the DSM over the past four decades. By tracing changes across six DSM editions, the analysis identified six overarching trends: ADHD has shifted from a primarily behavioral disorder to a neurodevelopmental condition, emphasizing biological and cognitive underpinnings. Initially conceptualized as a disorder of young boys, ADHD has evolved into a lifespan condition affecting individuals across genders. The reported prevalence in children increased from 3–7.2%, while the male-to-female ratio declined from 10:1 to 2:1. Simultaneously, DSM-5 formally recognized adult ADHD (with a 2.5% prevalence), lowered the symptom threshold for older individuals, and extended the required age of onset from 7 to 12 years. The definition of impairment has evolved, broadening beyond academic difficulties to include occupational, social, and emotional challenges. A range of negative consequences were introduced including, for example, incarceration, traffic accidents, overall mortality, and suicide. At the same time, the latest editions have also lowered the required threshold for impairment, allowing for the diagnosis of individuals with minor difficulties. Diagnostic criteria have become more flexible, moving from rigid subtypes to a classification that accommodates changing symptom presentations, including cases that do not fully meet the diagnostic criteria (categorized as Other Specified and Unspecified ADHD). The scope of ADHD’s comorbidities and differential diagnoses has expanded significantly, with DSM-5-TR listing 17 differential diagnoses—excluding 'normal variations.' While the DSM increasingly frames ADHD as a neurobiological disorder, later editions also acknowledge the role of cultural and environmental factors (e.g., digital environments) in shaping symptom expressions. These findings were derived from a dual-model AI pipeline—based on GPT-4o and Claude 3.5 Sonnet—and subjected to cross-model validation (Fig. 1). All findings were grounded in verbatim quotations from the DSM and detailed in four Supplementary Materials. The contributions of this study are threefold. First, it leveraged the advanced language capabilities of large AI models to systematically detect subtle textual shifts across multiple and complex editions of the DSM. Second, it offers a relatively objective and reproducible alternative to traditional qualitative approaches, which—despite examining the same texts—have produced conflicting interpretations of ADHD’s evolving description (e.g., 8,13,14,17 ). Third, it provides a fully transparent account of its methods and findings (see the four Supplementary Materials), offering a comprehensive and replicable reference point for researchers, clinicians, and policymakers concerned with ADHD classification and its implications. We recommend that future discussions consider these findings in light of the DSM’s own definition of a mental disorder, which emphasizes “clinically significant disturbance” and cautions against diagnosing “an expectable or culturally approved response to a common stressor” or “conflicts that are primarily between the individual and society.” In this study, we intentionally refrained from taking a position on this issue. Instead, we aimed to present the analysis with minimal human interpretation, allowing readers to evaluate whether the observed conceptual shifts reflect improved identification of previously overlooked cases or a broadening of diagnostic boundaries that may carry a risk of overdiagnosis. 21,22 In either case, these shifts underscore the importance of continued reflection on ADHD’s classification and its implications for clinical practice and public health policy. Limitations While this study provides a well-validated analysis of ADHD-related texts, several limitations should be acknowledged. First, the study focused exclusively on DSM texts, which, while serving as the most authoritative psychiatric classification system, represent only one aspect of how ADHD is conceptualized. The evolution of ADHD is influenced by broader forces, including empirical research, clinical practice, educational policies, and public discourse. Second, while this study critically examined shifts in the DSM’s portrayal of ADHD, it did not empirically assess their impact on real-world clinical practice, such as changes in prevalence rates, prescribing patterns, or patient outcomes. Future research should integrate real-world data to examine how DSM modifications translate into clinical decision-making. Third, despite employing state-of-the-art AI language models (GPT-4o and Claude 3.5 Sonnet), their proprietary nature poses challenges for reproducibility, as ongoing model updates may alter their outputs. Fourth, AI models are susceptible to biases based on their training data, including frequency effects that may skew their focus toward more recent DSM editions. Finally, AI models function as black-box systems, meaning their internal decision-making processes remain opaque. Although cross-validation measures were implemented and models were required to support all findings with verbatim DSM quotations to mitigate biases and hallucinations, human expertise remains essential for interpreting findings and contextualizing them within the broader scientific and clinical landscape. Future studies may benefit from utilizing open-weight AI models, which make their parameters publicly available, improving transparency, reproducibility, and researcher control over the analysis. Additionally, integrating hybrid methodologies that combine AI analysis with traditional qualitative research may further strengthen findings by providing a more nuanced understanding of ADHD’s evolving conceptualization. Despite these limitations, this study establishes a structured, transparent, and replicable foundation for future research, offering valuable insights for researchers, clinicians, and policymakers seeking to better understand the shifting diagnostic landscape of ADHD. Declarations Ethics, Consent to Participate, and Consent to Publish Declarations: Not applicable. This study is not a clinical trial. It is based on a textual analysis of the Diagnostic and Statistical Manual of Mental Disorders (DSM) and does not involve human participants or require ethical approval. Funding: The authors received no funding for this study. Conflicts of Interest: The authors declare no conflicts of interest. Availability of Data: All materials generated and analyzed in this study are available in the four Supplementary Materials accompanying this article. Author Contribution All authors contributed meaningfully to the development of this article. Y.O. conceptualized the study, reviewed all DSM texts, conducted the AI-assisted analysis, and drafted the manuscript. Y.S.R. and R.T. conducted a comprehensive review of the existing literature, thoroughly examined the study materials, actively contributed to internal discussions, and provided critical feedback and revisions to the manuscript. All authors approved the final version. The authors would also like to thank Refael Reuveni from Ariel University for his invaluable assistance in laying the groundwork for the current study, particularly through his work on preliminary trials with the AI models. References Epstein JN, Loren REA. Changes in the definition of ADHD in DSM-5: subtle but important. 2013; 3 : 455. Anthropic. Claude 3.5 Sonnet. 2024. Crowe M. Constructing normality: a discourse analysis of the DSM‐IV. 2000; 7 : 69–77. Voort JLV, He J-P, Jameson ND, Merikangas KR. Impact of the DSM-5 attention-deficit/hyperactivity disorder age-of-onset criterion in the US adolescent population. 2014; 53 : 736–744. Sibley MH, Kuriyan AB. DSM-5 changes enhance parent identification of symptoms in adolescents with ADHD. 2016; 242 : 180–185. Freedman JE, Honkasilta JM. Dictating the boundaries of ab/normality: a critical discourse analysis of the diagnostic criteria for attention deficit hyperactivity disorder and hyperkinetic disorder. 2017; 32 : 565–588. APA. Diagnostic and Statistical Manual of Mental Disorders - Third Edition (DSM-III) . American Psychiatric Association (APA), 1980. Wilson M. DSM-III and the transformation of American psychiatry: A history. 1993; 150 : 399. Lefler EK, Stevens AE, Garner AM, Serrano JW, Canu WH, Hartung CM. Changes in college student endorsement of ADHD symptoms across DSM edition. 2020; 42 : 488–499. APA. Diagnostic and Statistical Manual of Mental Disorders, 5th edition, Text Revision (DSM-5-TR). 2022. Rigler T, Manor I, Kalansky A, Shorer Z, Noyman I, Sadaka Y. New DSM-5 criteria for ADHD—Does it matter? 2016; 68 : 56–59. Sklepníková J, Slezáčková A. Evolution of the ADHD concept and its relation to the hyperkinetic disorders: Narrative review. 2022; 66 : 255–271. Doernberg E, Hollander E. Neurodevelopmental disorders (asd and adhd): Dsm-5, icd-10, and icd-11. 2016; 21 : 295–299. WHO. World mental health report: Transforming mental health for all . World Health Organization. Last retrieved on December 17, 2024 from: https://iris.who.int/bitstream/handle/10665/356119/9789240049338-eng.pdf?sequence=1, 2022. Danielson ML, Bitsko RH, Ghandour RM, Holbrook JR, Kogan MD, Blumberg SJ. Prevalence of Parent-Reported ADHD Diagnosis and Associated Treatment Among U.S. Children and Adolescents, 2016. 2018; 47 : 199–212. Achiam J, Adler S, Agarwal S, Ahmad L, Akkaya I, Aleman FL et al. Gpt-4 technical report. 2023. Kazda L, Bell K, Thomas R, Hardiman L, Heath I, Barratt A. Attention deficit/hyperactivity disorder (ADHD) in children: more focus on care and support, less on diagnosis. 2024; 384 . Visser SN, Danielson ML, Bitsko RH, Holbrook JR, Kogan MD, Ghandour RM et al. Trends in the parent-report of health care provider-diagnosed and medicated attention-deficit/hyperactivity disorder: United States, 2003–2011. 2014; 53 : 34–46. Mallett CA, Natarajan A, Hoy J. Attention deficit/hyperactivity disorder: A DSM timeline review. 2014; 43 : 36–60. Hyman SE. The diagnosis of mental disorders: the problem of reification. 2010; 6 : 155–179. Deacon BJ. The biomedical model of mental disorder: a critical analysis of its validity, utility, and effects on psychotherapy research. 2013; 33 : 846–861. Kazda L, Bell K, Thomas R, McGeechan K, Sims R, Barratt A. Overdiagnosis of Attention-Deficit/Hyperactivity Disorder in Children and Adolescents: A Systematic Scoping Review. 2021; 4 : e215335. Additional Declarations No competing interests reported. Supplementary Files SupplementaryMaterial1FullGPT4oAnalysisTranscript.pdf SupplementaryMaterial2FullClaude3.5SonnetAnalysisTranscript.pdf SupplementaryMaterial3ComparisonofAgreedUponKeyDifferences.pdf SupplementaryMaterial4FinalStructuralComparisonoftheAgreedUponDifferences.pdf Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6405218","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":454999346,"identity":"210f1146-682a-4c23-9051-e6b4be5ac92a","order_by":0,"name":"Yaakov Ophir","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5ElEQVRIiWNgGAWjYHACZjBpwMzA+ABI8/CRooXZAKSFjXgtDAxsEiAGQS0Gxw8/NvjYZpNnzs77rPJrjp0MGwPzw0c38Gk5k2acOLMtrdiymd3stuy2ZKDD2IyNc/BpOZBgfJjnzOHEDYfZ2G5LbmMGauFhk8ar5fzzz3AtxZLb6onQciPHOJmnAqKF8eO2w4S1SN54U2w4oyKt2OAwG7M047bjPGzMBPzCdz59s8QHA5s8g/PHGD/+3FZtz8/e/PAxPi0KByB0Aohg5gGTeJSDgHwDkhbGHwRUj4JRMApGwcgEABkeRbh6pRudAAAAAElFTkSuQmCC","orcid":"","institution":"Ariel University","correspondingAuthor":true,"prefix":"","firstName":"Yaakov","middleName":"","lastName":"Ophir","suffix":""},{"id":454999347,"identity":"ab74806f-c533-43df-b231-3d6a57a96653","order_by":1,"name":"Yaffa Shir-Raz","email":"","orcid":"","institution":"Reichman University","correspondingAuthor":false,"prefix":"","firstName":"Yaffa","middleName":"","lastName":"Shir-Raz","suffix":""},{"id":454999348,"identity":"975fbb5e-c63b-4258-a129-044d48f5fa05","order_by":2,"name":"Refael Tikochinski","email":"","orcid":"","institution":"Ariel University","correspondingAuthor":false,"prefix":"","firstName":"Refael","middleName":"","lastName":"Tikochinski","suffix":""}],"badges":[],"createdAt":"2025-04-08 16:38:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6405218/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6405218/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":98624605,"identity":"c2511e6d-d935-4047-817a-1045b31c0e04","added_by":"auto","created_at":"2025-12-19 17:08:33","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":644404,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6405218/v1/f074d6bd-9313-41da-8dd8-0368f3222a7b.pdf"},{"id":82541783,"identity":"6db7a7ee-d77d-4aee-87df-f3303e538888","added_by":"auto","created_at":"2025-05-12 16:58:56","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":221712,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial1FullGPT4oAnalysisTranscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6405218/v1/41479ff44b6d973333009edc.pdf"},{"id":82542101,"identity":"d7e3378a-107b-442c-8eaf-ce03c80aeeb0","added_by":"auto","created_at":"2025-05-12 17:06:56","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":387023,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial2FullClaude3.5SonnetAnalysisTranscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6405218/v1/361757b7b73bf0afc82a451e.pdf"},{"id":82541259,"identity":"b7435173-5089-48cc-a71f-aab66314d8e4","added_by":"auto","created_at":"2025-05-12 16:50:56","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":361165,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial3ComparisonofAgreedUponKeyDifferences.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6405218/v1/56aceb3a10524e6ab56d4a46.pdf"},{"id":82541784,"identity":"93ad3df4-1038-4606-a9b6-412f82c8a63a","added_by":"auto","created_at":"2025-05-12 16:58:56","extension":"pdf","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":139655,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial4FinalStructuralComparisonoftheAgreedUponDifferences.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6405218/v1/c3f5b5ccedee7459d13fb494.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Four Decades of ADHD: A Systematic AI-assisted Analysis of Conceptual Shifts Across Six DSM Editions","fulltext":[{"header":"Introduction","content":"\u003cp\u003e \u003cem\u003eAttention-Deficit/Hyperactivity Disorder\u003c/em\u003e (ADHD)—one of the most common psychiatric diagnoses in children\u003csup\u003e1\u003c/sup\u003e—was introduced in the third edition of the \u003cem\u003eDiagnostic and Statistical Manual of Mental Disorders\u003c/em\u003e (DSM-III).\u003csup\u003e2\u003c/sup\u003e Widely regarded as a turning point in psychiatry, DSM-III adopted a symptom-based, atheoretical approach that prioritized diagnostic reliability and standardization while aligning psychiatry with the medical model.\u003csup\u003e3,4\u003c/sup\u003e As part of this shift, it consolidated several earlier childhood classifications (e.g., Hyperkinetic Reaction of Childhood) into a single diagnosis—Attention Deficit Disorder—stating that \"\u003cem\u003eattentional difficulties are prominent and virtually always present among children with these [earlier] diagnoses\u003c/em\u003e\".\u003csup\u003e2, p. 41\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eFour decades and five revisions later, the current edition (DSM-5-TR)\u003csup\u003e5\u003c/sup\u003e retains a closely related label (Attention-Deficit/Hyperactivity Disorder) but reflects substantial conceptual shifts in how ADHD is defined, classified, and understood.\u003csup\u003e6–8\u003c/sup\u003e These changes have been the focus of extensive research, not only because of the DSM’s authority as the leading psychiatric classification system but also due to its broader influence on clinical practice and societal discourse.\u003csup\u003e9,10\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eOn the one hand, studies supporting the DSM’s revisions typically interpret them as scientific advancements that enhance diagnostic accuracy. These studies often argue that the rising prevalence of ADHD (e.g., \u003csup\u003e11,12\u003c/sup\u003e) reflects improved recognition of previously underdiagnosed cases, particularly among adults and females.\u003csup\u003e13–16\u003c/sup\u003e On the other hand, critics question the empirical basis of these changes and suggest that the increasing prevalence may reflect diagnostic expansion, raising concerns about potential overdiagnosis.\u003csup\u003e8,17,18\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eThis divergence between supporters and critics of the DSM’s descriptions—despite analyzing the same texts—may stem from the inherent subjectivity and complexity of traditional text analysis. Human reviewers may struggle to systematically track subtle textual changes across multiple DSM editions, and their interpretations may be shaped by preexisting perspectives. Furthermore, to our knowledge, no prior study has systematically analyzed all DSM editions that have described ADHD.\u003c/p\u003e\n\u003ch3\u003eThe Current Study\u003c/h3\u003e\n\u003cp\u003eThis study presents the first systematic analysis of all available descriptions of ADHD across the six editions of the psychiatric manual (DSM-III to DSM-5-TR). To address the limitations of prior research, we developed a structured analytical pipeline that leveraged two state-of-the-art AI language models (Fig.\u0026nbsp;1). This approach enabled the identification of subtle textual changes across DSM editions, incorporated cross-validation between the models, and required all findings to be grounded in verbatim quotations from the DSM texts.\u003c/p\u003e \u003cp\u003eImportantly, we intentionally minimized potentially biased human interpretations throughout the process. All methodological steps—including the wording of prompts, model outputs, and validation procedures—are fully documented in the following sections and in the Supplementary Materials. In this way, this study provides a structured, transparent, and replicable reference point for clinical discussions about ADHD and its evolving conceptualization.\u003c/p\u003e \n\n\u003cp\u003e \u003c/p\u003e\n\n\n\n\n\n \u003cp\u003e\u003c/p\u003e \n\n \n\n "},{"header":"Methods","content":"\u003cp\u003eThe dataset included all relevant ADHD-related texts from DSM-III (1980) through DSM-5-TR (2022), totaling 17,580 words. We analyzed this corpus using a structured five-step pipeline developed to systematically track conceptual and linguistic shifts in ADHD descriptions (Fig.\u0026nbsp;1).\u003c/p\u003e\u003cp\u003eFigure 1\u003c/p\u003e\u003ch3\u003eOverview of the Five-Stage AI-Assisted Analytical Pipeline\u003c/h3\u003e\u003ch3\u003eStep 1: Preliminary Human Review\u003c/h3\u003e\u003cp\u003eTwo researchers manually examined the compiled DSM texts to familiarize themselves with key conceptual shifts and establish the first prompts for the subsequent AI-assisted analysis. This initial phase also provided a reference point for evaluating the accuracy, relevance, and consistency of the AI-generated results throughout the entire pipeline.\u003c/p\u003e\u003ch3\u003eStep 2: Initial AI-Assisted Comparative Analysis\u003c/h3\u003e\u003cp\u003eThe full ADHD corpus was analyzed using two leading AI language models, GPT-4o and Claude 3.5 Sonnet.\u003csup\u003e19,20\u003c/sup\u003e Each model was instructed to detect key differences across DSM editions while being strictly bound to the original textual corpus. All AI findings were required to be supported by direct verbatim quotations, which were then validated against the original DSM texts (\u003cem\u003eSupplementary Materials 1 and 2\u003c/em\u003e).\u003c/p\u003e\u003ch3\u003eStep 3: Refinement of the Analysis Through AI Self-Prompting\u003c/h3\u003e\u003cp\u003eTo enhance the depth and precision of the analysis, each AI model was tasked with generating its own set of analytical prompts, tailored to its linguistic strengths. This step addressed limitations of traditional qualitative analysis, particularly in systematically detecting subtle textual changes across large corpora. In response, GPT-4o and Claude 3.5 Sonnet generated 10 and 7 targeted prompts, respectively, focusing on aspects such as shifts in language and tone, as well as diagnostic uncertainties (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cem\u003eThematic Focus of AI Self-Prompting in GPT-4o and Claude 3.5 Sonnet\u003c/em\u003e\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e\u003ccolgroup cols=\"2\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGPT-4o\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eClaude 3.5\u003c/p\u003e \u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConceptual evolution of ADHD over time\u003c/p\u003e \u003cp\u003eChanges in diagnostic criteria\u003c/p\u003e \u003cp\u003eSubtle shifts in descriptive language and tone\u003c/p\u003e \u003cp\u003eEvolution of ADHD-related risk factors\u003c/p\u003e \u003cp\u003eChanges in ADHD’s relationship with comorbid conditions\u003c/p\u003e \u003cp\u003eDifferences in functional impairment and prognosis descriptions\u003c/p\u003e \u003cp\u003eCultural and gender bias in ADHD diagnosis\u003c/p\u003e \u003cp\u003eNeurobiological and genetic explanations over time\u003c/p\u003e \u003cp\u003eChanges in ADHD prevalence estimates\u003c/p\u003e \u003cp\u003eADHD in adults: Emerging recognition\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLanguage pattern analysis\u003c/p\u003e \u003cp\u003eContextual framework evolution\u003c/p\u003e \u003cp\u003eDiagnostic boundary analysis\u003c/p\u003e \u003cp\u003eGender and cultural considerations\u003c/p\u003e \u003cp\u003eTemporal language analysis\u003c/p\u003e \u003cp\u003eImpairment conceptualization\u003c/p\u003e \u003cp\u003eDiagnostic uncertainty\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\u003c/div\u003e\u003cp\u003eAll model-generated prompts were manually reviewed to ensure their relevance and accuracy before being reintroduced into their respective models. Importantly, all prompts explicitly required the models to substantiate their findings with verbatim quotations from the DSM texts, which were attached as a document file each time (\u003cem\u003eSupplementary Materials 1 and 2\u003c/em\u003e).\u003c/p\u003e\u003ch3\u003eStep 4: Thematic Synthesis by Each Model\u003c/h3\u003e\u003cp\u003eThe complete AI-generated findings were then categorized into structured thematic groups by the respective AI models. Each model was instructed to: (a) summarize the most significant changes in each DSM edition, (b) identify overarching trends in how ADHD has been defined, framed, and classified over time, (c) analyze the implications of these changes, and (d) support all responses with verbatim quotations.\u003c/p\u003e\u003cp\u003eAltogether, this process (Steps 2–4) produced two comprehensive files: one containing the full analysis from GPT-4o (16,512 words) and another from Claude 3.5 Sonnet (8,631 words). These files, which represent the complete AI-assisted analyses of ADHD in the DSM, are available in \u003cem\u003eSupplementary Material 1\u003c/em\u003e and \u003cem\u003eSupplementary Material 2\u003c/em\u003e.\u003c/p\u003e\u003ch3\u003eStep 5: Cross-Model Validation\u003c/h3\u003e\u003cp\u003eIn the final step, each model was provided with the two comprehensive files containing the full AI-assisted analyses and tasked with identifying key insights that both models independently recognized and agreed upon. The models were instructed to rely solely on the AI-generated analyses without introducing new interpretations or external knowledge (\u003cem\u003eSupplementary Material 3\u003c/em\u003e).\u003c/p\u003e\u003cp\u003eNext, the two sets of agreed-upon insights were compiled and fed into GPT-4o, which was tasked with synthesizing a final summary of the changes that both models determined to be consensual. This final comparison represents the culmination of the methodological process, systematically consolidating the findings and highlighting the major conceptual shifts in ADHD across six DSM editions (\u003cem\u003eSupplementary Material 4\u003c/em\u003e).\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eThe final results of the AI-assisted analytical pipeline (\u003cem\u003eSupplementary Material 4\u003c/em\u003e) were manually reviewed, systematically organized, and synthesized into six overarching shifts.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e1. Neurodevelopmental Framework\u003c/h2\u003e \u003cp\u003eIn its initial formulation (DSM-III), ADHD was conceptualized primarily as a behavioral disorder. Over time, its classification evolved into a neurodevelopmental framework, incorporating genetic, neurobiological, and cognitive components (DSM-5, DSM-5-TR). DSM-5 also marked the first introduction of the distinct diagnostic category \u003cem\u003eNeurodevelopmental Disorders\u003c/em\u003e, under which ADHD was placed. In all previous editions, ADHD had been classified under \u003cem\u003eDisorders Usually First Diagnosed in Infancy, Childhood, or Adolescence\u003c/em\u003e.\u003c/p\u003e \u003cp\u003eNotably, despite this shift, recent editions (DSM-5 and DSM-5-TR) also mention that \"\u003cem\u003eno biological marker is diagnostic for ADHD\u003c/em\u003e.\" DSM-5-TR further clarifies that \"\u003cem\u003emeta-analyses of all neuroimaging studies do not show differences between individuals with ADHD and control subjects\u003c/em\u003e\" and concludes that \"\u003cem\u003euntil these issues are resolved, no form of neuroimaging can be used for diagnosis of ADHD\u003c/em\u003e.\"\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e2. Lifespan Condition Across Genders\u003c/h2\u003e \u003cp\u003eIn its early editions, the DSM conceptualized ADHD as a childhood-specific disorder, with no formal recognition of persistence into adulthood (DSM-III, DSM-III-R). DSM-III estimated its prevalence at 3% among prepubertal children in the U.S, stating it was \"\u003cem\u003eten times more common in boys than in girls\u003c/em\u003e.\"\u003c/p\u003e \u003cp\u003eOver time, the reported prevalence among children steadily increased, reaching 7.2% in DSM-5-TR. Newer editions also acknowledged gender differences in ADHD presentation, with DSM-5-TR stating that \"\u003cem\u003efemales are more likely than males to present primarily with inattentive features\u003c/em\u003e.\" Correspondingly, the male-to-female prevalence ratio has gradually declined from 10:1 (DSM-III) to 6\u0026ndash;9:1 (DSM-III-R), 4\u0026ndash;9:1 (DSM-IV), and approximately 2:1 (DSM-5 and DSM-5-TR). Among adults the current reported ratio is 1.6:1.\u003c/p\u003e \u003cp\u003eIn parallel, DSM-5 formally recognized adult ADHD, lowered the diagnostic threshold for individuals aged 17 and older, and acknowledged its persistence across the lifespan. The required age of onset was also extended from before age 7 (DSM-III through DSM-IV-TR) to before age 12 (DSM-5). DSM-5 also introduced the first specific prevalence estimate for adult ADHD (2.5%).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3. Impairment\u003c/h2\u003e \u003cp\u003eEarly DSM editions primarily defined ADHD based on core symptoms of inattention and hyperactivity, with impairment framed mostly in terms of academic underperformance (DSM-III). Over time, later editions (DSM-5, DSM-5-TR) placed increasing emphasis on functional impairment across multiple domains, including occupational, social, and emotional regulation difficulties.\u003c/p\u003e \u003cp\u003eA designated section on \"\u003cem\u003eFunctional Consequences of Attention-Deficit/Hyperactivity Disorder\u003c/em\u003e\" was introduced in DSM-5 and DSM-5-TR. DSM-5 listed severe negative consequences such as antisocial personality disorder, substance use disorders, incarceration, traffic accidents, obesity, and peer rejection. DSM-5-TR further extended this list, adding poor job stability, lower self-esteem, increased risk of trauma and subsequent PTSD, as well as a higher overall mortality rate\u0026mdash;primarily due to accidents and injuries.\u003c/p\u003e \u003cp\u003eIn addition to functional impairment, the relationship between ADHD and suicide risk was first introduced in DSM-5 and further elaborated in DSM-5-TR. DSM-5 mentioned an \"\u003cem\u003eincreased risk of suicide attempt\u003c/em\u003e\" under the section \"\u003cem\u003eAssociated Features Supporting Diagnosis.\u003c/em\u003e\" DSM-5-TR expanded on this by introducing a dedicated section titled \"\u003cem\u003eAssociation With Suicidal Thoughts or Behavior\u003c/em\u003e,\" explicitly stating that \"\u003cem\u003eADHD is a risk factor for suicidal ideation and behavior in children.\u003c/em\u003e\"\u003c/p\u003e \u003cp\u003eNotably, however, the direct texts on the impairment criterion have fluctuated over the years in a non-linear trend:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eEarly editions (DSM-III, DSM-III-R) set relatively mild dysfunction thresholds for diagnosis. DSM-III stated that \"\u003cem\u003eacademic difficulties are common; and although impairment may be limited to academic functioning, social functioning may be impaired as well\u003c/em\u003e.\" DSM-III-R similarly noted that \"\u003cem\u003esome impairment in social and school functioning is common\u003c/em\u003e\" and introduced severity criteria, including a \"\u003cem\u003emild\u003c/em\u003e\" ADHD category, where symptoms caused \"\u003cem\u003eonly minimal or no impairment in school and social functioning\u003c/em\u003e.\"\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eDSM-IV and DSM-IV-TR appear to have made the impairment criterion more stringent. The severity criteria were omitted, and a stricter \"\u003cem\u003eCriterion D\u003c/em\u003e\" was introduced, requiring \"\u003cem\u003eclear evidence of clinically significant impairment in social, academic, or occupational functioning\u003c/em\u003e.\"\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eFinally, DSM-5 and DSM-5-TR reintroduced severity levels, once again allowing for a \"\u003cem\u003emild\u003c/em\u003e\" ADHD diagnosis. DSM-5-TR further introduced the category \"\u003cem\u003ein partial remission\u003c/em\u003e\" to describe cases where individuals no longer meet the full diagnostic criteria but continue to experience some symptoms. Correspondingly, Criterion D was revised to require \"\u003cem\u003eclear evidence that the symptoms interfere with, or reduce the quality of, social, academic, or occupational functioning\u003c/em\u003e,\" reflecting a more flexible approach to defining impairment.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e4. Diagnostic Flexibility\u003c/h2\u003e \u003cp\u003eThe current flexibility in the impairment criterion (Finding 3) appears to be part of a broader trend toward a more adaptable diagnostic framework. ADHD was initially categorized into subtypes, with DSM-III distinguishing between ADD with and without hyperactivity. DSM-IV restructured these classifications into three subtypes: Predominantly Inattentive, Predominantly Hyperactive-Impulsive, and Combined Type. DSM-5 later replaced subtypes with \u003cem\u003epresentations\u003c/em\u003e, allowing for greater diagnostic flexibility and acknowledging that symptom profiles may change over time.\u003c/p\u003e \u003cp\u003eThis shift is also reflected in the DSM\u0026rsquo;s evolving approach to subthreshold cases of ADHD.\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eDSM-III (1980) introduced \u003cem\u003eResidual Type\u003c/em\u003e of the disorder, which referred to individuals who had previously met full diagnostic criteria but continued to experience some symptoms in a reduced form.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eDSM-III-R (1987) removed Residual Type without a direct replacement but introduced a severity scale, including a \"\u003cem\u003emild\u003c/em\u003e\" ADHD category as mentioned above (Finding 3) \u0026ndash; a category, which may have compensated for the removal of the Residual Type.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eDSM-IV and DSM-IV-TR replaced Residual Type with \u003cem\u003eADHD Not Otherwise Specified (NOS)\u003c/em\u003e, broadening the framework to include clinically significant cases that did not fully meet ADHD criteria, even if they had never met the full criteria before.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eDSM-5 and DSM-5-TR refined this structure further, replacing NOS with two distinct diagnostic alternatives: \u003cem\u003eOther Specified ADHD\u003c/em\u003e (used when clinicians specify why a presentation does not fully meet ADHD criteria) and \u003cem\u003eUnspecified ADHD\u003c/em\u003e (used when clinicians either do not specify the reason or lack sufficient information).\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e5. Comorbidities and Differential Diagnoses\u003c/h2\u003e \u003cp\u003eIn DSM-III, ADHD was primarily linked to conduct disorder and learning disabilities. Over time, DSM-5 and DSM-5-TR expanded its comorbidity profile to include autism spectrum disorder (ASD), mood disorders, substance use disorders, and anxiety disorders, reflecting a broader and more complex clinical framework (see also Finding 3).\u003c/p\u003e \u003cp\u003eSimultaneously, the differential diagnosis section expanded considerably. DSM-III and DSM-III-R listed 5 and 6 conditions, respectively, including \u0026ldquo;\u003cem\u003eage-appropriate overactivity\u003c/em\u003e\u0026rdquo; as a diagnostic consideration. DSM-IV and DSM-IV-TR increased this to 10, while DSM-5 and DSM-5-TR further expanded it to 16 and 17. Notably, the updated lists excluded \u0026ldquo;\u003cem\u003eage-appropriate overactivity\u003c/em\u003e\u0026rdquo; and did not include \u0026ldquo;\u003cem\u003enormal variations\u003c/em\u003e,\u0026rdquo; a consideration found in the differential diagnosis of some other neurodevelopmental disorders.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e6. Contextual and Cultural Considerations\u003c/h2\u003e \u003cp\u003eFinally, despite the shift towards a neurodevelopmental framework (Finding 1), contemporary DSM editions have expanded their acknowledgment of environmental and contextual factors in shaping symptom expression. While earlier DSM editions made little to no reference to cultural considerations, DSM-5-TR explicitly recognizes racial and ethnic disparities, highlighting patterns of underdiagnosis in minority populations and the potential influence of clinician bias on diagnostic practices. Additionally, recent editions have incorporated environmental influences on symptom manifestation, acknowledging the roles of family dynamics, digital environments, and workplace settings in modulating the presentation and severity of ADHD symptoms.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study provides the first systematic, AI-assisted analysis of how ADHD has been described and classified in the DSM over the past four decades. By tracing changes across six DSM editions, the analysis identified six overarching trends:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eADHD has shifted from a primarily behavioral disorder to a neurodevelopmental condition, emphasizing biological and cognitive underpinnings.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eInitially conceptualized as a disorder of young boys, ADHD has evolved into a lifespan condition affecting individuals across genders. The reported prevalence in children increased from 3\u0026ndash;7.2%, while the male-to-female ratio declined from 10:1 to 2:1. Simultaneously, DSM-5 formally recognized adult ADHD (with a 2.5% prevalence), lowered the symptom threshold for older individuals, and extended the required age of onset from 7 to 12 years.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eThe definition of impairment has evolved, broadening beyond academic difficulties to include occupational, social, and emotional challenges. A range of negative consequences were introduced including, for example, incarceration, traffic accidents, overall mortality, and suicide. At the same time, the latest editions have also lowered the required threshold for impairment, allowing for the diagnosis of individuals with minor difficulties.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eDiagnostic criteria have become more flexible, moving from rigid subtypes to a classification that accommodates changing symptom presentations, including cases that do not fully meet the diagnostic criteria (categorized as Other Specified and Unspecified ADHD).\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eThe scope of ADHD\u0026rsquo;s comorbidities and differential diagnoses has expanded significantly, with DSM-5-TR listing 17 differential diagnoses\u0026mdash;excluding 'normal variations.'\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eWhile the DSM increasingly frames ADHD as a neurobiological disorder, later editions also acknowledge the role of cultural and environmental factors (e.g., digital environments) in shaping symptom expressions.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003eThese findings were derived from a dual-model AI pipeline\u0026mdash;based on GPT-4o and Claude 3.5 Sonnet\u0026mdash;and subjected to cross-model validation (Fig.\u0026nbsp;1). All findings were grounded in verbatim quotations from the DSM and detailed in four Supplementary Materials.\u003c/p\u003e \u003cp\u003eThe contributions of this study are threefold. First, it leveraged the advanced language capabilities of large AI models to systematically detect subtle textual shifts across multiple and complex editions of the DSM. Second, it offers a relatively objective and reproducible alternative to traditional qualitative approaches, which\u0026mdash;despite examining the same texts\u0026mdash;have produced conflicting interpretations of ADHD\u0026rsquo;s evolving description (e.g., \u003csup\u003e8,13,14,17\u003c/sup\u003e). Third, it provides a fully transparent account of its methods and findings (see the four Supplementary Materials), offering a comprehensive and replicable reference point for researchers, clinicians, and policymakers concerned with ADHD classification and its implications.\u003c/p\u003e \u003cp\u003eWe recommend that future discussions consider these findings in light of the DSM\u0026rsquo;s own definition of a mental disorder, which emphasizes \u0026ldquo;clinically significant disturbance\u0026rdquo; and cautions against diagnosing \u0026ldquo;an expectable or culturally approved response to a common stressor\u0026rdquo; or \u0026ldquo;conflicts that are primarily between the individual and society.\u0026rdquo; In this study, we intentionally refrained from taking a position on this issue. Instead, we aimed to present the analysis with minimal human interpretation, allowing readers to evaluate whether the observed conceptual shifts reflect improved identification of previously overlooked cases or a broadening of diagnostic boundaries that may carry a risk of overdiagnosis.\u003csup\u003e21,22\u003c/sup\u003e In either case, these shifts underscore the importance of continued reflection on ADHD\u0026rsquo;s classification and its implications for clinical practice and public health policy.\u003c/p\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eWhile this study provides a well-validated analysis of ADHD-related texts, several limitations should be acknowledged. First, the study focused exclusively on DSM texts, which, while serving as the most authoritative psychiatric classification system, represent only one aspect of how ADHD is conceptualized. The evolution of ADHD is influenced by broader forces, including empirical research, clinical practice, educational policies, and public discourse. Second, while this study critically examined shifts in the DSM\u0026rsquo;s portrayal of ADHD, it did not empirically assess their impact on real-world clinical practice, such as changes in prevalence rates, prescribing patterns, or patient outcomes. Future research should integrate real-world data to examine how DSM modifications translate into clinical decision-making.\u003c/p\u003e \u003cp\u003eThird, despite employing state-of-the-art AI language models (GPT-4o and Claude 3.5 Sonnet), their proprietary nature poses challenges for reproducibility, as ongoing model updates may alter their outputs. Fourth, AI models are susceptible to biases based on their training data, including frequency effects that may skew their focus toward more recent DSM editions. Finally, AI models function as black-box systems, meaning their internal decision-making processes remain opaque. Although cross-validation measures were implemented and models were required to support all findings with verbatim DSM quotations to mitigate biases and hallucinations, human expertise remains essential for interpreting findings and contextualizing them within the broader scientific and clinical landscape.\u003c/p\u003e \u003cp\u003eFuture studies may benefit from utilizing open-weight AI models, which make their parameters publicly available, improving transparency, reproducibility, and researcher control over the analysis. Additionally, integrating hybrid methodologies that combine AI analysis with traditional qualitative research may further strengthen findings by providing a more nuanced understanding of ADHD\u0026rsquo;s evolving conceptualization. Despite these limitations, this study establishes a structured, transparent, and replicable foundation for future research, offering valuable insights for researchers, clinicians, and policymakers seeking to better understand the shifting diagnostic landscape of ADHD.\u003c/p\u003e "},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics, Consent to Participate, and Consent to Publish Declarations:\u0026nbsp;\u003c/strong\u003eNot applicable.\u003cbr\u003eThis study is \u003cu\u003enot\u003c/u\u003e a clinical trial. It is based on a textual analysis of the \u003cem\u003eDiagnostic and Statistical Manual of Mental Disorders\u003c/em\u003e (DSM) and does not involve human participants or require ethical approval.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e The authors received no funding for this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of Interest:\u003c/strong\u003e The authors declare no conflicts of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of Data:\u003c/strong\u003e All materials generated and analyzed in this study are available in the four Supplementary Materials accompanying this article.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eAll authors contributed meaningfully to the development of this article. Y.O. conceptualized the study, reviewed all DSM texts, conducted the AI-assisted analysis, and drafted the manuscript. Y.S.R. and R.T. conducted a comprehensive review of the existing literature, thoroughly examined the study materials, actively contributed to internal discussions, and provided critical feedback and revisions to the manuscript. All authors approved the final version. The authors would also like to thank Refael Reuveni from Ariel University for his invaluable assistance in laying the groundwork for the current study, particularly through his work on preliminary trials with the AI models.\u003c/p\u003e"},{"header":"References","content":"\u003cp\u003eEpstein JN, Loren REA. Changes in the definition of ADHD in DSM-5: subtle but important. 2013; \u003cstrong\u003e3\u003c/strong\u003e: 455.\u003c/p\u003e\n\u003cp\u003eAnthropic. Claude 3.5 Sonnet. 2024.\u003c/p\u003e\n\u003cp\u003eCrowe M. Constructing normality: a discourse analysis of the DSM‐IV. 2000; \u003cstrong\u003e7\u003c/strong\u003e: 69\u0026ndash;77.\u003c/p\u003e\n\u003cp\u003eVoort JLV, He J-P, Jameson ND, Merikangas KR. Impact of the DSM-5 attention-deficit/hyperactivity disorder age-of-onset criterion in the US adolescent population. 2014; \u003cstrong\u003e53\u003c/strong\u003e: 736\u0026ndash;744.\u003c/p\u003e\n\u003cp\u003eSibley MH, Kuriyan AB. DSM-5 changes enhance parent identification of symptoms in adolescents with ADHD. 2016; \u003cstrong\u003e242\u003c/strong\u003e: 180\u0026ndash;185.\u003c/p\u003e\n\u003cp\u003eFreedman JE, Honkasilta JM. Dictating the boundaries of ab/normality: a critical discourse analysis of the diagnostic criteria for attention deficit hyperactivity disorder and hyperkinetic disorder. 2017; \u003cstrong\u003e32\u003c/strong\u003e: 565\u0026ndash;588.\u003c/p\u003e\n\u003cp\u003eAPA. \u003cem\u003eDiagnostic and Statistical Manual of Mental Disorders - Third Edition (DSM-III)\u003c/em\u003e. American Psychiatric Association (APA), 1980.\u003c/p\u003e\n\u003cp\u003eWilson M. DSM-III and the transformation of American psychiatry: A history. 1993; \u003cstrong\u003e150\u003c/strong\u003e: 399.\u003c/p\u003e\n\u003cp\u003eLefler EK, Stevens AE, Garner AM, Serrano JW, Canu WH, Hartung CM. Changes in college student endorsement of ADHD symptoms across DSM edition. 2020; \u003cstrong\u003e42\u003c/strong\u003e: 488\u0026ndash;499.\u003c/p\u003e\n\u003cp\u003eAPA. Diagnostic and Statistical Manual of Mental Disorders, 5th edition, Text Revision (DSM-5-TR). 2022.\u003c/p\u003e\n\u003cp\u003eRigler T, Manor I, Kalansky A, Shorer Z, Noyman I, Sadaka Y. New DSM-5 criteria for ADHD\u0026mdash;Does it matter? 2016; \u003cstrong\u003e68\u003c/strong\u003e: 56\u0026ndash;59.\u003c/p\u003e\n\u003cp\u003eSklepn\u0026iacute;kov\u0026aacute; J, Slez\u0026aacute;čkov\u0026aacute; A. Evolution of the ADHD concept and its relation to the hyperkinetic disorders: Narrative review. 2022; \u003cstrong\u003e66\u003c/strong\u003e: 255\u0026ndash;271.\u003c/p\u003e\n\u003cp\u003eDoernberg E, Hollander E. Neurodevelopmental disorders (asd and adhd): Dsm-5, icd-10, and icd-11. 2016; \u003cstrong\u003e21\u003c/strong\u003e: 295\u0026ndash;299.\u003c/p\u003e\n\u003cp\u003eWHO. \u003cem\u003eWorld mental health report: Transforming mental health for all\u003c/em\u003e. World Health Organization. Last retrieved on December 17, 2024 from: https://iris.who.int/bitstream/handle/10665/356119/9789240049338-eng.pdf?sequence=1, 2022.\u003c/p\u003e\n\u003cp\u003eDanielson ML, Bitsko RH, Ghandour RM, Holbrook JR, Kogan MD, Blumberg SJ. Prevalence of Parent-Reported ADHD Diagnosis and Associated Treatment Among U.S. Children and Adolescents, 2016. 2018; \u003cstrong\u003e47\u003c/strong\u003e: 199\u0026ndash;212.\u003c/p\u003e\n\u003cp\u003eAchiam J, Adler S, Agarwal S, Ahmad L, Akkaya I, Aleman FL \u003cem\u003eet al.\u003c/em\u003e Gpt-4 technical report. 2023.\u003c/p\u003e\n\u003cp\u003eKazda L, Bell K, Thomas R, Hardiman L, Heath I, Barratt A. Attention deficit/hyperactivity disorder (ADHD) in children: more focus on care and support, less on diagnosis. 2024; \u003cstrong\u003e384\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eVisser SN, Danielson ML, Bitsko RH, Holbrook JR, Kogan MD, Ghandour RM \u003cem\u003eet al.\u003c/em\u003e Trends in the parent-report of health care provider-diagnosed and medicated attention-deficit/hyperactivity disorder: United States, 2003\u0026ndash;2011. 2014; \u003cstrong\u003e53\u003c/strong\u003e: 34\u0026ndash;46.\u003c/p\u003e\n\u003cp\u003eMallett CA, Natarajan A, Hoy J. Attention deficit/hyperactivity disorder: A DSM timeline review. 2014; \u003cstrong\u003e43\u003c/strong\u003e: 36\u0026ndash;60.\u003c/p\u003e\n\u003cp\u003eHyman SE. The diagnosis of mental disorders: the problem of reification. 2010; \u003cstrong\u003e6\u003c/strong\u003e: 155\u0026ndash;179.\u003c/p\u003e\n\u003cp\u003eDeacon BJ. The biomedical model of mental disorder: a critical analysis of its validity, utility, and effects on psychotherapy research. 2013; \u003cstrong\u003e33\u003c/strong\u003e: 846\u0026ndash;861.\u003c/p\u003e\n\u003cp\u003eKazda L, Bell K, Thomas R, McGeechan K, Sims R, Barratt A. Overdiagnosis of Attention-Deficit/Hyperactivity Disorder in Children and Adolescents: A Systematic Scoping Review. 2021; \u003cstrong\u003e4\u003c/strong\u003e: e215335.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"ADHD, DSM, psychiatric classification, AI-assisted text analysis","lastPublishedDoi":"10.21203/rs.3.rs-6405218/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6405218/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eConsidering the central role of the \u003cem\u003eDiagnostic and Statistical Manual of Mental Disorders\u003c/em\u003e (DSM) in psychiatric classification, multiple studies have examined how it describes \u003cem\u003eAttention-Deficit/Hyperactivity Disorder\u003c/em\u003e (ADHD)\u0026mdash;one of the most common psychiatric diagnoses. However, despite analyzing the same DSM texts, these studies yielded conflicting conclusions, likely influenced by the subjectivity of qualitative research and the challenge of systematically tracking subtle changes in large textual corpora. This study addresses these limitations by providing the first systematic, Artificial Intelligence (AI)-assisted analysis of all ADHD-related texts across six DSM editions (\u003cem\u003eDSM-III\u003c/em\u003e to \u003cem\u003eDSM-5-TR\u003c/em\u003e).\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eThe analysis employed two AI models (\u003cem\u003eGPT-4o\u003c/em\u003e and \u003cem\u003eClaude 3.5 Sonnet\u003c/em\u003e) and followed five steps: (A) preliminary human review, (B) AI-assisted comparison, (C) refinement through self-prompting to detect subtle linguistic changes, (D) thematic synthesis by each model, and (E) cross-model validation. Strict adherence to DSM texts ensured that all findings were grounded in verifiable evidence. To reduce interpretive bias, human interpretations were intentionally minimized.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe complete analysis is available in the Supplementary Materials. Overall, the findings revealed six overarching shifts toward: (1) a neurodevelopmental framework, (2) a lifespan condition across genders, (3) a broader concept of impairment, (4) greater diagnostic flexibility, (5) expanded comorbidities and differential diagnoses, and (6) cultural and contextual influences.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThese trends, together with the full study materials, systematically map how ADHD has been described and classified in the DSM over four decades, offering a structured and transparent foundation for future discussions.\u003c/p\u003e","manuscriptTitle":"Four Decades of ADHD: A Systematic AI-assisted Analysis of Conceptual Shifts Across Six DSM Editions","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-12 16:50:51","doi":"10.21203/rs.3.rs-6405218/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"10f80375-77a1-4753-8424-9dcb0c6eaf40","owner":[],"postedDate":"May 12th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-12-18T08:40:25+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-12 16:50:51","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6405218","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6405218","identity":"rs-6405218","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}