Gender dysphoria refers to the psychological distress that arises when a person’s gender identity does not align with their sex observed at birth. For decades, the clinical literature described two broad developmental pathways. The early-onset pathway typically begins in early childhood, is marked by pervasive gender-nonconforming behavior, and is strongly associated with same-sex attraction upon sexual maturity. The late-onset pathway, documented extensively in males, emerges during or after puberty and is historically associated with an absence of childhood gender nonconformity and with attraction to the opposite sex. What has changed dramatically over the past fifteen years is the demographic composition of gender clinics: referrals are now dominated by adolescent females, many of whom report no childhood history of gender-nonconforming behavior and many of whom carry significant comorbid psychiatric diagnoses.

The research team, led by Hollie Hammond of Western Sydney University together with colleagues including James S. Morandini, recruited participants through social media platforms such as Facebook and Reddit, as well as word of mouth, rather than through gender clinics. This non-clinical sampling strategy was a deliberate methodological choice. Most previous evidence on the subject came either from clinic-referred samples or from parent-report studies, the latter of which have drawn sustained criticism for relying exclusively on parental accounts and for allegedly conflating transgender identity with a psychiatric diagnosis. By surveying young people directly, the researchers aimed to capture the experiences of a population that rarely appears in clinical datasets. Participants completed an anonymous online survey lasting an average of 37 minutes, and the team applied quality controls including bot detection, minimum completion times, and manual review of responses.

To classify onset type, participants answered eight symptom items adapted from the DSM-5-TR diagnostic criteria for childhood gender dysphoria, recalling their experiences between the ages of 3 and 11. Those who reported meeting at least six symptoms for a minimum of six months—the DSM-5-TR childhood diagnostic threshold—were placed in the early-onset group; those below the threshold were classified as late-onset. The results were lopsided: 29 participants (21 percent) fell into the early-onset category while 109 (79 percent) were late-onset. The late-onset group reported an average of fewer than two childhood symptoms, with the lowest endorsement on internal items such as dislike of one’s sexual anatomy and desire for the physical attributes of the other sex. The early-onset group, by contrast, most often recalled preferences related to clothing and cross-sex roles in make-believe play.

The statistical contrasts between the groups were substantial. Early-onset participants reported significantly more childhood symptoms, with an effect size of Cohen’s d = 3.05—one of the largest reported in this literature. They also adopted a non-cisgender identity earlier, at an average age of 12.68 years compared with 15.01 years for the late-onset group, and came out to others roughly two years sooner. Yet even the early-onset group diverged from classic clinical cohort studies, in which children typically presented to clinics around age six and socially transitioned by age eight. The authors suggest their early-onset subgroup may better be described as a pseudo-early-onset, or late childhood and early adolescent onset, pathway—one that does not replicate the developmental signature of the historical cohorts on which much of the evidence base for early gender-affirming care was built.

Gender identity itself differentiated the groups. Late-onset participants were significantly more likely to identify as non-binary, whereas early-onset participants more frequently endorsed binary identities such as man or male. The authors report that this association between onset type and gender identity has not been documented before, and they flag it as clinically significant given emerging evidence that non-binary identity is associated with poorer mental health outcomes in some meta-analyses and with higher odds of discontinuing gender-affirming medical treatment in others. Transmasculine was the most commonly endorsed identity in both groups, chosen by more than half of participants in each.

Sexual orientation told a different story from the male-dominated literature. Using an adapted Kinsey scale, over half of participants in both groups—51.7 percent of the early-onset and 57.8 percent of the late-onset—reported at least partial same-sex attraction, and only about a fifth to a quarter described themselves as exclusively attracted to males. This represents a dramatic overrepresentation of same-sex attraction relative to general population estimates, in which roughly nine in ten women are heterosexual. Notably, however, orientation did not differ between onset groups, challenging the traditional male-pattern association in which the early-onset pathway is linked to homosexuality and the late-onset pathway to heterosexuality. The authors suggest that typologies developed for males with gender dysphoria may simply not transfer to females.

The clinical picture also split along onset lines. Early-onset participants reported roughly twice as many diagnosed comorbid psychiatric conditions as their late-onset counterparts, with significantly higher rates of attention-deficit/hyperactivity disorder, generalized anxiety disorder, and panic disorder. They also scored significantly higher on measures of dysphoria severity and anxiety, though the two groups did not differ on depression, stress, or well-being—both scoring in the moderate to extremely severe range across distress measures. The late-onset group was far from symptom-free: 56.9 percent reported at least one comorbid diagnosis, a figure that overlaps with the 62.5 percent reported in the controversial 2018 parent-report study by Lisa Littman that first proposed the rapid-onset gender dysphoria hypothesis. The authors note that this self-reported overlap offers some indirect support for aspects of that hypothesis, while stopping short of endorsing it, given that their data cannot establish whether distress preceded dysphoria or followed it.

Peer relationships offered a notable null finding. Participants reported having approximately two to three transgender or gender-diverse friends before realizing their own dysphoria, with no significant differences between onset groups in the number of friends, online friends, or in-person friends. If the late-onset group were primarily composed of the hypothesized sociogenic cohort driven by peer influence, the authors reason, one would expect them to report more transgender friends prior to onset than the early-onset group. That pattern did not emerge, though the authors caution that the sociogenic hypothesis remains deeply contested and that cross-sectional data cannot resolve questions of causality or directionality.

The study’s limitations are considerable and the authors are candid about them. The sample was a small, uneven convenience sample reliant on retrospective self-report, vulnerable to recall bias—particularly given evidence that distressed and neurotic individuals tend to overestimate negative childhood affect, which could inflate early-onset classification among the most distressed participants. A longitudinal design tracking children from early development into adulthood, ideally with corroboration from parents, teachers, or clinicians, would be needed to firmly establish the temporal relationships the present data can only suggest. Still, the authors argue the findings carry practical implications: clinicians should take careful histories of symptom onset, since onset type may signal different comorbidity profiles, different dysphoria severity, and even different identity trajectories. Most adolescent females experiencing gender dysphoria, the study concludes, report little childhood gender nonconformity, and the assumption that historical clinical patterns observed mostly in males apply to today’s predominantly female adolescent population is one that future research can no longer afford to make.

Subject of Research: Early- versus late-onset gender dysphoria in adolescent and young adult birth-assigned females

Article Title: Demographic and Clinical Features of Early-Onset versus Late-Onset Gender Dysphoria Among Adolescent and Young Adult Birth-Assigned Females

Article References: Hammond, H., Walter, E., Daly, M., Morandini, J. S., & Smith, E. (2026). Demographic and Clinical Features of Early-Onset versus Late-Onset Gender Dysphoria Among Adolescent and Young Adult Birth-Assigned Females. Archives of Sexual Behavior. https://doi.org/10.1007/s10508-026-03571-6

Image Credits: AI Generated

DOI: 10.1007/s10508-026-03571-6

Keywords: gender dysphoria, adolescents, late-onset, early-onset, non-binary, transgender, mental health, DSM-5-TR, sexual orientation, comorbidity, Archives of Sexual Behavior, birth-assigned females

The research, led by Mahrukh Shakoor and Tamim Ahmed Khan of the Department of Software Engineering at Bahria University in Islamabad, together with Hina Ghafoor of the university’s Department of Professional Psychology, describes a system that integrates reflective listening, large language models, knowledge graph reasoning, and clinical guidelines drawn from the DSM-5-TR, the current edition of the Diagnostic and Statistical Manual of Mental Disorders. The goal is not to replace clinicians but to give them a structured, interpretable tool for the initial assessment process, where human resources are scarce and early signals are easy to miss.

At the heart of the framework is a two-stage screening workflow designed to mimic, in a limited way, the flow of a real clinical intake. In the first stage, a reflective-listening conversational agent engages the user in empathetic dialogue. Reflective listening is a counseling technique in which the listener mirrors and restates what the speaker has said, validating their experience and encouraging deeper disclosure. By building this technique into the language model, the researchers aimed to create an agent that does not simply interrogate the user with checklist questions but instead draws out clinically relevant symptoms through conversation that feels supportive and engaging.

During this first-stage dialogue, the system performs contextual reasoning over what the user says, extracting symptom features and representing them within a knowledge graph, a structured network of entities and relationships that allows the AI to connect scattered conversational cues into a coherent clinical picture. Rather than treating each mention of sleeplessness, worry, or low mood as an isolated data point, the knowledge graph links symptoms to one another and to diagnostic categories, enabling an initial mental health risk screening that can be traced and explained. This emphasis on explainability is central to the framework: clinicians can see how the system arrived at its conclusions rather than receiving an opaque risk score from a black-box model.

The second stage brings quantitative rigor to the conversation. The framework incorporates standardized psychometric instruments, including the Penn State Worry Questionnaire, a widely used measure of pathological worry; the Beck Depression Inventory-II, one of the most established tools for assessing the severity of depressive symptoms; and the McLean Screening Instrument for Borderline Personality Disorder, a validated screening tool for that condition. By embedding these instruments into the workflow, the system grounds its conversational impressions in validated, disorder-specific measures, producing a quantitative assessment that supports more fine-grained risk estimation for particular conditions.

To train and evaluate the system, the team constructed a structured dataset that is notable for its clinical grounding. It comprises AI-generated reflective conversations, screening labels informed by DSM-5-TR criteria, extracted symptom features, knowledge graph representations, and, critically, annotations reviewed by clinicians. This dataset was then used to fine-tune the proposed language model, aligning its behavior not just with general linguistic patterns but with the specific vocabulary and logic of mental health assessment. The clinician-reviewed component of the dataset provides a human benchmark against which the AI’s outputs can be measured.

Experimental results reported in the study show encouraging agreement between the framework’s AI-assisted screening outcomes and assessments reviewed by clinicians. The authors highlight that integrating reflective listening, psychometric evaluation, knowledge graph reasoning, and generative AI improved not only screening performance but also interpretability and user engagement, two factors that often pull in opposite directions in automated health tools. Systems that are highly engaging are frequently opaque, while systems that are transparent can feel clinical and cold. The reflective-listening approach appears to soften that trade-off, keeping users in conversation while the underlying reasoning remains inspectable.

The ethical scaffolding around the study is unusually detailed for work in this space. The research was reviewed and approved by the Ethics Review Committee of Bahria School of Professional Psychology at Bahria University, with approval dated 26 January 2026, and all procedures were conducted in accordance with the Declaration of Helsinki. Participation was voluntary and informed consent was obtained from all participants, with data anonymized before analysis and personally identifiable information removed. Importantly, the framework was evaluated solely as an AI-assisted screening and decision-support prototype and was not used for clinical diagnosis or treatment. Participants identified as potentially high-risk, including those expressing suicidal ideation, self-harm, psychotic symptoms, or abuse-related concerns, were advised to discontinue the assessment and seek immediate evaluation from qualified mental health professionals or emergency services, a safeguard that acknowledges the real dangers of deploying conversational AI in sensitive domains.

The significance of the work lies in how it addresses the structural bottlenecks of mental healthcare. Screening is the first and often most neglected step in the care pathway: a person must be identified as needing help before any treatment can begin. In many parts of the world, the ratio of mental health professionals to population is so low that systematic early screening is simply impossible at scale. A conversational agent that can conduct an empathetic intake, administer validated psychometric measures, and hand clinicians a structured, explainable summary of risk could extend the reach of overstretched services, flagging individuals who would otherwise remain invisible until crisis point.

The researchers are careful to position the framework as a support tool rather than an autonomous diagnostician. The system’s outputs are intended to assist healthcare professionals during initial assessment, with final clinical judgment remaining in human hands. That framing, combined with the framework’s explainability features and its grounding in DSM-5-TR criteria and validated instruments, reflects a growing consensus in medical AI research: the most promising systems are not those that attempt to replace clinical expertise, but those that structure information, surface early signals, and make their reasoning transparent enough for professionals to trust, verify, and act upon. As generative AI continues to mature, studies like this one suggest a future in which the first conversation a struggling person has about their mental health, often the hardest one to start, might be made a little easier, and a great deal more likely to lead to help.

Subject of Research: An explainable AI framework for early mental health screening and clinical decision support using reflective listening and generative AI

Article Title: An explainable AI framework for early mental health screening and clinical decision support using reflective listening and generative AI

Article References: Shakoor, M., Khan, T. A., & Ghafoor, H. (2026). An explainable AI framework for early mental health screening and clinical decision support using reflective listening and generative AI. Discover Psychology. https://doi.org/10.1007/s44202-026-00877-3

Image Credits: AI Generated

DOI: 10.1007/s44202-026-00877-3

Keywords: mental health screening, explainable AI, reflective listening, generative AI, large language models, knowledge graph reasoning, DSM-5-TR, clinical decision support, psychometric assessment, psychological disorders, early identification, conversational agents

The study cohort comprised 4,250 adults diagnosed under DSM-5-TR criteria, including 1,515 with bipolar disorder—839 with bipolar I and 676 with bipolar II—and 2,735 with major depressive disorder. Patients had been ill for an average of 14.1 years and were followed prospectively and systematically for 4.73 years, giving the investigators an unusually rich longitudinal window into the relationship between substance misuse and the course of mood illness. Across the full sample, the overall prevalence of lifetime substance abuse averaged 11.5 percent, but that single figure conceals dramatic differences between diagnostic groups and between the sexes.

The headline comparison is unambiguous: substance abuse was 4.36 times more prevalent among patients with bipolar disorder than among those with major depressive disorder, affecting 22.8 percent of the bipolar group versus just 5.23 percent of the depressed group. Within the bipolar spectrum, the gradient continued. Bipolar I patients showed a lifetime substance abuse rate of 26.6 percent, roughly 1.45 times the 18.4 percent observed in bipolar II patients. This pattern reinforces a growing body of evidence that risk of comorbid substance misuse scales with the severity and mania load of the mood syndrome, rather than being a uniform feature of mood disorders in general.

Gender emerged as one of the most powerful risk factors in the entire dataset. Among the 1,583 men studied, 20.2 percent had a lifetime substance abuse history, compared with only 6.37 percent of the 2,667 women—a 3.17-fold difference that dwarfs the gender gaps typically reported in general population surveys. The authors suggest this pronounced male excess may interact with the biological and social vulnerabilities specific to mood disorders, compounding a baseline gender difference in substance abuse risk that exists even in people without psychiatric illness.

Equally striking was the finding on polyabuse, defined as misuse of multiple substances. Polyabuse averaged 6.35 percent across the cohort overall, but it was 5.60 times more common in bipolar disorder, where 13.5 percent of patients reported it, than in major depression, where the rate was 2.41 percent. When the researchers ranked specific abused substances, alcohol dominated at 20.9 percent, followed by polyabuse at 6.35 percent, cannabis at 3.29 percent, with opioids and stimulants each at 0.31 percent and benzodiazepines at a negligible 0.02 percent. For every category, rates were higher in bipolar disorder than in major depressive disorder, with alcohol and multi-substance misuse driving the overwhelming majority of the comorbidity burden.

Beyond diagnosis and sex, the team mapped a detailed clinical and psychosocial signature of patients prone to substance abuse. Those with lifetime substance misuse were more likely to be male, less likely to be married, more likely to be divorced, had fewer children, and were more often unemployed. They smoked more, experienced an earlier onset of mood illness, and carried more psychiatric comorbidity—although notably less somatic or medical comorbidity—than their peers without substance problems. Most sobering of all, substance-abusing patients showed markedly more suicidal behavior, a finding that echoes prior meta-analytic work linking co-occurring bipolar and substance use disorders to dramatically elevated suicide attempt risk.

The study also probed temperament, using structured affective temperament ratings, and found that an irritable temperament was associated with lifetime substance abuse. Intriguingly, however, substance abuse was not linked to greater affective morbidity—that is, substance-abusing patients did not simply accumulate more mood episodes. This dissociation is scientifically important because it argues against the simplistic interpretation that patients abuse substances merely in proportion to how often their mood illness relapses. Instead, the data point toward trait-like vulnerabilities, including temperament, impulsivity-related traits, and early-onset illness, as the more proximal drivers of substance misuse in this population.

To isolate the factors that independently predict substance abuse, the researchers applied multivariable regression analysis. The ranked order of independently associated factors was: smoking, male gender, younger age at illness onset, a bipolar diagnosis, never having married, an irritable temperament, and having fewer children. Cigarette smoking topping the list is particularly noteworthy from a mechanistic standpoint; smoking may serve as both a marker of broader impulsivity and reward-system dysregulation and as a gateway behavior that facilitates progression to other substance misuse. The prominence of early onset likewise suggests that a neurodevelopmentally earlier form of mood illness carries heightened vulnerability to addictive comorbidity, consistent with shared genetic liability hypotheses supported by recent twin, family, and population-based studies.

From a clinical standpoint, the implications are concrete. Because alcohol accounts for the large majority of abused substances in this population, routine screening for alcohol misuse should be considered standard practice in mood disorder clinics, with particularly vigilant surveillance of male patients, smokers, and those with early-onset or bipolar I illness. The strong association with suicidal behavior argues that dual-diagnosis patients warrant intensified safety planning and may benefit from integrated treatment models that address mood and substance problems simultaneously rather than sequentially. Previous research has shown, for example, that comorbid substance use disorder can impair recovery from depression during standard antidepressant treatment and reduce responsiveness to mood stabilizers in bipolar patients, underscoring the cost of ignoring the comorbidity.

The study’s scale and prospective design distinguish it from much of the existing literature, which has relied heavily on cross-sectional surveys, meta-analyses of heterogeneous samples, or national registries that lack detailed clinical characterization. By directly comparing bipolar I, bipolar II, and major depressive disorder within a single, systematically assessed cohort followed for years, the McLean-Harvard team has provided what may be the most granular stratification of substance abuse risk across the major mood disorders to date. As substance-related disorders continue to impose an enormous global disease burden, and as rates of cannabis and other drug use climb among young adults, identifying which mood disorder patients are at highest risk—and intervening early—could pay substantial dividends in reducing disability, hospitalization, and suicide in this vulnerable population.

The diagnostic framework used in the study deserves some attention. By applying DSM-5-TR criteria uniformly across all 4,250 participants, the investigators reduced the diagnostic heterogeneity that has plagued earlier comparisons of substance misuse across mood disorders. This matters because prevalence estimates for comorbid substance use disorders have varied enormously in prior research—sometimes ranging from under 10 percent to over 50 percent in bipolar samples—depending largely on whether studies relied on self-report, registry data, or structured diagnostic interviews, and on whether abuse and dependence were distinguished from mere use.

The prospective element of the design also strengthens causal interpretation in one specific respect. Because participants were followed systematically for nearly five years after baseline characterization, the researchers could examine whether substance abuse predicted subsequent affective morbidity. The finding that it did not—that substance-abusing patients did not experience more mood episodes over time—challenges the widespread self-medication hypothesis in its simplest form. That hypothesis holds that patients turn to alcohol or drugs to dampen painful mood symptoms, implying that heavier substance use should track with more frequent or severe episodes. While self-medication may still operate in individual cases, and prior work has shown that drinking to relieve mood symptoms does predict later alcohol dependence, the present data suggest that in this cohort, trait vulnerabilities rather than episode frequency best explain who develops substance problems.

The near-absence of benzodiazepine misuse, at just 0.02 percent, is a notable detail given clinical concerns about tranquilizer dependence in psychiatric populations. It may reflect the specific composition and treatment setting of the cohort, or genuine patterns of preference for alcohol and cannabis among mood disorder patients, but it contrasts with population surveys in the United States that have documented substantial rates of prescription benzodiazepine misuse. Similarly, the low rates of opioid and stimulant misuse may partly reflect the era and region in which participants were recruited, reminding readers that substance availability and local drug markets shape comorbidity patterns as much as underlying psychopathology does.

Finally, the temperamental finding invites further research. Irritable temperament, measured with validated self-report instruments derived from the Akiskal temperamental framework, has previously been linked to impulsivity and interpersonal conflict, both plausible pathways into substance misuse. If replicated, temperament assessment could become a low-cost screening tool, allowing clinicians to flag newly diagnosed mood disorder patients—especially young men with early-onset illness and a smoking history—for early preventive counseling before problematic use takes hold.

Subject of Research: Lifetime substance abuse prevalence and clinical correlates in bipolar I, bipolar II, and major depressive disorder patients

Article Title: Lifetime Substance Use Disorder in 4250 Bipolar and Major Depressive Disorder Patients

Article References: Miola, A., Tondo, L., & Baldessarini, R. J. (2026). Lifetime Substance Use Disorder in 4250 Bipolar and Major Depressive Disorder Patients. International Journal of Mental Health and Addiction. https://doi.org/10.1007/s11469-026-01718-z

Image Credits: AI Generated

DOI: 10.1007/s11469-026-01718-z

Keywords: bipolar disorder, major depressive disorder, substance abuse, alcohol abuse, polyabuse, comorbidity, suicidal behavior, smoking, affective temperament, dual diagnosis, psychiatry, DSM-5-TR