The research, led by Mimi Liu and Zifu Shi of Hunan Normal University together with colleagues at Huanghuai University, the Neusoft Institute of Guangdong, and Hunan University of Chinese Medicine, tracked 1,243 Chinese college students across two measurement waves separated by six months. Participants, who averaged 18.54 years of age and were 71.3 percent female, were drawn from six universities. Rather than treating mind wandering as a single undifferentiated tendency, the team measured two distinct flavors of it: spontaneous mind wandering, in which thoughts drift off involuntarily, and deliberate mind wandering, in which a person purposefully redirects attention toward off-task thoughts.

That distinction matters. Spontaneous mind wandering is often framed as a failure of attentional control, an unbidden escape of cognition from the task at hand. Deliberate mind wandering, by contrast, looks more like an intentional choice, a mental vacation the thinker chooses to take. Prior work had mostly examined short-term dynamics between boredom and these tendencies, using designs that could not establish which construct predicted the other over meaningful stretches of time. By adopting a two-wave cross-lagged design, the researchers could ask a sharper question: does being prone to boredom today make you more likely to drift tomorrow, and vice versa, six months down the line?

The answer, for spontaneous mind wandering, is a resounding yes in both directions. After controlling for covariates, students who reported higher boredom proneness at the first wave showed significantly higher spontaneous mind wandering at the second wave, and students whose minds wandered spontaneously at the first wave were significantly more boredom prone six months later. This bidirectional, reciprocal pattern is the statistical signature of a self-reinforcing feedback loop. It dovetails with theoretical models such as the boredom feedback model and the MAC model of boredom and cognitive engagement, which portray boredom as a state in which attention is neither fully engaged nor meaningfully deployed, leaving cognition untethered to wander off in search of stimulation.

Deliberate mind wandering told a more asymmetric story. Boredom proneness at the first wave still predicted more deliberate mind wandering at the second wave, suggesting that people who chronically experience boredom come to treat off-task thinking as a deliberate coping strategy, an exploratory response to an unsatisfying present. But the reverse path was not significant: students who deliberately daydreamed at the first wave did not become more boredom prone by the second. In other words, boredom appears to push people toward intentional mental escape, but choosing to escape does not, by itself, breed boredom. The door swings one way only.

The most novel contribution of the study lies in its treatment of trait mindfulness, the stable dispositional tendency to attend to present-moment experience with awareness. Using cross-lagged mediation analyses, the researchers found that boredom proneness at the first wave was indirectly associated with both spontaneous and deliberate mind wandering at the second wave through trait mindfulness measured at the second wave. The same held in the other direction: higher spontaneous and deliberate mind wandering at baseline were indirectly linked to later boredom proneness through reduced mindfulness. Mindfulness, in effect, acted as a conduit carrying the influence of each construct forward in time.

This mediating role makes theoretical sense. Mindfulness is conceptualized as an attentional and awareness skill, and individuals high in trait mindfulness are better at noticing when attention has drifted and at sustaining contact with the present. A bored person who lacks this skill may have few resources to re-engage with the moment, letting attention slip into spontaneous drift or strategic daydreaming. Conversely, a habitual mind wanderer who never cultivates present-moment awareness may find ordinary experience increasingly dull, deepening boredom proneness. Prior research has linked mindfulness to lower boredom in workplace settings, to reduced media multitasking-related wandering, and to improved metacognition, but this study is among the first to position mindfulness longitudinally between the two constructs over a six-month horizon.

The findings carry practical weight, particularly for educational settings. Boredom proneness among Chinese college students has been rising across birth cohorts according to earlier meta-analytic work, and boredom is reliably associated with poorer academic outcomes, lower satisfaction, and even counterproductive behaviors. If boredom and mind wandering form a reciprocal loop that erodes mindfulness over time, interventions that bolster trait mindfulness, such as mindfulness meditation training, could plausibly interrupt the cycle from either end. The authors’ results are consistent with existing reviews suggesting that mindfulness practice reduces mind wandering, though a longitudinal observational design cannot itself establish that training would work here.

Caveats are worth keeping in view. The sample consisted of young adults at six Chinese universities, and the study relied on self-report scales, including short boredom proneness and Chinese-validated deliberate and spontaneous mind wandering measures, which leaves open the possibility of shared method bias that the authors took steps to address. Two-wave designs also cannot fully disentangle within-person change from stable between-person differences, a limitation that methodologists have raised about cross-lagged panel models generally. Six months is long enough to demonstrate durable associations but still too coarse to reveal the moment-to-moment dynamics by which boredom sparks drift and drift curdles into boredom.

Even so, the study adds an important longitudinal plank to a growing body of work portraying boredom not as a trivial annoyance but as a meaningful state of attentional disengagement with downstream cognitive costs. The picture that emerges is of a feedback system: a boredom-prone mind drifts, a drifting mind grows bored, and diminishing mindfulness lubricates both transitions. The hopeful corollary is that the system has an identified lever. If present-moment awareness can be cultivated, the loop that ties boredom to the wandering mind may be a cycle that can, with practice, be broken.

Subject of Research: Longitudinal associations between boredom proneness, spontaneous and deliberate mind wandering, and trait mindfulness in college students

Article Title: The longitudinal associations between boredom proneness and mind wandering: the role of trait mindfulness

Article References: Liu, M., Chen, H., Xie, J., Xie, J., Ma, X., & Shi, Z. (2026). The longitudinal associations between boredom proneness and mind wandering: the role of trait mindfulness. Current Psychology, 45(17), Article 1471. https://doi.org/10.1007/s12144-026-10032-9

Image Credits: AI Generated

DOI: 10.1007/s12144-026-10032-9

Keywords: boredom proneness, mind wandering, spontaneous mind wandering, deliberate mind wandering, trait mindfulness, longitudinal study, cross-lagged panel model, attention, college students, Current Psychology, meditation, cognitive engagement

The study’s central innovation is a four-part classification framework that separates adolescent engagement into active social interaction, passive browsing, content creation, and social comparison. This typology emerged from latent class analysis applied at each of the three measurement waves, and it held up consistently across the entire observation period. At baseline, passive browsing was the largest group, accounting for 35.2 percent of participants, followed by active social engagement at 28.7 percent, social comparison at 21.4 percent, and content creation at 14.7 percent. Over the twelve months, the proportion of passive browsers crept upward to 37.8 percent while content creators shrank to 11.3 percent—a drift the author tentatively links to mounting academic pressure, though the design cannot confirm that explanation.

Measurement was unusually careful for this field. Social media patterns were captured through a 24-item scale cross-validated against seven-day usage diaries and, for a subsample of 720 participants, passive screen-time monitoring software installed on their devices. Emotion regulation was assessed with two complementary instruments covering emotional awareness, understanding, cognitive reappraisal, expressive suppression, and strategy flexibility. All scales exceeded conventional reliability thresholds, and measurement invariance across gender and school level was confirmed, allowing meaningful group comparisons across the full sample.

The analytical strategy paired two statistical approaches that are rarely used together with this degree of care. Latent growth curve models traced how regulatory capacity changed over the year, while a random-intercept cross-lagged panel model separated stable between-person differences from genuine within-person change, probing the temporal ordering of media behavior and emotional outcomes. Critically, the author estimated these models separately rather than forcing their incompatible assumptions into a single structure. The growth models revealed modest average improvement in regulation across the cohort, but with substantial individual variation in both starting points and rates of change—and it was here that usage patterns made their mark.

Relative to passive browsers, active social users showed a steeper positive trajectory in emotion regulation, while social comparison users displayed a significantly flatter, nearly stagnant growth rate. The cross-lagged analyses sharpened the picture further. Passive browsing prospectively predicted slower gains specifically in cognitive reappraisal and strategy flexibility, the capacities that depend most on active cognitive engagement. Active engagement, by contrast, predicted small gains in emotional awareness and understanding. The author applied a smallest effect size of interest of 0.10 in standardized units, treating smaller paths as unsupported even when statistically significant—a discipline that keeps the conclusions honest in a sample large enough to flag trivially small effects.

Perhaps the most striking result concerns directionality. A Cross-Lag Asymmetry Index, comparing the strength of media-to-regulation paths against the reverse, was highest for social comparison use at 0.52, indicating that the prospective association ran predominantly from comparison-oriented platform behavior toward weaker regulation rather than the other way around. Passive browsing showed a similar but weaker asymmetry at 0.38, and active use a more reciprocal pattern at 0.24. This suggests that comparison-heavy engagement may be more than a mere symptom of pre-existing emotional difficulty—though the author is careful to note that even the largest cross-lagged coefficient remains small, and no observational design can establish causation.

The mediation analyses identified the mechanisms behind these links. For passive browsing, social comparison tendency was the dominant mediator, followed by self-esteem, while perceived online social support played no meaningful role in that pathway. For active social engagement, the pattern reversed: online social support carried the protective indirect effect, and social comparison contributed nothing. Gender moderated only one chain—the passive browsing to comparison to regulation pathway—with the conditional indirect effect roughly twice as large for girls as for boys. The author interprets this cautiously: both estimates are small, the girls’ effect barely clears the pre-specified threshold, and the moderation emerged on just one of four pathways, offering thin support for popular narratives about girls as a uniquely vulnerable group.

The study is notable as much for what it resists claiming as for what it finds. The author explicitly warns against reading the protracted maturation of adolescent prefrontal circuits as evidence that teenagers are uniquely defenseless, notes that recent meta-analyses find associations between social media use and adolescent well-being to be small and context-dependent, and declines to endorse population-level social media bans on the strength of effects this modest. Time-varying confounds—changes in school, peer groups, or academic strain—go unmeasured, and heritable traits like neuroticism, which was controlled rather than treated as a moderator, could still account for part of the observed covariation. Attrition reached 17.8 percent over the year, though analyses indicated the missing data did not depart detectably from a random pattern.

Still, the practical implications, however tentative, are pointed. If the associations reflect genuine processes—which this design cannot establish—then guiding adolescents toward interactive, reciprocal use rather than comparison-heavy scrolling would be a more promising target than uniform screen-time limits. Schools might embed digital-literacy modules that help students recognize comparison traps, and platform designers face renewed questions about architectures that optimize for exactly the passive consumption mode most consistently linked to slower regulatory growth. The study also flags a behavioral transition worth watching: nearly one in five social comparison users migrated toward passive browsing over the year, suggesting comparison-driven engagement may mark a transitional phase rather than a settled style.

Future work, the author argues, should extend tracking to three or more years, incorporate neuroimaging measures of prefrontal-amygdala connectivity, replicate across cultures, and—most importantly—move toward randomized interventions that experimentally shift usage patterns and measure downstream effects on regulation. Until then, the study stands as a methodological benchmark for a field long criticized for collapsing richly varied digital behaviors into a single hours-per-day number. Its message is measured but clear: the question of whether social media helps or harms adolescent emotional development is badly posed. The better question—which patterns of use matter, through what mechanisms, and for whom—now has its first serious longitudinal answer, and the answer is that not all screen time is created equal.

Subject of Research: Longitudinal associations between adolescent social media use patterns and emotion regulation development

Article Title: How social media use patterns shape adolescent emotion regulation: A three-wave longitudinal study of mediating mechanisms and moderating conditions

Article References: Li, H. (2026). How social media use patterns shape adolescent emotion regulation: A three-wave longitudinal study of mediating mechanisms and moderating conditions. Current Psychology, 45(18), Article 1500. https://doi.org/10.1007/s12144-026-09983-w

Image Credits: AI Generated

DOI: 10.1007/s12144-026-09983-w

Keywords: social media, adolescents, emotion regulation, passive browsing, social comparison, longitudinal study, cross-lagged panel model, online social support, self-esteem, screen time, adolescent development, digital well-being

The research, published in the Journal of Autism and Developmental Disorders, was led by Hatice Gundeslioglu of the University of Warwick together with colleagues at the University of Birmingham and Monash University. The team followed UK undergraduates across two waves of an online survey spaced six months apart, comparing autistic students, identified through self-reported clinical diagnosis, with their non-autistic peers. Their central question was deceptively simple: do anxiety, depression and suicidality feed into one another over time, and does being autistic, or being a particular gender, change those dynamics?

To answer it, the researchers used a statistical technique called a cross-lagged panel model, or CLPM. Unlike a simple correlation, which cannot tell you whether A drives B or B drives A, a cross-lagged design measures the same variables at two time points and asks whether each construct predicts change in the others after accounting for its own stability. In this study, the first wave recruited 747 undergraduates between November 2022 and June 2023; six months later, 209 participants completed the follow-up, and 179 provided complete data for the final models. Mental health was measured with well-validated instruments, including the GAD-7 for anxiety and the PHQ-9 for depression, alongside items on self-harm and suicidal ideation.

An exploratory factor analysis distilled the questionnaire items into three distinct dimensions: anxiety and worry, behavioural symptoms of depression and anxiety, and suicidality. The resulting models showed an exceptionally good statistical fit, with fit indices at or near their ideal values. That matters because it gives confidence that the patterns the researchers report are not artefacts of a poorly specified model, but reflect genuine structure in the data.

The headline finding is directional. Suicidality at the first wave significantly predicted increases in both anxiety and worry and in behavioural symptoms of depression and anxiety at the second wave, with standardised coefficients of 0.21 and 0.25 respectively. These are small to moderate effects, but in a longitudinal design even modest effects can accumulate into clinically meaningful change. In other words, a student who reported suicidal thoughts or self-harm at baseline was more likely to be more anxious and more depressed six months later, even after the researchers statistically accounted for how stable those problems already were.

Equally striking was the sheer stability of mental health across the six-month window. The autoregressive paths, which capture how well a construct predicts itself over time, ranged from roughly 0.70 to 0.80 across the three outcomes. Values in that range indicate high but not perfect stability: most of what determines a student’s mental health at the follow-up is simply where they stood six months earlier. Students who were struggling at the start of the study were, by and large, still struggling at the end, a pattern consistent with earlier work showing that psychological distress among undergraduates tends to persist across an academic year rather than resolve on its own.

Where does autism fit in? At baseline, autistic undergraduates reported significantly higher levels of anxiety, depression and suicidality than their non-autistic peers, confirming what previous cross-sectional research had suggested. Yet, perhaps counterintuitively, an autism diagnosis did not predict change in mental health over the six-month period. Once baseline levels were taken into account, autistic and non-autistic students followed similar trajectories. The authors suggest several possible explanations: some autistic students may have received targeted support or developed effective coping strategies; the autistic group is highly heterogeneous in support needs and circumstances; or six months may simply be too short a window to capture divergence that emerges over years.

Gender told a similar story of null effects. Neither gender nor autism diagnosis was significantly associated with mental health outcomes at follow-up in the cross-lagged models. That finding runs against some prior studies reporting that female students experience more stress, anxiety and depression during university, and against evidence that autistic females face heightened suicide risk. The authors caution that their sample was modest, skewed by attrition, and recruited largely through online communities, which may have masked underlying gender differences. Notably, they could not test whether autism and gender interact, for example whether autistic female students fare worse over time than autistic male students, because the follow-up sample of 179 fell well short of the 300 participants their pre-registered analysis plan called for to support such complex models.

The study’s limitations deserve honest attention. Attrition was substantial: only about 28 percent of the original wave-one sample completed the follow-up, and those retained differed from those lost, with autistic students retained at higher rates. Missing data were not completely at random, forcing the researchers to rely on complete-case analysis, which reduces statistical power and may limit generalisability. Autism status rested on self-report of a clinical diagnosis, and eleven participants reported receiving their diagnosis during the study interval itself. The six-month follow-up is also short by longitudinal standards, and the authors recommend future studies with at least three waves and longer intervals to capture change more reliably.

Even so, the practical implications are hard to ignore. If suicidality predicts subsequent increases in anxiety and depression, then screening for suicidal thoughts and self-harm, and intervening early, could blunt the development of broader mental health problems in the student population. The stability of symptoms over time argues for regular, long-term monitoring rather than one-off assessments, so that universities can identify students whose difficulties persist and intervene before they deepen. And although no overall gender differences emerged in this analysis, the wave-one data from the same cohort showed that autistic female students reported elevated suicidality, leading the authors to suggest that university well-being services should pay particular attention to this subgroup. For a generation of students, autistic and non-autistic alike, the message from this research is clear: suicidal distress is not just a symptom to be managed after the fact, but a signal that deserves immediate, sustained attention.

Subject of Research: Longitudinal bidirectional relationships between suicidality, anxiety and depression among autistic and non-autistic UK undergraduates

Article Title: A Cross-Lagged Panel Model of Mental Health Amongst UK Undergraduates: The Role of Autism and Gender

Article References: Gundeslioglu, H., Gray, K. M., Thompson, P. A., & Langdon, P. E. (2026). A Cross-Lagged Panel Model of Mental Health Amongst UK Undergraduates: The Role of Autism and Gender. Journal of Autism and Developmental Disorders. https://doi.org/10.1007/s10803-026-07541-8

Image Credits: AI Generated

DOI: 10.1007/s10803-026-07541-8

Keywords: autism, undergraduates, mental health, suicidality, anxiety, depression, cross-lagged panel model, longitudinal study, gender, UK universities, GAD-7, PHQ-9

The research team focused on a construct known as suicide cognitions: deeply embedded, maladaptive thought patterns such as perceived burdensomeness, thwarted belongingness, and the belief that death is a solution to one’s problems. Unlike suicidal ideation, which refers to the momentary presence of suicidal thoughts, suicide cognitions represent the underlying belief architecture that can sustain risk even when acute thoughts subside. The researchers asked whether these cognitions might indirectly set the stage for depression sensitivity — a tendency to interpret everyday negative experiences as evidence of personal failure and impending depressive episodes — through their effects on nighttime arousal and resilience.

To test this, the team recruited university students in Türkiye and collected data at two time points separated by a five-month interval. After matching participants across both waves, the final sample comprised 280 Turkish university students with a mean age of 22.35 years and a standard deviation of 3.88. Each participant completed four validated instruments: the Brief Suicide Cognitions Scale, the Depression Sensitivity Index, the Pre-Sleep Arousal Scale, and the Brief Resilience Scale. The Pre-Sleep Arousal Scale, originally developed by Nicassio and colleagues in 1985, captures both somatic arousal — a racing heart, tense muscles, shallow breathing — and cognitive arousal, the intrusive, spiraling thoughts that keep people awake long after the lights go out.

Methodologically, the study employed structural equation modeling to estimate two separate two-wave cross-lagged panel models, or CLPMs. This statistical framework is particularly valuable for longitudinal research because it allows investigators to examine whether a variable measured at the first time point predicts change in another variable at the second time point, above and beyond the stable, baseline level of the outcome itself. In this study, the models controlled for baseline levels of all study variables as well as participants’ age and gender, sharpening the ability to detect genuine prospective associations rather than mere stability in traits over time.

The results revealed a striking chain of prospective relationships. Students who reported higher suicide cognitions at the first wave showed significantly higher pre-sleep arousal and lower resilience five months later, even after accounting for their initial levels of arousal and resilience. In turn, students with higher pre-sleep arousal and lower resilience at the first wave reported greater depression sensitivity at the second wave, controlling for their baseline depression sensitivity. In other words, the data suggested a pathway in which maladaptive suicidal beliefs appear to fuel nighttime physiological and cognitive agitation while simultaneously eroding the psychological resources that normally buffer against stress.

Perhaps the most intriguing finding was what the models did not show. The direct associations between suicide cognitions at the first wave and depression sensitivity at the second wave were not statistically significant in either cross-lagged panel model. Yet when the researchers used bootstrapping — a resampling technique that provides robust estimates of indirect effects — the indirect pathways through both pre-sleep arousal and resilience were statistically significant. This pattern suggests that suicide cognitions may not erode emotional resilience to depression directly; instead, their influence appears to travel through the body’s nighttime state and through the individual’s capacity to recover from setbacks.

These findings resonate with a broader literature linking sleep disturbances to suicide risk. Prior research has shown that sleep problems can outperform depression and hopelessness as predictors of suicidal ideation and behavior in young adults, and cognitive models of insomnia, such as Harvey’s influential 2002 framework, describe how nighttime arousal perpetuates itself through worry and monitoring of sleep-related threats. The new study extends this work by positioning pre-sleep arousal as a prospective bridge between suicidal belief systems and vulnerability to depression, rather than merely a symptom that accompanies both conditions. It also aligns with evidence that rumination and repetitive negative thinking are associated with suicidal ideation across depressive disorders.

The role of resilience in the models is equally consequential. Resilience, as measured by the Brief Resilience Scale, reflects the ability to bounce back from stress and adversity, and a growing body of research links it to lower levels of psychological distress, depression, and suicidal ideation. In the present study, higher suicide cognitions predicted lower resilience over time, and lower resilience in turn predicted greater depression sensitivity. This suggests that interventions designed to strengthen resilience — from cognitive-behavioral skills training to self-compassion and mindfulness-based approaches — might interrupt the pathway by which suicidal thinking deepens vulnerability to depressive episodes in young adults.

The study’s authors are careful to frame their results as preliminary and indicative rather than definitive. The sample consisted of university students in a single country, and the two-wave design, while stronger than a cross-sectional snapshot, cannot fully establish causal direction or rule out unmeasured confounding variables. The five-month interval between assessments is meaningful but may miss faster-moving dynamics, and the study was not pre-registered. Nonetheless, the use of matched longitudinal data, validated Turkish-language instruments, and rigorous structural equation modeling with bootstrapped indirect effects lends considerable credibility to the pattern of findings.

The practical implications are significant for campus mental health services and clinicians working with young adults. Screening for suicide cognitions may identify students at risk not only for suicidal behavior but also for a slower-burning vulnerability to depression that operates through sleep and resilience. Sleep-focused interventions that reduce pre-sleep cognitive and somatic arousal — including relaxation training, stimulus control, and mindfulness practices — could serve as accessible entry points for prevention, while resilience-building programs may offer a complementary protective buffer. As the authors conclude, their findings provide preliminary evidence about the cognitive, physiological, and emotional processes that may be associated with vulnerability to depression in university students who experience suicide cognitions, pointing toward integrated approaches that treat disturbed sleep and diminished resilience as modifiable targets in suicide prevention.

Subject of Research: Prospective associations among suicide cognitions, pre-sleep arousal, resilience, and depression sensitivity in university students

Article Title: Pre-sleep arousal and resilience in suicide cognitions and depression sensitivity: indications from two-wave cross-lagged panel models

Article References: Akbulut, Ö. F., Kütük, H., Akyıl, Y., Okur, S., & Erdinç, B. (2026). Pre-sleep arousal and resilience in suicide cognitions and depression sensitivity: indications from two-wave cross-lagged panel models. Current Psychology, 45(19), Article 1560. https://doi.org/10.1007/s12144-026-10127-3

Image Credits: AI Generated

DOI: 10.1007/s12144-026-10127-3

Keywords: suicide cognitions, depression sensitivity, pre-sleep arousal, resilience, cross-lagged panel model, university students, longitudinal study, sleep, mental health, structural equation modeling, depression, suicide risk

Emerging adulthood, typically defined as the span from the late teens through the twenties, is a life stage characterized by identity exploration, instability, and a heightened susceptibility to substance and behavioral addictions. It is also the period during which social media use tends to peak. Young adults in this age group are, in many respects, digital natives whose social lives, romantic relationships, academic identities, and professional aspirations are all mediated to a substantial degree by online platforms. The study’s authors, Hugo Sanchez-Sanchez and Konstanze Schoeps of the University of Valencia and Frances Hoferichter of Greifswald University, set out to disentangle the direction of the relationships between social media addiction, emotion regulation strategies, and multiple domains of quality of life in this population, moving beyond the cross-sectional snapshots that dominate the field.

To do so, the team recruited 620 German university students and followed them across three waves of data collection. At each wave, participants completed validated measures of social media addiction, their habitual use of two distinct emotion regulation strategies, and several dimensions of health-related quality of life, including psychological well-being, physical well-being, and self-esteem. The researchers then applied a cross-lagged panel model, a structural equation modeling technique specifically designed to examine temporal and reciprocal associations between variables measured repeatedly over time. By statistically controlling for each construct’s own prior level, cross-lagged models allow researchers to ask whether changes in one variable predict later changes in another, providing far stronger grounds for inferring directional effects than single-timepoint correlations can offer.

The results were striking in their consistency. Social media addiction at one wave reliably predicted lower levels of physical well-being, psychological well-being, and self-esteem at the following wave, even after accounting for participants’ baseline standing on each of these outcomes. In other words, students whose relationship with social media displayed addictive characteristics, such as salience, tolerance, mood modification, relapse, and withdrawal-like symptoms, tended to experience subsequent deterioration across all three well-being domains examined. Notably, the reverse pathways were weaker: declines in well-being did not consistently predict later increases in social media addiction, suggesting that the primary directional flow runs from problematic use toward diminished well-being rather than the other way around.

The physical health findings deserve particular attention, because much of the existing literature on social media effects has concentrated on mood, anxiety, and depression while paying less attention to somatic outcomes. The new results align with experimental work showing that deliberate reductions in social media usage can produce measurable improvements in physical health and daily functioning. Plausible mechanisms include the displacement of sleep, exercise, and face-to-face activity by compulsive scrolling, as well as the physiological stress associated with constant connectivity, notification-driven interruptions, and the pressure of perpetual social comparison. Whatever the precise pathway, the longitudinal pattern indicates that heavy, compulsive engagement with social platforms is followed by worse self-reported physical condition, not simply co-occurring with it.

The study’s second major contribution concerns emotion regulation, the repertoire of processes by which individuals influence which emotions they have, when they have them, and how they experience and express those emotions. Drawing on the influential process model of emotion regulation, the researchers distinguished between two strategies with very different consequences. Expressive suppression, which involves inhibiting the outward expression of ongoing emotional states, predicted decreases in every dimension of well-being measured, including psychological health, physical health, and self-esteem. Cognitive reappraisal, by contrast, which involves reinterpreting emotionally charged situations in ways that alter their meaning and emotional impact, predicted improvements over time in psychological well-being and self-esteem. These findings replicate a robust pattern in the emotion regulation literature, in which suppression is associated with physiological arousal, strained social relationships, and poorer mental health, while reappraisal is linked to resilience and adaptive functioning.

The theoretical framing of the study draws on the Interaction of Person-Affect-Cognition-Execution model of addictive behaviors, which conceptualizes problematic use of internet-related applications as the product of interactions between an individual’s predisposing characteristics, affective states, cognitive processes, and executive control capacities. Within this framework, social media platforms can serve as a maladaptive coping strategy, offering immediate but shallow relief from negative emotions while undermining the development of more sustainable regulatory skills. The new longitudinal data give this model empirical teeth: young adults who rely on suppression to manage their emotions fare worse over time, while those who can reframe situations maintain or improve their psychological resources, and addictive social media use chips away at well-being regardless.

For researchers, the methodological rigor of the design matters as much as the headline findings. Much of the public debate about social media harms has been fueled by correlational studies that cannot establish temporal precedence, leaving room for skeptics to argue that unhappy people simply use social media more. The three-wave cross-lagged panel approach substantially strengthens the causal interpretation, although the authors and methodologists alike caution that even longitudinal panel designs cannot fully rule out confounding by unmeasured variables, and that alternative analytic techniques such as random-intercept cross-lagged models continue to refine how such data should be interpreted. The study also benefited from established measurement instruments, including the Bergen Social Media Addiction Scale, whose longitudinal measurement invariance has been demonstrated in prior work, and a German-language version of the Emotion Regulation Questionnaire.

The practical implications reach well beyond the laboratory. The authors argue that their findings underscore the need for public health initiatives that promote healthier digital habits and simultaneously strengthen emotion regulation skills during emerging adulthood. Universities, which sit at the epicenter of this developmental stage, are natural settings for such interventions. Programs that teach cognitive reappraisal and other adaptive regulation strategies could serve a dual purpose, both directly supporting students’ psychological resources and potentially reducing the appeal of social media as an emotional escape hatch. At the same time, the consistent link between addictive use and physical well-being suggests that digital wellness efforts should address sleep, activity, and somatic health rather than focusing exclusively on mood.

As social media platforms continue to embed themselves in the daily routines of young people worldwide, evidence of this kind shifts the conversation from whether social media matters for well-being to which patterns of use, and which individual skills, determine its impact. The message emerging from this three-wave study of German students is twofold and unusually clear: compulsive, addiction-like engagement with social media is a longitudinal risk factor for declining physical health, psychological well-being, and self-esteem in emerging adults, but the trajectory is not fixed. Adaptive emotion regulation, particularly the capacity to reappraise difficult situations rather than suppress their expression, functions as a protective factor that can bend the curve toward resilience. For a generation navigating identity, intimacy, and ambition through a screen, cultivating that capacity may be one of the most consequential health investments available.

Subject of Research: Longitudinal effects of social media addiction and emotion regulation on quality of life in emerging adults

Article Title: Social Media Addiction, Emotion Regulation, and Quality of Life in Emerging Adults: A Cross-Lagged Panel Model

Article References: Sanchez-Sanchez, H., Schoeps, K., & Hoferichter, F. (2026). Social Media Addiction, Emotion Regulation, and Quality of Life in Emerging Adults: A Cross-Lagged Panel Model. International Journal of Mental Health and Addiction. https://doi.org/10.1007/s11469-026-01728-x

Image Credits: AI Generated

DOI: 10.1007/s11469-026-01728-x

Keywords: social media addiction, emerging adulthood, emotion regulation, cognitive reappraisal, expressive suppression, quality of life, self-esteem, psychological well-being, physical well-being, cross-lagged panel model, longitudinal study, university students

The research, conducted by Jin-long Liang and Xin-ling Cui of the College of Education Sciences at Northwest Normal University in Lanzhou, China, drew on caregiver-reported data from the KIT longitudinal study, formally known as the Kids in Taiwan: National Longitudinal Study of Child Development and Care. The analytic sample comprised 1,694 children, each assessed at approximately 36, 48, 60, and 72 months of age. By tracking the same children at four points across three years, the researchers could ask a question that single-snapshot studies cannot answer: does memory at one age predict literacy skills a year later, above and beyond what the child’s literacy level already suggested?

Emergent literacy, the construct at the heart of the study, refers to the precocious skills that precede formal reading instruction. In this dataset, it was measured with six caregiver-rated items capturing a child’s understanding of common signs and pictorial symbols such as traffic lights, the ability to orient a book correctly, awareness that a picture book’s cover carries information such as the title and author, knowledge of where print begins on a page, the capacity to recognize simple Chinese characters such as one’s own name, and the ability to express ideas through drawings or written symbols like drawing diaries or object lists. Memory, by contrast, was assessed with a broader 18-item scale that included pretend play, recalling and stating personal details such as age, birth month, and telephone numbers, counting accurately to 21 and later to 100, identifying previously seen storybooks, and distinguishing living from nonliving things.

At the simplest descriptive level, the results confirmed what developmental scientists would expect: average scores on both constructs climbed steadily with age, and children who scored high on memory at any given wave also tended to score high on emergent literacy at that same wave, and at every other wave. But the study’s real contribution lies in how it disentangled the direction of these associations over time. The researchers employed two complementary statistical frameworks. The first, a traditional cross-lagged panel model, or CLPM, estimates overall longitudinal associations between the two constructs across the four waves. The second, a random-intercept cross-lagged panel model, or RI-CLPM, is a more recent refinement that separates stable, trait-like differences between children from time-specific deviations within each child, a distinction that has become central to modern longitudinal methodology.

The traditional CLPM told a strikingly consistent story about memory. In all three one-year intervals, memory scores prospectively predicted subsequent emergent literacy scores, even after accounting for prior literacy levels. Emergent literacy showed the reverse pattern in two of the three intervals, prospectively predicting later memory. In other words, when the two constructs were analyzed with the conventional approach, the pathway running from memory to literacy appeared somewhat more robust than the pathway running from literacy back to memory, though both directions were present in the data.

The RI-CLPM, however, is where the study delivers its most technically important insight. When the model separated each child’s stable overall standing from their wave-specific deviations, the correlation between the stable between-child components of memory and emergent literacy was a remarkable 0.81. Children who are generally stronger in memory relative to their peers are, to an extraordinary degree, also the children who are generally stronger in emergent literacy. This single statistic captures the substantial overlap between the two capacities at the level of stable individual differences, and it suggests that much of the association reported in earlier cross-sectional work may reflect this shared underlying profile rather than dynamic developmental influence.

Yet the within-child results demonstrate that the story does not end with stable traits. At the within-child level, deviations in memory were positively associated with subsequent deviations in emergent literacy across all three intervals, and deviations in emergent literacy were likewise positively associated with subsequent deviations in memory across all three intervals. This means that when a child performed better than their own typical level in memory at one assessment, they tended to perform better than their own typical level in emergent literacy at the next assessment, and vice versa. The bidirectional relationship, in other words, is not merely an artifact of some children being generally more advanced than others; it unfolds dynamically within individual developmental trajectories.

One temporal detail stands out. The largest estimated coefficient for the pathway from emergent literacy to memory appeared in the interval from 60 to 72 months, spanning the transition from age five to age six. This is precisely the period during which many children begin encountering formal or semi-formal literacy instruction, and the finding hints that as children’s print-related skills consolidate, they may increasingly feed back into the memory processes that support learning more broadly. The authors note that the prospective association from memory to emergent literacy was more consistent across the two modeling approaches than the reverse pathway, suggesting that memory may serve as the somewhat more reliable developmental engine during this age window, even as literacy skills exert their own growing influence.

Methodologically, the study is a textbook demonstration of why the choice of longitudinal model matters. Had the researchers relied solely on the traditional CLPM, they might have overinterpreted cross-lagged paths that partly reflect stable trait-level covariation. Had they relied solely on the RI-CLPM, they might have dismissed genuine within-child dynamics. By running both models on the same four-wave dataset, Liang and Cui were able to show that the longitudinal relation between memory and emergent literacy involves both stable between-child differences and time-specific within-child variation, a dual conclusion that neither model alone could have delivered. The analysis used full information maximum likelihood to handle missing data and robust estimation procedures, and the work was supported by research funding from the College of Education Sciences at Northwest Normal University.

The practical implications reach into homes, preschools, and early intervention programs. If memory skills reliably foreshadow literacy growth from age three onward, then activities that exercise young children’s recall, sequencing, and symbolic reasoning, from pretend play to storybook recognition, may pay dividends in later reading readiness. Conversely, the feedback loop from emergent literacy to memory suggests that early exposure to books, print conventions, and symbolic expression is not merely preparation for reading but an active contributor to broader cognitive development. Because the study relied on caregiver reports and secondary analysis of an existing cohort, the authors are careful to frame the findings as associations rather than causal proof, and they call for continued work with direct assessments and additional populations. Still, for parents and educators wondering whether the toddler memorizing songs and the preschooler scribbling a drawing diary are on connected journeys, this research offers a resounding answer: the two journeys are woven together, each one helping to pull the other forward through the critical years before formal schooling begins.

Subject of Research: Bidirectional longitudinal associations between memory and emergent literacy in preschool children aged 3 to 6 years.

Article Title: Longitudinal associations between memory and emergent literacy from 3 to 6 years of age: a four-wave random-intercept cross-lagged panel study

Article References: Liang, J.-L., & Cui, X.-L. (2026). Longitudinal associations between memory and emergent literacy from 3 to 6 years of age: a four-wave random-intercept cross-lagged panel study. BMC Psychology. https://doi.org/10.1186/s40359-026-05598-9

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05598-9

Keywords: memory, emergent literacy, early childhood, longitudinal study, cross-lagged panel model, random-intercept CLPM, cognitive development, language development, preschool children, KIT study, child development, developmental psychology

The study, conducted by a team led by Yasuo Murayama of Kanazawa University and published in Child Psychiatry & Human Development, tracked 634 students in grades 3 through 9 across all public elementary and junior high schools in Wajima City, a community that suffered extensive damage in the January 2024 Noto Peninsula Earthquake. Data were collected at three time points between July 2024 and February 2025, roughly three months apart, as part of a broader mental health support program run jointly by the Ishikawa Kokoro no Care Center and Kanazawa University’s Research Center for Child Mental Development. Participation was voluntary, anonymous, and approved by the Chukyo University Research Ethics Committee.

What makes the research technically distinctive is its modeling strategy. The investigators employed a three-wave cross-lagged panel model, a structural equation technique that estimates whether each variable predicts change in the others over time while statistically controlling for each variable’s own prior level. The researchers deliberately chose the traditional cross-lagged panel model over the newer random-intercept variant, arguing that in a period when situational stress was widely shared across the community, a random intercept would risk conflating stable trait-like differences with persistent contextual effects of the disaster. The final model fit the data exceptionally well, with a comparative fit index of 1.00 and a root mean square error of approximation of just 0.011.

The study’s central and most striking finding concerns an asymmetry in the relationship between loneliness and depressive symptoms. While the two constructs showed bidirectional longitudinal associations, the pathway running from depressive symptoms to later loneliness was substantially stronger and more consistent than the reverse. Depressive symptoms at one wave predicted loneliness at the next with standardized coefficients ranging from 0.15 to 0.23, both statistically significant at p less than 0.001 and qualifying as large effect sizes under recent guidelines for cross-lagged effects. In contrast, loneliness predicted subsequent depressive symptoms with coefficients of only 0.06 to 0.08, one of which was merely marginally significant.

This directional pattern partially diverges from earlier meta-analytic work, which reported relatively comparable reciprocal effects between loneliness and depression in the general population. The authors propose that the post-disaster context may explain the difference. Cognitive models of depression hold that depressive symptoms are associated with negative biases in how people interpret their social environments, and in a community where nearly everyone is struggling, awareness that others face similar hardships may blunt the usual impact of loneliness on mood. Meanwhile, elevated depressive symptoms may still distort individuals’ perceptions of their relationships, gradually deepening feelings of isolation. Consistent with this interpretation, research conducted during the COVID-19 pandemic similarly found stronger longitudinal effects from depression to loneliness than the reverse in collectively affected populations.

The behavioral findings were equally revealing. Aggressive behavior consistently predicted higher subsequent loneliness, with standardized coefficients of roughly 0.08 to 0.09 at both intervals, and it also forecast declines in prosocial behavior with coefficients between negative 0.11 and negative 0.12. Notably, the reverse pathway from loneliness to later aggression was not supported, a result that contrasts with laboratory and animal research suggesting social isolation heightens aggression. Instead, the data suggest that in the complex social ecology of real classrooms, it is aggressive conduct that erodes future social standing and connection, echoing a five-year longitudinal study of first-grade children that found aggression predicted reductions in prosocial behavior while the reverse was not significant.

One unexpected result concerns prosocial behavior. In the first interval, higher prosocial behavior actually predicted a small increase in subsequent loneliness, a finding that ran counter to the common assumption that helping others buffers against isolation. The researchers offer a tentative hypothesis: in a post-disaster community under collective stress, children who frequently engage in helping behavior, or who perceive themselves as doing so, may focus excessively on others’ needs at the expense of their own, potentially receiving less support in return. They note prior evidence that self-sacrificing altruism is sometimes linked with interpersonal difficulties, but they emphasize this interpretation should be regarded as a preliminary hypothesis, since individual disaster exposure was not directly assessed.

The temporal stability of the constructs added further nuance. Between the first and second waves, depressive symptoms and aggressive behavior showed notably high stability, with autoregressive coefficients of 0.680 and 0.697 respectively, while loneliness was considerably less stable at 0.345. Loneliness, the authors suggest, is a socially constructed and cognitively mediated experience that is more state-dependent and therefore more malleable over time, whereas emotional and behavioral problems may have been particularly entrenched during the recovery period. Analyses comparing students with complete data to those with missing waves found no significant baseline differences on any measure, suggesting attrition did not bias the results, and exploratory multi-group analyses indicated the longitudinal associations were largely invariant across sex and grade level.

The findings carry concrete implications for practitioners supporting young people after disasters. Because the study’s evidence aligns with the developmental cascade framework, which views development as a system of mutually reinforcing interactions across emotional, behavioral, and social-cognitive domains, interventions targeting a single domain may produce cascading benefits elsewhere. The authors argue that early identification and treatment of depressive symptoms should be a priority, given their central and stable role, while the lower temporal stability of loneliness suggests that efforts to foster a shared sense of social connectedness within affected communities may be especially valuable. Even modest, coordinated multi-domain support, they contend, may contribute meaningfully to psychosocial recovery in resource-limited settings.

The researchers also acknowledge important limitations. The observational design precludes causal conclusions, the traditional cross-lagged panel model cannot cleanly separate within-person change from stable between-person differences, and the exclusive reliance on self-report measures may have inflated associations through shared method variance. Socioeconomic status was not assessed, a deliberate ethical choice given the sensitivity of income questions for families whose livelihoods were disrupted by the earthquake, leaving open the possibility of residual confounding. Because the sample was drawn exclusively from a disaster-affected, ethnically homogeneous Japanese community and lacked a comparison group, the authors caution that the observed asymmetry between depression and loneliness may reflect post-disaster circumstances or broader developmental processes, and they frame several of their interpretations as hypothesis-generating rather than definitive. Future research incorporating perceived social support, multi-informant measures, and earlier post-disaster time points could help clarify how young people’s emotional lives and social worlds reshape one another in the long aftermath of catastrophe.

The Wajima findings gain additional weight when placed against the broader epidemiological picture of youth loneliness. A meta-analysis drawing on data from 113 countries has estimated that roughly one in ten adolescents aged 12 to 17 experiences loneliness, while earlier surveys in Japan suggested that nearly one-third of children and adolescents reported the feeling. Loneliness, in this literature, is understood not as objective isolation but as a subjective aversive state arising from a perceived gap between desired and actual relationships, which helps explain why it can persist even in densely connected communities like Wajima.

Adolescence is considered a period of particular vulnerability because peer relationships, social evaluation, and identity formation all reach heightened salience, making the grade 3 through 9 span covered by the study developmentally informative. The pandemic-era research cited by the authors offers a useful parallel: adolescents worldwide reported elevated loneliness during COVID-19, and studies of collectively affected populations likewise found stronger longitudinal effects running from depression to loneliness than the reverse, mirroring the asymmetry observed after the earthquake.

The developmental cascade framework that anchors the study emerged from empirical demonstrations that internalizing problems, externalizing problems, and interpersonal quality influence one another reciprocally over time, meaning no single domain should be examined in isolation. Prior prospective work has shown depressive symptoms predicting later aggression, prosocial behavior buffering later depressive symptoms, and externalizing problems eroding later prosociality. By simultaneously modeling all four constructs, the Wajima study addresses a gap left open by research that considered these variables only in pairs, and it suggests that the cascade processes documented in general populations remain operative even under the unusual strain of a large-scale natural disaster.

Subject of Research: Longitudinal associations among loneliness, depressive symptoms, and social behaviors in youth after a major earthquake

Article Title: Longitudinal Interrelations Among Loneliness, Depressive Symptoms, and Social Behaviors in Youth: A Cross-Lagged Panel Study in a Post-Disaster Context

Article References: Murayama, Y., Nishiura, M., Soma, D., Kameya, M., Tanaka, S., Tsujii, M., & Kikuchi, M. (2026). Longitudinal Interrelations Among Loneliness, Depressive Symptoms, and Social Behaviors in Youth: A Cross-Lagged Panel Study in a Post-Disaster Context. Child Psychiatry & Human Development. https://doi.org/10.1007/s10578-026-02096-2

Image Credits: AI Generated

DOI: 10.1007/s10578-026-02096-2

Keywords: loneliness, depressive symptoms, aggressive behavior, prosocial behavior, children and adolescents, cross-lagged panel model, developmental cascade, post-disaster mental health, psychosocial adjustment, Noto Peninsula Earthquake, Japan, longitudinal study

Perinatal depression, a debilitating mood disorder occurring during pregnancy and up to one year postpartum, has long been a focus of mental health research due to its multifactorial origins and potential impact on both maternal and infant outcomes. Previous studies often relied on static correlational analyses, which limited the ability to detect causal or bidirectional relationships between depressive symptoms and potential risk factors. The study by Lin et al. importantly addresses this methodological limitation by utilizing a cross-lagged panel design that allows for the temporal sequencing of variables, thereby enabling researchers to infer more robust directional influences.

The cross-lagged panel model (CLPM) is a structural equation modeling technique that analyzes the reciprocal relationships between variables measured at multiple time points. Unlike traditional regression methods, which typically analyze data at a single time point, CLPM accounts for stability in constructs over time and captures how one variable may predict changes in another across successive waves of data collection. This methodological strength is crucial when assessing psychological phenomena like depression, which are inherently dynamic and influenced by a complex constellation of biopsychosocial factors.

Lin and colleagues meticulously tracked a cohort of high-risk perinatal women over several critical time points during pregnancy and postpartum periods. High-risk designation was based on pre-existing medical, psychological, and sociodemographic factors known to increase vulnerability to mood disorders. By integrating repeated assessments of depressive mood alongside related psychosocial variables such as stress levels, social support, and hormonal changes, the researchers were able to construct a comprehensive model depicting not just correlations but potential causal pathways shaping depressive trajectories.

One of the pivotal revelations from the study was the bidirectional influences observed between depressive mood and perceived social support. Rather than a simple unidirectional effect whereby lack of social support exacerbates depression, the analysis uncovered a feedback loop in which worsening depressive symptoms also lead individuals to perceive or experience diminished social support over time. This cyclical dynamic highlights the necessity of interventions that simultaneously bolster social networks while directly addressing mood symptoms to disrupt this pernicious cycle.

Hormonal fluctuations, particularly involving cortisol and estrogen levels, were also incorporated into the CLPM framework to elucidate their temporal effects on mood states. The findings indicated that cortisol elevations during late pregnancy predicted subsequent increases in depressive symptoms postpartum. However, the reverse pathway was not significant, suggesting that biological stress mechanisms may act as precursors rather than consequences of mood deterioration in this context. This insight corroborates the growing body of evidence implicating dysregulated hypothalamic-pituitary-adrenal (HPA) axis function in perinatal mood disorders.

Apart from biological and social variables, psychological constructs such as coping strategies and cognitive appraisal styles were integral to Lin et al.’s model. Their data demonstrated that maladaptive coping not only predicted an increase in depressive symptoms at follow-up but that elevated depression also impaired effective coping ability, reinforcing the concept of reciprocal causation. Importantly, the timing and magnitude of these effects varied according to the perinatal stage, underscoring the need for developmental sensitivity in clinical assessment and intervention planning.

By combining biological markers, psychosocial factors, and temporal sequencing, the study offers a nuanced view of the etiology and persistence of perinatal depression, moving beyond one-dimensional causal explanations. It underscores the heterogeneity and complexity in at-risk populations, thereby challenging the notion of universal intervention models. The implications for personalized medicine and precision psychiatry are profound, suggesting that treatment plans should be tailored to the individual temporal dynamics uncovered through sophisticated longitudinal analyses such as the cross-lagged panel model.

Another considerable strength of this investigation is its potential to inform preventative strategies during pregnancy. Understanding the early predictive markers for depression enables clinicians to identify those individuals who would most benefit from timely psychosocial support, stress reduction techniques, and possibly pharmacological interventions before the full onset of mood episodes. The temporal insights from the CLPM also facilitate monitoring of treatment efficacy, as shifts in key variables can be tracked over time to adjust intervention intensity or modality.

At a broader level, Lin et al.’s research project contributes to destigmatizing conversations about maternal mental health by highlighting the biological underpinnings and contextual risk factors in a manner accessible to interdisciplinary stakeholders. Policymakers, healthcare providers, and caretakers can leverage these findings to advocate for integrated screening programs and allocate resources toward comprehensive perinatal mental healthcare services, potentially reducing the long-term societal and familial burdens associated with untreated depression.

From a methodological standpoint, the study exemplifies cutting-edge applications of latent variable modeling in psychological epidemiology. The rigorous use of cross-lagged panel analysis provides a blueprint for future research endeavors aiming to parse out directionalities in complex psychosocial phenomena. Moreover, the inclusion of diverse mediators and moderators within the model enhances explanatory power and ecological validity, setting a new standard for psychometric and longitudinal research designs in perinatal mental health.

The authors also address potential limitations candidly, noting the challenges inherent in capturing self-reported data on mood and social variables, possible attrition biases, and the generalizability of findings to broader populations beyond the high-risk cohort studied. They propose future work involving larger sample sizes, incorporation of neuroimaging biomarkers, and cross-cultural validations to enhance applicability and refine mechanistic understanding further.

Given the escalating global concern over mental health disorders in vulnerable populations, this study represents a significant milestone in perinatal psychiatry research. It seamlessly integrates theoretical rigor, clinical relevance, and methodological innovation to reshape how depressive mood and its influencing factors are conceptualized, assessed, and managed across the perinatal timeline. As these insights permeate clinical practice, the hope is that more women will receive timely, effective support, ultimately improving maternal and child health outcomes worldwide.

The advent of this cross-lagged panel approach signals a paradigm shift toward embracing temporally sensitive models that acknowledge the bidirectional, multifaceted nature of psychological disorders. Lin et al.’s contribution reverberates beyond perinatal depression, offering a versatile analytical template that can be adapted to study dynamic relationships in various mental health conditions, thereby broadening its impact on psychiatric research and care.

Collectively, the work of Lin and colleagues advances both the science of perinatal mental health and the practical frameworks necessary for combating depressive disorders in high-risk populations. Their findings herald a new era of personalized, temporally informed mental health care, underscored by the nuanced realities of depression’s evolution during the critical perinatal period.

Subject of Research: Dynamics of depressive mood and influencing psychosocial and biological factors in high-risk perinatal women using a cross-lagged panel model approach.

Article Title: Cross-lagged panel model of depressive mood and influencing factors in high-risk perinatal depression.

Article References:
Lin, S., Hong, Y., Hong, H. et al. Cross-lagged panel model of depressive mood and influencing factors in high-risk perinatal depression. BMC Psychol (2026). https://doi.org/10.1186/s40359-026-03970-3

Image Credits: AI Generated