Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Sep 5.
Published before final editing as: Nat Ment Health. 2025 Aug 12:10.1038/s44220-025-00481-9. doi: 10.1038/s44220-025-00481-9

Cognitive and Global Morphometry Trajectories as Predictors of Youth Persistent Distressing Psychotic-Like Experiences

Nicole R Karcher 1, Fanghong Dong 1, Sarah E Paul 2, Emma C Johnson 1, Can M Kilciksiz 1,3, Hans Oh 4, Jason Schiffman 5, Arpana Agrawal 1, Ryan Bogdan 2, Joshua J Jackson 2, Deanna M Barch 1,2
PMCID: PMC12411066  NIHMSID: NIHMS2105627  PMID: 40917767

Abstract

Psychotic-like experiences (PLEs) may arise from genetic and environmental risk leading to worsening cognitive and morphometry metrics over time, which in turn lead to worsening PLEs. Analyses used three waves of unique longitudinal Adolescent Brain Cognitive Development Study data (ages 9–13) to test whether changes in cognition and global morphometry metrics attenuate associations between genetic and environmental risk with persistent distressing PLEs. Multigroup univariate latent growth models examined three waves of cognitive metrics and global morphometry separately for three PLE groups: persistent distressing PLEs (n=356), transient distressing PLEs (n=408), and low-level PLEs (n=7901). Persistent distressing PLEs showed greater decreases (i.e., more negative slopes) of cognition and morphometry metrics over time compared to those in low-level PLE groups. Analyses also provided novel evidence for extant theories that worsening cognition and global morphometry metrics may partially account for associations between environmental risk with persistent distressing PLEs.

Keywords: Adolescent Brain Cognitive Development Study, psychotic-like experiences, adolescence, trajectories, genetic liability, environment

Psychotic-like experiences (PLEs) are subclinical abnormal thought content and perception abnormalities which are on the continuum of psychosis spectrum1,2. Studies indicate evidence for the clinical relevance of distressing PLEs, with 57–92% of those with significantly distressing PLEs developing a diagnosable mental health problem later in adulthood, including psychotic disorders or other diagnosable conditions, such mood and behavioral disorders3,4. Although research indicates PLEs are not necessarily associated with later risk for psychotic disorders5 or sustained negative outcomes2, other research finds childhood and adolescent PLEs are associated with a range of impairments in domains including cognition and pathophysiology, including global morphometrymetrics6–10. Our recent work using the Adolescent Brain Cognitive Development℠ (ABCD) Study data found evidence that persistent (i.e., occurring over multiple years) and distressing PLEs showed, as expected, generally small effect-sized associations with a range of risk factors (e.g., environmental exposures, cognitive impairments) measured at the start of adolescence9. This present work represents an important expansion from prior cross-sectional work by examining the extent to which these identified cognitive and global morphometry factors over time represent mechanisms underlying associations between genetic and environmental risk with persistent distressing PLEs.

Literature shows that longitudinal changes in cognition, including executive function, are both risk factors for and consequences of general psychopathology11. There is evidence that longitudinal changes in cognition may be particularly associated with the development of schizophrenia spectrum symptoms12, and that decline in cognition is present at the subclinical phase of psychosis spectrum in early ages13. Although results on cognitive deficits among high risk or prodromal populations are promising for early detection of psychosis, there is limited data about changes in cognitive domains in populations experiencing PLE even before the risk phase12, with some evidence for development lags in cognition13, although other evidence for weak declines over time14. Following changes in cognition may be used as an important tool for tracking worsening functioning.

Early childhood through adolescence is a key developmental period for brain development and maturation. Brain changes during this time include but are not limited to general decreases in cortical thickness, surface area, and volume15. This peak of brain maturation and neuroplasticity during adolescence makes this a period of vulnerability for the development of psychopathology16. Studies examining longitudinal changes in global morphometry metrics associated with the psychosis symptom spectrum find individuals who convert to psychosis generally show reductions in grey matter in a range of frontal17,18), temporal17,19), and cingulate and cerebellar regions19). One recent review indicates that this accelerated decline in a range of areas (e.g., frontal, temporal, cingulate, and parietal) extends to individuals experiencing PLEs20). Overall, this research indicates psychosis spectrum symptoms are associated with longitudinal global morphometry deviations spanning a diverse range of neural regions.

Theory indicates that PLEs may arise as a result of a combination of genetic and environmental factors leading to worsening pathophysiological factors (e.g., cognition, global morphometry metrics), which in turn lead to worsening PLEs21). The present work therefore also explored whether cognitive or global morphometry metrics over time attenuated associations between genetic (quantified using schizophrenia polygenic scores and family history of psychosis) and environmental risk to persistent distressing PLEs. Our previous work found evidence that significantly distressing PLEs are associated with genetic liability (as measured by polygenic risk scores for schizophrenia, PLEs, broad cross-disorder psychopathology, and lowered educational attainment)22, as well as environmental risk (as measured by a composite of previously implicated environmental risk factors (i.e., poverty, lack of neighborhood safety, deprivation, and estimated lead risk exposure)9,23. Showing changes in cognitive and global morphometry metrics may alter links between genetic and environmental factors with worsening PLEs informs both our understanding of the pathogenesis of PLEs and the possibility that altered cognitive and global morphometry metrics may affect PLEs trajectories, information useful for early identification and intervention (e.g., cognitive therapy and remediation24) efforts.

The Adolescent Brain Cognitive Development (ABCD) study is a longitudinal study that aims to understand the neurodevelopmental trajectories of a diverse sample of over 11,000 children across the United States. It was hypothesized that impairments in cognitive metrics and deviations of global morphometry metrics over time would be most strongly associated with persistent distressing PLEs compared to low-level PLEs or transient distressing PLEs, and that these metrics would attenuate associations between both genetic and environmental factors with PLEs.

RESULTS

Multi-group growth curve models estimated slopes for cognitive and global morphometry metrics over time (i.e., baseline, 2-year-follow-up and 4-year follow-up), simultaneously modeling and freely estimated for each of the three PLE groups [i.e., persistent distressing (n=456), transient distressing (n=408), low-level PLEs (n=7901)]. Information about the persistent distressing, transient distressing, and low-level PLE groups can be found in Table 1.

Table 1.

Participant Characteristics

Persistent Distressing (n=356) Transient Distressing (n=408) Low-Level PLEs (n=7901) Low PLEs Matcheda (n=762)
% N % N % N % N
Sex (% female) 59.83% 213 55.88% 228 45.41% 3588 57.74% 440
Race/Ethnicity (%)
Black 28.37% 101 26.72% 109 10.78% 852 12.07% 92
Hispanic 24.72% 88 24.51% 100 17.80% 1406 17.85% 136
Asian 0.56% 2 1.72% 7 2.53% 200 2.89% 22
Multiracial/Multiethnic 13.48% 48 12.50% 51 10.02% 792 10.76% 82
White 59.83% 213 55.88% 228 45.41% 3588 57.74% 440
Mean SD Mean SD Mean SD Mean SD
Age at Baseline 9.81 0.61 9.83 0.62 9.93 0.63 9.85 0.61
Baseline Distressing PLEs 20.56 19.16 13.04 14.60 3.94 7.33 3.86 6.84
1-Year Distressing PLEs 25.08 18.95 12.98 17.15 1.75 4.09 1.52 3.58
2-Year Distressing PLEs 25.92 15.49 10.80 13.95 0.82 1.62 0.79 1.60
3-Year Distressing PLEs 23.56 13.75 7.78 10.96 0.62 1.36 0.73 1.50
4-Year Distressing PLEs 23.58 16.41 10.63 12.72 0.54 1.18 0.51 1.20
Baseline Pubertal Status 2.31 0.86 2.32 0.80 2.01 0.82 2.04 0.85
2-Year Pubertal Status 3.04 0.92 3.00 0.90 2.62 0.97 2.70 0.98
4-Year Pubertal Status 3.73 0.80 3.81 0.79 3.55 0.78 3.60 0.81
Open in a new tab

Abbreviations. PLEs=psychotic-like experiences; N=sample size; %=percentage; SD=standard deviation.

a

Low PLEs group was matched on age, sex, and race/ethnicity using SPSS Case-Control Matching function (low PLEs matched to a combined PLEs group). Matching was done without replacement.

Cognition Trajectories in PLE groups

As can be seen in Figure 1, the persistent distressing PLEs group relative to the low-level PLEs group showed lower initial scores (i.e., negative intercepts at ages 9–10; Supplemental Table 2) and worsening performance over time (i.e., negative slopes) across cognitive tasks (difference between slope estimates≤−0.093, FDRps<.02, except for Flanker inhibitory control: difference between slope estimates=−0.141, FDRp=.07; see Table 2 for group differences; see Supplemental Table 3 for estimates for each PLE group). The transient distressing PLEs group showed lower initial values across tasks (i.e., negative intercepts; Supplemental Table 2) and significant worsening performance over time (i.e., negative slopes; Figure 1) for the picture vocabulary, picture sequence memory, and working memory tasks compared to the low-level PLEs groups (difference between slope estimates≤−0.16, FDRps<.01).

Figure 1.

Figure 1.

Open in a new tab

Cognitive and global morphometry metric slope estimates from growth curve models (z-scored across the entire sample; error bars represent standard errors) for low-level psychotic-like experiences (PLEs; n=7901), persistent distressing PLEs (n=356), and transient distressing PLEs (n=408) groups when including covariates (i.e., age, sex, pubertal status). Negative slope estimates indicate decreasing scores over time. Data points reflect model-derived estimates. Error bars represent the standard error of the estimated slope from the growth curve model.

Table 2.

Difference between slope estimates Comparisons of Each of the PLE Groups

Persistent Distressing vs. Low-Level PLEs Transient Distressing vs. Low-Level PLEs Persistent Distressing vs. Transient Distressing PLEs
Est. lower 95% CI upper 95% CI Z value FDRp Est. lower 95% CI upper 95% CI Z value FDRp Est. lower 95% CI upper 95% CI Z value FDRp
Pattern −0.181 −0.287 −0.075 −3.342 0.002* −0.083 −0.196 0.031 −1.433 0.17 0.098 −0.056 0.252 1.251 0.88
Picture Vocabulary −0.171 −0.256 −0.086 −3.934 <.001* −0.168 −0.229 −0.107 −5.417 <.001* 0.003 −0.1 0.106 0.059 0.95
Flanker −0.141 −0.293 0.011 −1.82 0.07 −0.115 −0.24 0.011 −1.791 0.09 0.026 −0.169 0.222 0.265 0.94
Picture Sequence −0.173 −0.283 −0.063 −3.08 0.003* −0.16 −0.24 −0.08 −3.908 <.001* 0.013 −0.122 0.147 0.187 0.94
Reading −0.093 −0.168 −0.018 −2.442 0.02* −0.075 −0.149 0 −1.969 0.07* 0.018 −0.085 0.122 0.35 0.94
List Sorting −0.28 −0.484 −0.077 −2.698 0.009* −0.198 −0.348 −0.049 −2.602 0.01* 0.082 −0.167 0.331 0.645 0.94
Thickness −0.144 −0.055 −0.233 −3.169 0.003* −0.046 −0.032 −0.124 −1.156 0.25 0.098 −0.018 0.213 1.656 0.88
Volume −0.175 −0.107 −0.244 −5.026 <.001* −0.122 −0.062 −0.182 −3.992 <.001* 0.054 −0.036 0.143 1.177 0.88
Subcortical Volume −0.11 −0.051 −0.169 −3.67 <.001* −0.1 −0.05 −0.149 −3.949 <.001* 0.011 −0.065 0.086 0.278 0.94
Surface Area −0.115 −0.048 −0.182 −3.349 0.002* −0.096 −0.045 −0.147 −3.689 <.001* 0.019 −0.064 0.102 0.444 0.94
Intracranial Volume −0.162 −0.045 −0.28 −2.702 0.009* −0.11 −0.028 −0.192 −2.623 0.01* 0.052 −0.089 0.194 0.726 0.94
Open in a new tab

Abbreviations. Est=estimate; PLE=psychotic-like experiences; 95% CI= 95% confidence interval; Z=Z statistics; p=p-value; FDRp=false discovery-rate p-value. Two-sided Z-tests were used to compare slope estimates between groups, all estimates were FDR-corrected across 11 test for each group comparison.

*

p<.05

MRI Metrics Trajectories in PLE groups

For global morphometry metrics, as can be seen in Figure 1, the persistent distressing group relative to the low-level PLEs group showed lower initial values (i.e., negative intercepts at ages 9–10; Supplemental Table 2) and accelerated decreases over time (i.e., negative slopes) across all metrics, including cortical and subcortical volume, cortical surface area, cortical thickness, and intracranial volume (difference between slope estimates≤−0.11, FDRp<.009, Table 2; see Supplemental Table 3 for estimates for each PLE group). The transient distressing PLEs group relative to the low-level PLEs group showed lower initial values (i.e., negative intercepts; Supplemental Table 2) and accelerated decreases over time (i.e., negative slopes, Figure 1) for cortical and subcortical volume, cortical area, and intracranial volume, but not for cortical thickness.

Follow-up Analyses

The analyses above incorporated a priori definitions of PLEs groups, and so we also conducted latent profile analyses to examine data-driven definitions of PLE groups, with results highly consistent (Supplemental Results). Analyses also examined whether results remained consistent with matching demographics between the PLEs groups with the low-level PLEs group (Supplemental Table 4), as well as when specificity of models by incorporating internalizing and externalizing symptoms over time in models (Supplemental Table 5), with results remaining consistent.

Models Testing Attenuation of Associations between Genetic and Environmental Factors and PLE Groups.

First, we examined associations between either genetic liability (PGS, family history of psychosis) or environmental risk with PLE group comparisons. Greater environmental risk scores and greater family history of psychosis (FDRps<.05; Supplemental Table 6), but not schizophrenia PGS scores, were associated with being in the persistent distressing PLE versus the low-level PLEs group. When family history and environmental risk scores were entered simultaneously, only environmental risk scores remained associated with the persistent distressing PLEs (95%CI:0.174,0.311).

Genetic and Environment Risk with Cognition and Imaging Slope Metrics.

Given the associations between both family history of psychosis and environmental risk with PLE group comparisons, we examined associations between either environmental risk scores or family history with cognition and imaging slope metrics. Greater environmental risk scores were associated with lower cognitive scores over time (i.e., negative slopes; βs<−.08, FDRps<.004; Supplemental Table 7, except for pattern processing speed).

Higher environmental risk scores were associated with lower imaging slopes (βs<−0.08, FDRps<.001) across all imaging metrics (i.e., thickness, cortical volume, subcortical volume, cortical area, and intracranial volume; Supplemental Table 7). Associations between family history of psychosis with cognitive or imaging scores over time did not survive correction for multiple comparisons (ps>.01, FDRps>.10).

Evidence for Cognitive and Neural Trajectories Attenuating Environmental Risk with PLEs.

Given the above results, we next explored whether inclusion of cognition/global morphometry slopes attenuated (i.e., mediated) associations between environmental risk with persistent distressing versus low-level PLEs group comparison. As can be seen in Figure 2, the inclusion of cognitive slopes and to a lesser extent imaging slopes attenuated associations between environmental risk with persistent distressing versus low-level PLEs PLEs (i.e., slopes attenuated between 2.4% (intracranial volume) – 26.38% (picture vocabulary)).

Figure 2.

Figure 2.

Open in a new tab

(A) Slope estimates for growth curve models examining associations between environmental risk with Persistent Distressing PLEs (n=356) vs. Low-level PLEs (n=7901) either with or without accounting for cognitive or imaging metric slopes (error bars represent standard errors). Positive estimates indicate greater associations between environment risk with Persistent Distressing PLEs. Bars reflect model-derived slopes estimates. Error bars represent the standard error of the estimated indirect effect from the growth curve model.

(B) Estimated proportion of the association between environmental risk with Persistent Distressing PLEs vs. Low-level PLEs attenuated when accounting for each cognitive or imaging metric slope metric.

DISCUSSION

The present work provides several important novel advances towards understanding the pathogenesis of early PLEs including: a) providing novel evidence for longitudinal changes over six years in cognitive and global morphometry metrics predicting persistent distressing PLEs, b) providing insight into whether associations between genetic and environmental factors and PLEs are attenuated by cognition and global morphometry metrics over time, consistent with a plausible mediation pathway. Deviations in cognitive and cortical thickness metrics were evident in a gradient across PLE groups, with persistent distressing PLEs compared to low-level PLEs showing the largest impairments, followed by the transient distressing PLEs group, even when accounting for youth-reported internalizing and externalizing symptoms (Supplemental Table 5). Although the effect sizes were smaller than those generally seen in clinical high-risk studies12, they were larger than previous cross-sectional work in this sample9, providing evidence for the importance of understanding declines over time. Further, worsening cognitive and to a lesser degree imaging metrics attenuated links between higher environmental risk and persistent distressing PLEs. The findings provide key information that youth who show persistent distressing PLEs are characterized by demonstrable decreases in cognition and global morphometry metrics, and that these impairments in cognition and structural metrics may partially account for associations between environmental risk with persistent distressing PLEs.

Across cognitive tests, the distressing PLEs groups showed worsening performance over time, with the strongest findings for persistent distressing PLEs. The persistent distressing PLEs group showed evidence of worsening performance over time across cognitive metrics, whereas the transient distressing PLEs group showed worsening performance in a subset of tests including picture vocabulary, picture sequence, and working memory. These results are generally consistent with previous cross-sectional work, including our work in the ABCD study25, where findings have pointed to broad ranging cognitive impairments associated with PLEs25,26 spanning both crystallized and fluid cognition impairments, with effect strongest for persistent distressing PLEs.

For global structural imaging metrics, as with the cognitive metrics, the distressing PLEs groups showed broad accelerated declines over time, with the strongest findings for persistent distressing PLEs. Although volume, area, and thickness generally decline across adolescence15 especially the persistent distressing PLEs group showed evidence of accelerated declines in cortical metrics over time compared to the low-level PLEs group. Persistent distressing PLEs showing evidence of greater reductions in cortical thickness over time is consistent with prior work indicating that accelerated cortical thinning often precedes the onset of psychosis17. The most robust global morphometry difference between the groups was the accelerated decrease in cortical thickness within individuals in the persistent distressing PLEs group, a finding that may be consistent with accelerated pruning or inflammation processes, both of which have been implicated in psychosis risk 27,28. Overall, the persistent distressing PLEs group showed widespread accelerated declines in global structural metrics over time, and there was evidence for that cortical thickness showed the most robust associations, such that only the persistent distressing PLEs group showed accelerated declines in cortical thickness when compared to the low-level PLEs group.

Attenuation analyses point to the potential role of worsening cognition and brain metrics in linking environmental risk to persistent distressing PLEs. Genetic liability analyses indicated that family history of psychosis was associated with persistent distressing PLEs, but polygenic scores for schizophrenia were not. Transient distressing PLEs were not strongly associated with family history of psychosis or schizophrenia polygenic scores, which may point the heterogeneity of transient distressing PLEs. Consistent with our prior work, persistent distressing PLEs were strongly associated with greater environmental risk scores23, bolstering evidence for the importance of environmental factors including poverty and adverse life events in relation to persistent PLEs23,29. Worsening cognitive and global morphometry metrics over time attenuated associations between environmental risk and persistent distressing PLEs, indicating that the robust link of the environment with PLEs may be through diffuse worsening of cognitive and global morphometry metrics over time. These data provide important empirical evidence consistent with previous theoretical models regarding the development and worsening of PLEs30, namely that both genetic liability and environmental risk play a role in adolescents experiencing persistent distressing PLEs, and that these environmental associations are partially attributable to worsening cognition and by worsening global morphometry metrics. These findings are consistent with stress-sensitization and developmental cascade theories, where exposure to environmental risk factors leads to changes in neurobiological functioning to make youth more vulnerable to worsening symptoms 31,32.

This work had several limitations and points to consider. The groups were created based on a priori (versus data driven) definitions of group membership, as we were specifically interested in following up on our previous work examining these definitions of PLEs25, although it is likely that heterogeneity exists within these groups (e.g., remitting trajectories), as evidenced by the greater variability in cognitive and global morphometry metric slopes for PLE groups compared to the low-level PLE group (see Figure 1). Further, latent profile analyses (Supplemental Table 9) generally supported these results, with 98% of the persistent distressing group falling within the high PLEs latent profile, whereas only 31% of the transient distressing PLEs group fell into the high PLEs latent profile. Variables chosen for analyses relied on our past cross-sectional work examining genetic liability, cognition, structural metrics, and environment9,22,23. Results with polygenic scores warrant caution.33 Polygenic scores models were limited to ancestries to ancestral populations where we have previously validated PGS. Important genetic effects on PLEs in individuals of other genetic ancestries might have been missed, however, this is likely partially recovered through our family history measure which in part assesses genetic liability. Genetic liability analyses therefore have incomplete generalizability, which may have the potential to increase structural inequities including health disparities 34, necessitating future research to understand associations between genetic liability, structural inequities, and risk for psychosis.

Future work should examine the role of other potentially influential factors, including drug use and mental health service utilization in potentially exacerbating or attenuating these associations, respectively. Research should also examine the influence of factors that may vary by marginalized communities, including structural racism, as well as service utilization and potential targets for intervention, including physical health-related variables (e.g., sleep, physical activity, nutrition, substance use) in interrelationships between environment, cognitive, global morphometry, and PLEs over time. Future work should investigate early intervention strategies that incorporate staged approaches identifying at-risk youth endorsing environment risk factors and monitoring pathophysiological declines and symptoms whenever possible, providing individualized care targeting both environmental context and individual vulnerability. Efficient and inexpensive methods for assessing pathophysiological changes over time should also be investigated.

Overall, there was evidence that compared to the low-level PLEs group, persistent distressing PLEs showed broad evidence of worsening of cognitive functioning and accelerated declines in global morphometry over the course of approximately six years. The present work provides important evidence for extant theory, providing some of the first evidence for general cognitive and global morphometry metrics attenuate links between environment with worsening psychosis spectrum symptoms over the course of middle childhood and adolescence. This research provides important advances over previous cross-sectional work (Supplemental Table 10), indicating that declines in cognitive and global morphometry are already evident in middle childhood, declines that are generally smaller than in high-risk populations12, and therefore this developmental time period may represent a key period for early intervention. This work provides information regarding the nature of worsening PLEs in adolescence, including the role of cognition and the brain as potential mechanisms linking genes and environment to worsening PLEs.

METHODS

Participants

The ABCD study aimed to recruit a sample reflecting the demographic variation of the U.S. population, recruiting children using probability sampling from both public and private elementary schools. Study-wide exclusionary criteria were as follows: child not fluent in English, MRI contraindication (e.g., irremovable ferromagnetic implants or dental appliances, claustrophobia, pregnant), major neurological disorder, gestational age less than 28 weeks or birthweight less than 1,200 grams, history of traumatic brain injury, or had a current diagnosis of schizophrenia, autism spectrum disorder (moderate, severe), mental retardation/intellectual disability, or alcohol/substance use disorder.35–37 Caregivers provided written informed consent and all children provided assent. Centralized institutional review board approval was obtained from the University of California, San Diego Institutional Review Board (protocol # 160091).

Each participant’s past 3 waves of data were used to define PLE groups (see below). Cognitive and MRI data was obtained at baseline, 2-year follow-up and 4-year follow-up. All available data were used in analyses and missing data were handled using full information maximum likelihood (fiml) estimation (Supplemental Table 12 for models using imputed data; Supplemental Table 1 for sample sizes for included variables).

Measures

Prodromal Questionnaire-Brief Child Version (PQ-BC)

Participants completed the previously validated Prodromal Questionnaire-Brief Child Version (PQ-BC)8). As the present work followed up our prior cross-sectional work, we used the same PLE group designations as in our prior work9,38, with group definitions based on previous research39. The distress score reflects both the presence and intensity of distress for each of the 21 items: responses are scored as 0 (not endorsed), 1 (endorsed but no distress), or 2–6 (endorsed with increasing levels of distress, based on the distress scale; range:0–126). Analyses examined the following groups (Table 1):

1) a persistent distressing PLEs group that scored>=1.96SDs above the mean for distressing PLEs for 2 or more of the 3 waves of data based on the participant’s last three waves of data from baseline through 4-year follow-up (n=356, range of mean PLEs across waves: 20.56–25.92(27); means and standard deviation were examined across the entire sample for each assessment wave); 2) a transient distressing PLEs group that scored >=1.96 SDs above the mean for distressing PLEs for 1 wave of data and scored <=0.50 SDs above the mean for distressing PLEs for the other 2 waves of data (n=408, range of mean PLEs across waves: 7.78–13.04); 3) a low-level PLEs group that scored <=0.50SDs above the mean for distressing PLEs for all three waves of data (n=7901, range of mean PLEs across waves: 0.54–3.94).

Neuropsychological Test Battery

National Institutes of Health Toolbox Cognitive Battery (NIHTB-CB) tests included flanker inhibitory control, pattern processing speed, picture vocabulary, and readings tests, administered at baseline, 2-year-follow-up and 4-year follow-up assessments, as well as list sorting working memory, which was administered at baseline and 4-year follow-up (for included tests, administration time is approximately 22 minutes; for more details, see40.

Structural MRI Measures

Consistent with prior work9, structural MRI measures included intracranial volume, total cortical and subcortical volumes41, total surface area42, and mean cortical thickness43. All data were acquired on a 3T scanner (Siemens, General Electric, or Phillips) with a 32-channel head coil and completed T1-weighted and T2-weighted structural scans (1mm isotropic). Structural neuroimaging processing was completed using FreeSurfer version 5.3.0 through standardized processing pipelines44 (Supplemental Methods for additional details). Participants that did not pass FreeSurfer Quality Control measures (i.e., at least one T1 scan that passed all quality control metrics) were excluded from analyses (n=142). We used longitudinal ComBat harmonization45, with age and sex added as biological covariates to the design matrix, and individual, site, and family nested as a random effects, to estimate and remove individual scanner effects (i.e., using Siemens, Phillips, GE device serial numbers) from MRI measures prior to entry into models.

Genetic and Environmental Risk Scores

To account for genetic liability, we examined schizophrenia polygenic scores (PGS)22 and family history of psychosis. Summary statistics from well-powered discovery GWASs of schizophrenia for youth most genetically similar to individuals from European reference populations (N=69,369 cases+236,642 controls)46 and youth most genetically similar to individuals from African reference populations (N=6152 cases+3918 controls)47 were used. PGS were generated using polygenic risk scores-continuous shrinkage (PGS-CS; see Supplemental Methods for more information)48. These analyses were done in a subset of individuals from European ancestry (n=4533) and from African ancestry (n=1201), due to the sample compositions of the discovery GWASs and evidence that polygenic risk scores do not translate well across genetic ancestries in an unbiased manner49. To examine a proxy for potential genetic (and non-genetic) liability across the entire ABCD sample, we also examined whether caregivers endorsed any family history of psychosis. For environmental risk scores, based on previous work9,23 we created a principal component analysis (PCA) using baseline metrics of caregiver-rated perception of neighborhood safety, number of years at current residence, and based on primary address: drug crime exposure, overall deprivation, rate of poverty, and lead exposure risk estimates (see Supplemental Methods for more details). The extracted first principal component was used as a predictor in analyses.

Statistical Analysis

Multi-group (i.e., persistent distressing, transient distressing, low-level PLEs) growth curve models (growth function in lavaan package in R version 0.6–15) estimated slopes for cognitive and global morphometry metrics over time (i.e., baseline, 2-year-follow-up and 4-year follow-up), simultaneously modeled and freely estimated for each of the three PLE groups. Differences between the groups were specified in each model as difference scores between individual parameter estimates. Given the focus of the present work on understanding persistent distressing PLEs, analyses primarily focused on the slope difference of persistent distressing vs. low-level PLEs. To examine specificity of findings, analyses additionally examined transient distressing vs. low-level PLEs and persistent distressing vs. transient distressing PLEs. To account for multiple comparisons, these difference between slope estimates were false discovery-rate (FDR) corrected (11 FDR-corrected tests: 6 cognition metrics, 5 MRI metrics) for each group comparison. Models also examined mean initial scores (i.e., as indexed by the intercept which was set at the baseline assessment) separately for each of the three PLE groups. Analyses included sex at baseline as a time invariant covariate, with pubertal status and age in months as time varying covariates. Variables were examined for skew, and all continuous variables with skew ≥|1.96| were winsorized to 3SDs.

Follow-up analyses examined whether changes in cognition and global morphometry metrics (ncognition metric=6; nMRI metric=5) attenuated (i.e., mediated) associations between genetic indices and environmental risk as predictors and with a group comparison as outcome (e.g., persistent distressing versus low-level PLEs). We first examined associations between genetic indices (i.e., schizophrenia PGS or family history of psychosis) with a two-level group variable (e.g., persistent distressing versus low-level PLEs), as well as a separate model between environmental risk score with a two-level group variable (e.g., persistent distressing versus low-level PLEs). We next examined whether the association with between either genetic or environmental risk factors with cognition and global morphometry slopes were significant. Since only environmental risk was associated with both cognitive/global morphometry metrics and persistent distressing PLEs vs. low-level PLEs, we focused our attenuation analyses on these metrics. For each of the 11 cognitive/global morphometry metrics, we fit a model estimating all mediation paths (i.e., between environment risk and each plausible mediator, as well as paths between environmental risk and persistent distressing versus low-level PLEs, between potential mediators and persistent distressing vs. low-level PLEs). We examined indirect (mediated), direct, and total paths within these models. These analyses were conducted using the lavaan R package growth with FDR adjustment for multiple testing (n=11). Covariates in these models were identical to those used in our primary analyses. While our analyses formally tested indirect paths linking genetic and environmental risk factors to persistent distressing PLEs through longitudinal changes in cognition and brain morphometry, we acknowledge an important limitation: the variables overlapped in terms of temporal ordering. This means that the temporal ordering required for definitive mediation testing is not strictly met. As such, our findings should be interpreted as consistent with—but not conclusive evidence of—indirect pathways. Nonetheless, our approach provides a reasonable, developmentally informed framework to examine how longitudinal neurocognitive patterns relate to distinct PLE groups

We followed up analyses of the a priori PLE groups by conducting data-driven analyses of PLEs group using latent profile analyses [LPA; using tidyLPA (version 1.1.0) and mclust (version 6.1.1) packages].50,51 Latent profile models with one through five profiles were estimated using PLE distress scores at the five time points. Models were evaluated based on fit indices, and when there was disagreement between fit indices in terms of which model showed the best fit, model selection was based on the lowest BIC 52. Latent growth curves compared extracted latent profile assignments on trajectories of cognitive and neural data.

Supplementary Material

Supplemental Methods and Results

Acknowledgements

Data used in the preparation of this article were obtained from the Adolescent Brain Cognitive Development (ABCD) Study (https://abcdstudy.org), held in the NIMH Data Archive (NDA). This is a multisite, longitudinal study designed to recruit more than 10,000 children age 9–10 and follow them over 10 years into early adulthood. The ABCD Study is supported by the National Institutes of Health and additional federal partners under award numbers U01DA041022, U01DA041025, U01DA041028, U01DA041048, U01DA041089, U01DA041093, U01DA041106, U01DA041117, U01DA041120, U01DA041134, U01DA041148, U01DA041156, U01DA041174, U24DA041123, and U24DA041147. A full list of supporters is available at https://abcdstudy.org/nih-collaborators. A listing of participating sites and a complete listing of the study investigators can be found at https://abcdstudy.org/principal-investigators.html. ABCD consortium investigators designed and implemented the study and/or provided data but did not necessarily participate in analysis or writing of this report. This manuscript reflects the views of the authors and may not reflect the opinions or views of the NIH or ABCD consortium investigators. The ABCD data repository grows and changes over time.

This manuscript is the result of funding in whole or in part by the National Institutes of Health (NIH). It is subject to the NIH Public Access Policy. Through acceptance of this federal funding, NIH has been given a right to make this manuscript publicly available in PubMed Central upon the Official Date of Publication, as defined by NIH.

This work was supported by National Institute of Health grants U01 DA041120 (DMB), K23 MH121792 (NRK), R01-MH139880 (NRK), R01-DA054869, (AA), K01-DA051759 (ECJ), R01-DA054750 (AA, RB), and F31-AA029934 (SEP).

Footnotes

Competing Interests

The authors do not report any competing interests.

Data Availability

Data used in the preparation of this article were obtained from the Adolescent Brain Cognitive Development (ABCD) Study (https://abcdstudy.org), held in the NIMH Data Archive (NDA). This is a multisite, longitudinal study designed to recruit more than 10,000 children aged 9–10 and follow them over 10 years into early adulthood. The ABCD data repository grows and changes over time. The ABCD data used in this report came from https://nda.nih.gov/study.html?id=2313.

References

  • 1.Karcher NR Psychotic-like experiences in childhood and early adolescence: Clarifying the construct and future directions. Schizophr Res 246, 205–206 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Staines L et al. Psychotic experiences in the general population, a review; definition, risk factors, outcomes and interventions. Psychol Med 52, 3297 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Fisher HL et al. Specificity of childhood psychotic symptoms for predicting schizophrenia by 38 years of age: a birth cohort study. Psychol Med 43, 2077–2086 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Poulton R et al. Children’s Self-Reported Psychotic Symptoms and Adult Schizophreniform Disorder: A 15-Year Longitudinal Study. Arch Gen Psychiatry 57, 1053–1058 (2000). [DOI] [PubMed] [Google Scholar]
  • 5.Zammit S et al. Psychotic experiences and psychotic disorders at age 18 in relation to psychotic experiences at age 12 in a longitudinal population-based cohort study. American Journal of Psychiatry 170, 742–750 (2013). [DOI] [PubMed] [Google Scholar]
  • 6.Evermann U, Gaser C, Besteher B, Langbein K & Nenadić I Cortical Gyrification, Psychotic-Like Experiences, and Cognitive Performance in Nonclinical Subjects. Schizophr Bull 46, 1524–1534 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Satterthwaite TD et al. Structural Brain Abnormalities in Youth With Psychosis Spectrum Symptoms. JAMA Psychiatry 73, 515–524 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Karcher NR et al. Assessment of the Prodromal Questionnaire-Brief Child Version for Measurement of Self-reported Psychoticlike Experiences in Childhood. JAMA Psychiatry 75, 853–861 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Karcher NR et al. Persistent and distressing psychotic-like experiences using adolescent brain cognitive developmentSM study data. Mol Psychiatry 27, 1490–1501 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Schoorl J et al. Grey and white matter associations of psychotic-like experiences in a general population sample (UK Biobank). Transl Psychiatry 11, 21 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Romer AL & Pizzagalli DA Associations Between Brain Structural Alterations, Executive Dysfunction, and General Psychopathology in a Healthy and Cross-Diagnostic Adult Patient Sample. Biological Psychiatry Global Open Science 2, 17 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Karcher NR, Merchant J, Pine J & Kilciksiz CM Cognitive Dysfunction as a Risk Factor for Psychosis. Curr Top Behav Neurosci 63, 173–203 (2023). [DOI] [PubMed] [Google Scholar]
  • 13.Gur RC et al. Neurocognitive growth charting in psychosis spectrum youths. JAMA Psychiatry 71, 366–374 (2014). [DOI] [PubMed] [Google Scholar]
  • 14.Dickson H et al. Trajectories of cognitive development during adolescence among youth at-risk for schizophrenia. J Child Psychol Psychiatry 59, 1215–1224 (2018). [DOI] [PubMed] [Google Scholar]
  • 15.Tamnes CK et al. Development of the Cerebral Cortex across Adolescence: A Multisample Study of Inter-Related Longitudinal Changes in Cortical Volume, Surface Area, and Thickness. Journal of Neuroscience 37, 3402–3412 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Thorup AAE et al. The Danish High-Risk and Resilience Study—VIA 15 – A Study Protocol for the Third Clinical Assessment of a Cohort of 522 Children Born to Parents Diagnosed With Schizophrenia or Bipolar Disorder and Population-Based Controls. Front Psychiatry 13, 809807 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Collins MA et al. Accelerated cortical thinning precedes and predicts conversion to psychosis: The NAPLS3 longitudinal study of youth at clinical high-risk. Mol Psychiatry 28, 1182 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Cannon TD et al. Progressive reduction in cortical thickness as psychosis develops: a multisite longitudinal neuroimaging study of youth at elevated clinical risk. Biol Psychiatry 77, 147–157 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Pantelis C et al. Neuroanatomical abnormalities before and after onset of psychosis: a cross-sectional and longitudinal MRI comparison. Lancet 361, 281–288 (2003). [DOI] [PubMed] [Google Scholar]
  • 20.Merritt K, Luque Laguna P, Irfan A & David AS Longitudinal Structural MRI Findings in Individuals at Genetic and Clinical High Risk for Psychosis: A Systematic Review. Front Psychiatry 12, 49 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.van Os J, Linscott RJ, Myin-Germeys I, Delespaul P & Krabbendam L A systematic review and meta-analysis of the psychosis continuum: evidence for a psychosis proneness-persistence-impairment model of psychotic disorder. Psychol Med 39, 179–195 (2009). [DOI] [PubMed] [Google Scholar]
  • 22.Karcher NR et al. Psychotic-like Experiences and Polygenic Liability in the Adolescent Brain Cognitive Development Study. Biol Psychiatry Cogn Neurosci Neuroimaging 7, 45–55 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Karcher NR, Schiffman J & Barch DM Environmental Risk Factors and Psychotic-like Experiences in Children Aged 9–10. J Am Acad Child Adolesc Psychiatry 60, 490–500 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Glenthøj LB, Hjorthøj C, Kristensen TD, Davidson CA & Nordentoft M The effect of cognitive remediation in individuals at ultra-high risk for psychosis: a systematic review. NPJ Schizophr 3, 20 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Karcher NR et al. Persistent and distressing psychotic-like experiences using adolescent brain cognitive developmentSM study data. Mol Psychiatry 1–12 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Karcher NR, Merchant J, Pine J & Kilciksiz CM Cognitive Dysfunction as a Risk Factor for Psychosis. Curr Top Behav Neurosci 63, 173–203 (2023). [DOI] [PubMed] [Google Scholar]
  • 27.Sekar A et al. Schizophrenia risk from complex variation of complement component 4. Nature 2016 530:7589 530, 177–183 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Shaw P et al. Neurodevelopmental Trajectories of the Human Cerebral Cortex. The Journal of Neuroscience 28, 3586 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Fusar-Poli P et al. Deconstructing vulnerability for psychosis: Meta-analysis of environmental risk factors for psychosis in subjects at ultra high-risk. European Psychiatry 40, 65–75 (2017). [DOI] [PubMed] [Google Scholar]
  • 30.Linscott RJ & van Os J An updated and conservative systematic review and meta-analysis of epidemiological evidence on psychotic experiences in children and adults: on the pathway from proneness to persistence to dimensional expression across mental disorders. Psychol Med 43, 1133–1149 (2013). [DOI] [PubMed] [Google Scholar]
  • 31.Masten AS & Cicchetti D Developmental cascades. Dev Psychopathol 22, 491–495 (2010). [DOI] [PubMed] [Google Scholar]
  • 32.Myin-Germeys I & van Os J Stress-reactivity in psychosis: Evidence for an affective pathway to psychosis. Clin Psychol Rev 27, 409–424 (2007). [DOI] [PubMed] [Google Scholar]
  • 33.Meyer MN et al. Wrestling with Social and Behavioral Genomics: Risks, Potential Benefits, and Ethical Responsibility. Hastings Center Report 53, (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Martin AR et al. Clinical use of current polygenic risk scores may exacerbate health disparities. Nat Genet 51, 584–591 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Karcher NR & Barch DM The ABCD study: understanding the development of risk for mental and physical health outcomes. Neuropsychopharmacology 2020 46:1 46, 131–142 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Barch DM et al. Demographic and mental health assessments in the adolescent brain and cognitive development study: Updates and age-related trajectories. Dev Cogn Neurosci 52, 101031 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Garavan H et al. Recruiting the ABCD sample: Design considerations and procedures. Dev Cogn Neurosci 32, 16–22 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Karcher NR, Modi H, Kochunov P, Gao S & Barch DM Regional Vulnerability Indices in Youth With Persistent and Distressing Psychoticlike Experiences. JAMA Netw Open 6, e2343081–e2343081 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Chapman LJ, Chapman JP, Kwapil TR, Eckblad M & Zinser MC Putatively psychosis-prone subjects 10 years later. J Abnorm Psychol 103, 171–183 (1994). [DOI] [PubMed] [Google Scholar]
  • 40.Weintraub S et al. Cognition assessment using the NIH Toolbox. Neurology 80, S54–64 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Fischl B, Sereno MI & Dale AM Cortical surface-based analysis. II: Inflation, flattening, and a surface-based coordinate system. Neuroimage 9, 195–207 (1999). [DOI] [PubMed] [Google Scholar]
  • 42.Chen CH et al. Hierarchical genetic organization of human cortical surface area. Science (1979) 335, 1634–1636 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Fischl B & Dale AM Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proc Natl Acad Sci U S A 97, 11050–11055 (2000). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Hagler DJ Jr et al. Image processing and analysis methods for the Adolescent Brain Cognitive Development Study. Neuroimage 202, 116091 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Beer JC et al. Longitudinal ComBat: A method for harmonizing longitudinal multi-scanner imaging data. Neuroimage 220, 117129 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Trubetskoy V et al. Mapping genomic loci implicates genes and synaptic biology in schizophrenia. Nature 2022 604:7906 604, 502–508 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Bigdeli TB et al. Contributions of common genetic variants to risk of schizophrenia among individuals of African and Latino ancestry. Mol Psychiatry 25, 2455–2467 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Ge T, Chen CY, Ni Y, Feng YCA & Smoller JW Polygenic prediction via Bayesian regression and continuous shrinkage priors. Nature Communications 2019 10:1 10, 1–10 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Martin AR et al. Clinical use of current polygenic risk scores may exacerbate health disparities. Nat Genet 51, 584–591 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Rosenberg JM, Beymer PN, Anderson DJ, Lissa C. j. van & Schmidt JA. tidyLPA: An R Package to Easily Carry Out Latent Profile Analysis (LPA) Using Open-Source or Commercial Software. J Open Source Softw 3, 978 (2019). [Google Scholar]
  • 51.Scrucca L, Fop M, Murphy TB & Raftery AE mclust 5: Clustering, Classification and Density Estimation Using Gaussian Finite Mixture Models. R J 8, 289 (2016). [PMC free article] [PubMed] [Google Scholar]
  • 52.Chakrabarti A & Ghosh JK AIC, BIC and Recent Advances in Model Selection. Philosophy of Statistics: Volume 7 in Handbook of the Philosophy of Science 7, 583–605 (2011). [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Methods and Results
NIHMS2105627-supplement-Supplemental_Methods_and_Results.pdf (13.7MB, pdf)

Data Availability Statement

Data used in the preparation of this article were obtained from the Adolescent Brain Cognitive Development (ABCD) Study (https://abcdstudy.org), held in the NIMH Data Archive (NDA). This is a multisite, longitudinal study designed to recruit more than 10,000 children aged 9–10 and follow them over 10 years into early adulthood. The ABCD data repository grows and changes over time. The ABCD data used in this report came from https://nda.nih.gov/study.html?id=2313.

RESOURCES