Longitudinal hippocampal subfield development associated with psychotic experiences in young people

Subthreshold psychotic experiences (PE) affect a significant portion of young people, increasing their risk for psychosis and other psychopathology. While individuals with PE may not meet criteria for a formal psychotic disorder, they often present with subtle neurocognitive impairments and moderate neuroanatomical changes, notably in the hippocampus. The hippocampus plays a critical role in various cognitive functions and the stress response, all frequently impaired in psychosis. Previous research has implicated hippocampal abnormalities in psychosis, often focusing on whole hippocampal volume. However, recent advancements in neuroimaging allow for investigation of smaller, functionally distinct hippocampal subfields (CA1-4, dentate gyrus (DG), subiculum, presubiculum). Meta-analyses suggest that focusing solely on whole hippocampal volume may overlook subtle subfield deficits. Studies indicate that CA1 volume deficits are prominent in those with prodromal psychosis, potentially becoming more widespread with disease progression. This study aimed to investigate the longitudinal developmental trajectory of hippocampal subfields in young people reporting PE, providing insights into the mechanisms underlying psychotic psychopathology and vulnerability without the confounding factors of formal diagnosis or medication.

The literature extensively documents hippocampal volumetric reductions across the psychosis spectrum, with a growing interest in the localization of these reductions within the hippocampal subfields. Deficits in the CA1 subfield, in particular, have been strongly implicated in the neuropathophysiology of psychotic disorders. Studies have shown associations between hippocampal size deficits and the expression of schizotypy traits in adolescents. These findings align with studies of individuals in the prodromal stages of psychosis, suggesting hippocampal reductions may be a marker of psychosis vulnerability. Early research focused on whole hippocampal structures, but with improvements in neuroimaging, finer-grained analyses of hippocampal subfields have become possible, revealing subfield-specific abnormalities. Existing research largely comprises cross-sectional studies of treatment-seeking individuals, limiting understanding of longitudinal trajectories of hippocampal subfields in the absence of formal diagnosis or medication. This gap underscores the need for longitudinal investigations of adolescents and young people reporting PE to better understand the development of these regions without the confounding influence of illness or medication.

This longitudinal study used data from the Adolescent Brain Development (ABD) study, which initially recruited 211 young people (aged 11-13). The sample was enriched for adolescents with at least two points on the Adolescent Psychotic Symptom Screener. Neuroimaging (structural MRI) was conducted at three time points over approximately six years. PE classification was determined by expert consensus at each timepoint. The final sample included 78 participants (33 with PE at least once, 45 controls). Hippocampal subfield volumes (CA1, CA2/3, CA4/DG, presubiculum, subiculum) were extracted using FreeSurfer image analysis software. Linear Mixed Effects (LME) models were used to analyze longitudinal volumetric data. Fixed effects included group, Group × Time interaction, intracranial volume (ICV), time, age at baseline, gender, handedness, and DSM-5 lifetime diagnosis. Subjects were treated as random effects, accounting for within-person correlations and individual variations. A top-down model selection process was employed to determine the best-fitting models for each ROI. False discovery rate (FDR) procedures were used to correct for multiple comparisons. For any subfields showing significant uncorrected differences, associations between subfield volume and Structured Interview for Psychosis-Risk Syndromes (SIPS) scores were also explored using LME.

Before FDR correction, the right CA1 and CA2/3 subfields showed significant Group × Time interactions: volumes increased in the control group and decreased in the PE group over time (large effect sizes). Before correction, the PE group also displayed smaller right subiculum and left presubiculum volumes compared to controls (moderate effect sizes). However, none of these effects survived FDR correction. No significant associations were found between symptom severity (SIPS scores) and subfield volumes. The PE group had a significantly higher proportion of males and a significantly higher rate of lifetime DSM-5 diagnoses compared to the control group. No significant differences were observed between groups for the number of scans, age, handedness, socioeconomic status, or childhood adversity.

This study found evidence of differential developmental trajectories in specific hippocampal subfields between adolescents with and without PE, although these effects did not consistently survive correction for multiple comparisons. The moderate to large effect sizes observed suggest the differences could be practically significant or may achieve statistical significance with a larger sample. The findings do not solely implicate the CA1 subfield as predicted, but rather show more widespread involvement, particularly in the right CA2/3 and subiculum complex. These results align with some findings in individuals with diagnosed psychotic disorders, suggesting shared underlying neurobiological mechanisms. The lack of association between symptom severity and hippocampal subfield volumes might be due to the subtle nature of volumetric differences in this non-clinical sample. Developmental delay could underlie the observed subfield growth trajectories in the PE group, possibly due to disordered synaptic pruning and gray matter maturation during adolescence. Future research should investigate potential neurophysiological differences relating to subjective appraisals of psychotic experiences and consider spectral approaches to categorise the range of psychotic experiences to better understand those who transition to persistent illness versus those who do not.

This longitudinal study provides novel insights into hippocampal subfield development in adolescents with subthreshold psychotic experiences. While findings did not consistently survive correction for multiple comparisons, the observed effect sizes highlight the potential significance of alterations in the CA1, CA2/3, subiculum, and presubiculum regions in the pathophysiology of psychosis. Future research with larger samples is needed to confirm these findings, explore the relationship between symptom severity and subfield volumes more comprehensively, and examine the specific role of these subfields in the transition to persistent psychotic illness.

The study’s limitations include limited data on symptom severity over time, lifetime DSM-5 diagnoses, and childhood adversity, potentially leading to an underestimation of the relationship between these factors and brain morphology. The binary categorization of PE (definite PE vs. no PE) limits a more nuanced spectral analysis of psychotic experiences. The accuracy of the FreeSurfer version used has been questioned in previous literature, which should be considered in future studies. Despite the use of LME models that accommodates missing data, attrition still poses a potential limitation.