Brain-age in psychosis: exploring its links with clinical characteristics, genetic liability, and subthreshold experiences

: (Alternative Format Thesis)

Abstract

Magnetic resonance imaging (MRI) studies have previously identified cortical and subcortical grey-matter alterations among those with schizophrenia and related psychotic disorders (SCZ), and at-risk groups. The extent to which those apparent structural brain abnormalities reflect an aberrant developmental and/or ageing trajectory remains elusive. Recent advancements combine structural (T1-weighted) MRI and multivariate machine learning techniques to estimate the underlying biological age of the brain, an approach commonly known as the “brain age” paradigm. The discrepancy between estimated brain age and chronological age (brain-predicted age difference, brain-PAD) is a putative index of quantifying an individual’s deviation from normative brain ageing or maturation. The first aim of the current thesis was to investigate cross-sectional associations between brain-PAD and SCZ status as well as its clinical characteristics across adulthood. Brain-PAD was further explored as a vulnerability marker in the context of SCZ by investigating its association with polygenic risk for SCZ (SCZ-PRS) or subclinical psychotic experiences (PEs) in late adolescence/early adulthood.

First, SCZ case-control differences in brain-PAD are presented using the largest (meta-analytical) study to date via the ENIGMA consortium, covering almost the entire adult lifespan (K = 26 cohorts; NSCZ = 2803; Ncontrols = 2598; age range: 18-73 years). SCZ was associated with a moderately higher brain-PAD relative to unaffected controls, potentially reflecting advanced structural brain ageing. Among SCZ patients, there was little to no evidence for an association between brain-PAD and clinical characteristics including positive/negative symptom severity, age of onset, length of illness, and antipsychotic use.

Next, the second study utilised a recall-by-genotype design nested within the ALSPAC population-based birth cohort to examine the association between SCZ-PRS and brain-PAD among unaffected young adults (Nhigh-PRS = 96; Nlow-PRS = 93; age range: 21-24 years). There was little evidence for an association, suggesting that brain-PAD may not be a vulnerability marker of common genetic risk for SCZ.

Last, the association between PEs and brain-PAD was examined using population-based, case-control sample of older adolescents nested within the ALSPAC birth cohort (NPE = 117, Nno-PE = 115, aged 19-21 years). There was some weak evidence for a negative (i.e., lower) brain-PAD among those with PEs relative to those without, indicative of delayed structural brain maturation.

Overall, the current thesis provides further evidence for older structural brain age in SCZ, yet this apparent deviation in brain ageing or maturation does not appear to be linked to clinical characteristics or SCZ polygenic risk. The thesis also provides some weak evidence for a younger brain age among those with PEs in late adolescence, suggesting that whether psychotic psychopathology is associated with higher or lower brain-PAD may partly depend on developmental timing and/or sub-clinical state. Key limitations and directions for future research including recent methodological developments are discussed.

Date of Award	25 Jun 2025
Original language	English
Awarding Institution	University of Bath
Sponsors	GW4 BioMed Medical Research Council
Supervisor	Esther Walton (Supervisor), Tom Freeman (Supervisor), Doretta Caramaschi (Supervisor) & Stanley Zammit (Supervisor)