Individualized brain biomarkers of late life depression: contributions to heterogeneity and resilience

Individualized brain biomarkers of late life depression: contributions to heterogeneity and resilience
Project Summary/ Abstract
Approximately 10% of people aged 60+ suffer from depression. Such late-life depression (LLD) is linked
to broader adverse health outcomes such as stroke and dementia. Many brain correlates of LLD have
been reported, but each explains only a small amount of interindividual variance in LLD symptoms, likely
due to a many-to-one mechanistic mapping in which multiple neural mechanisms contribute to similar
symptoms. Heterogeneity in clinical presentation arises from between-patient differences in acute severity,
symptoms, chronicity, and age of onset. Few patients are matched in all clinical domains and therefore
heterogeneity in conventional research samples is often unavoidable. Over and above clinical
heterogeneity, additional risk/resilience factors may alter the experience of LLD at the individual level.
Population data from the UK Biobank offers an unprecedented opportunity to fully disentangle clinical
heterogeneity by curating clinically homogeneous subject groups. We will distinguish brain profiles,
longitudinal changes, and resilience/vulnerability factors that are uniquely linked to LLD clinical domains
of: acute severity, mood symptoms, somatic symptoms, chronicity, and late onset LLD. Sixty brain
measures associated with LLD, including cortical thickness, gray matter volume, fractional anisotropy,
white matter hyperintensities, and resting state connectivity will be used for all aims. In Aim 1, we will curate
five homogeneous groups of UKB subjects with shared clinical presentation, focusing on: late onset LLD,
acute severity, lifetime chronicity, mood symptoms, and somatic symptoms. Using normative models
trained on never-depressed UKB subjects, we will distinguish normative brain deviation profiles associated
with these different domains of clinical heterogeneity. In Aim 2, we will curate new groups of UKB subjects
with shared longitudinal changes in acute severity, chronicity, mood symptoms, or somatic symptoms to
test whether changes over time in brain measures are linked to longitudinal changes in clinical
presentation. This aim therefore offers an independent validation of aim 1 and differentiates between state
and trait markers of LLD. In Aim 3, we will test the hypothesis that cumulative environmental and
psychological stressors alter the experience of LLD at the individual level. We will obtain individual-specific
statistical estimates of resilience/vulnerability based on the difference between predicted and actual
depression scores (‘brain depression gap’). The brain depression gap will be linked to stressors separately
in each homogeneous subject group (same as aim 1) to determine factors that promote LLD resilience or
vulnerability.
Public health significance: this proposal is expected to move the field closer to a full understanding of
LLD heterogeneity by combining theory-driven subject groups with data-driven population prediction
models. Gaining a better understanding of LLD heterogeneity may inform improved strategies for treatment
and prevention of LLD, which could positively influence broader health outcomes in older age.

Grant Number: 5R01MH128286-04
NIH Institute/Center: NIH
Principal Investigator: Janine Bijsterbosch