Brain iron distribution in transdiagnostic mental health burden

Abstract

Background: Psychiatric diseases are increasingly understood as a spectrum or continuous phenomena, ranging from healthy to severely affected, yet neurobiological correlates for these dimensions remain elusive. Regional alteration of iron levels in the central nervous system may reflect neuropathological processes that result in significant impairment of cognitive and behavioral functions. The aim of this study was to characterize brain iron distribution in individuals with common clinical psychiatric disorders and offspring of individuals affected by these disorders.

Methods: R2* magnetic resonance imaging (MRI) based multi-parameter mapping (MPM) algorithm was used to assess regional brain iron distribution in a) individuals diagnosed with bipolar disorder (BD), borderline personality disorder (BPD), attention-deficit/hyperactivity disorder (ADHD), and offspring (n = 80, age = 23 ± 7 y., 61% females), and b) healthy controls and offspring controls (n = 43, age = 25 ± 9, 56% females). Hierarchical cluster analysis (HCA) was used to identify group-divisive patterns of regional brain iron distribution.

Results: Three distinct clusters of regional brain iron distribution were found, differentiating patients and patient offspring from healthy controls and control offspring with 94% sensitivity (OR: 8.9; p = 0.038). Secondary analysis revealed no significant difference in brain iron distribution among BD, BPD, and ADHD diagnoses.

Conclusion: Our finding of a characteristic brain iron distribution pattern in individuals diagnosed with BD, BPD, and ADHD and offspring of diagnosed individuals supports that brain iron patterns may serve as neurobiological correlates for psychiatric disorders conceptualized as spectrum conditions. Further, longitudinal studies are needed to confirm whether pattern analysis of brain iron distribution may represent a transdiagnostic biomarker for a better understanding of underlying neuropathology in psychiatric disorders.

Keywords: brain iron; hierarchical clustering; iron homeostasis; multi‐parameter mapping; psychiatric disorders; transdiagnostic biomarkers.

Figure 1

Hierarchical clustering dendrogram. Clusters were derived using the DIANA algorithm, with data assigned into three clusters based on MPM values from distinct brain regions. The X‐axis represents subjects. Cluster 1 = purple, N = 15; Cluster 2 = green, N = 82; Cluster 3 = yellow, N = 26.

Figure 2

Cluster‐specific iron levels across brain regions. Clusters were derived using the DIANA hierarchical clustering algorithm, with data assigned into three clusters based on R2* values from distinct brain regions. Blue indicates lower R2* values, yellow and red indicate higher R2* values.

Figure 3

Iron level differences between clusters. Boxplots illustrating the R2* derived brain iron (ppm) for individual brain regions. Cluster 1 = purple. Cluster 2 = green, Cluster 3 = yellow.

Figure 4

Cluster‐wise group composition. Absolute number and cumulative percentage of patients/offspring and controls/control offspring within each cluster. To calculate differences between clusters, patients and patient offspring were grouped as one level of a binary factor, while controls and control offspring constituted the other level.

Figure 5

Iron differences between clusters. Median differences of brain iron (ppm) per region between Cluster 2 and Cluster 1 are shown.

Figure 6

Frequency of pathologies within clusters. ADHD, attention deficit/hyperactivity disorder; BD, bipolar disorder; BPD, borderline personality disorder; Offspr, offspring.