Neuroimaging alterations and relapse in early-stage psychosis

Marina Mihaljević, Anisha Nagpal, Semra Etyemez, Zui Narita, Anna Ross, Rebecca Schaub, Nicola G. Cascella, Jennifer M. Coughlin, Gerald Nestadt, Frederik C. Nucifora, Thomas W. Sedlak, Koko Ishizuka, Vince D. Calhoun, Andréia V. Faria, Kun Yang, Akira SawabioRxiv 2021 · DOI

TL;DR

This study investigates brain connectivity differences in early-stage psychosis patients who have experienced relapse, hypothesizing this subgroup is more homogenous. By analyzing resting-state functional connectivity (FC), they identified specific brain regions and connections associated with relapse, offering potential biomarkers for understanding and potentially predicting this critical aspect of the disorder.

What Problem Does This Paper Address?

Early-stage psychosis is a complex condition with a varied trajectory, and relapse—a worsening of symptoms after a period of improvement—is a major determinant of poorer long-term outcomes. Identifying the neurobiological underpinnings of relapse is crucial for developing better interventions and understanding the disease's progression. However, the broad category of "early-stage psychosis" may encompass diverse underlying biological signatures. This paper tackles the problem of identifying more specific neuroimaging markers by focusing on a subgroup of patients who have experienced relapse, hypothesizing that this subgroup might exhibit more consistent and distinct brain alterations compared to the broader patient population.

Methods

The researchers conducted a cross-sectional study using a retrospective approach. They recruited 85 patients with early-stage psychosis and 94 healthy controls. Relapse was defined by hospitalization due to psychotic symptom exacerbation, a common proxy in the literature. Resting-state functional connectivity (rs-fMRI) was used to examine brain network activity. Specifically, they acquired rs-fMRI data, processed it using MRICloud for segmentation and artifact correction, and then calculated seed-by-seed correlation matrices to derive 3,003 pairwise resting-state z-correlations between 78 brain regions of interest (ROIs). Statistical analyses included T-tests, chi-squared tests, ANCOVA (controlling for age, sex, race, handedness, smoking status, and framewise displacement), and correlation analyses to assess the impact of antipsychotic medication and duration of illness. Multiple comparison correction was applied using the Benjamini-Hochberg procedure, with q-values < 0.05 considered significant.

Datasets

The study utilized data from a cohort of 85 patients with early-stage psychosis and 94 healthy controls. The data was collected retrospectively, with medical records used to identify patients who had experienced relapse. No new benchmark datasets were introduced.

Key Insights and Ideas

  • Relapse Subgroup Homogeneity: The core hypothesis that patients who relapse are a more homogenous group in terms of neurobiological signatures was supported. This suggests that focusing on this specific subgroup can yield more robust and interpretable findings than studying the entire patient population.
  • Distinct FC Signatures: The study identified a significant number of altered functional connections (FCs) in patients with early-stage psychosis, with a particularly large set of these alterations being unique to or more pronounced in the relapse group.
  • Key Brain Regions for Relapse: The thalamus, precentral gyrus, and dorsal anterior cingulate cortex emerged as critical brain regions implicated in relapse, showing the highest number of significant FC alterations in the relapse group. These regions are known to be involved in various cognitive and emotional processes relevant to psychosis.
  • Independence from Medication/Illness Duration: Importantly, the identified FC alterations associated with relapse were not significantly correlated with antipsychotic medication dosage or duration of illness, suggesting these changes might reflect intrinsic vulnerability or consequences of relapse itself, rather than being solely medication effects.

Key Results

  • The study identified 131 significant FCs in the relapse group (R group) compared to healthy controls (HCs), and 60 significant FCs in the entire patient group compared to HCs.
  • Of the 60 FCs altered in the entire patient group, 51 were also significantly altered in the R group, indicating the relapse group's substantial contribution to the overall patient group's findings.
  • The R group showed significantly larger effect sizes and smaller variations in these 51 overlapping FCs compared to the entire patient group, supporting the hypothesis of greater homogeneity.
  • The top three brain regions with the most significant FCs in the R group were the thalamus, precentral gyrus, and dorsal anterior cingulate cortex, accounting for 64% of these significant FCs.

Limitations and Future Work

The authors acknowledge that hospitalization, while a practical proxy for relapse, may not perfectly capture all nuances of symptom exacerbation and could be influenced by other factors like suicidality or family pressure. They also note that this was a cross-sectional study, which limits the ability to establish causality or track changes over time.

Future work could involve longitudinal studies to better understand the temporal relationship between brain alterations and relapse, and to explore whether these identified brain changes reflect intrinsic vulnerability to relapse or are consequences of the relapse episodes themselves. The authors suggest that their cross-sectional approach can serve as a prototype for larger, multi-institutional projects to further investigate these brain alterations in early-stage psychosis.

Summary updated

Abstract

ABSTRACT Recent reports have indicated that the occurrence of symptom exacerbation in early-stage psychosis could result in brain changes. Such a symptom exacerbation is frequently called relapse, which underlies a poorer disease outcome. Thus, it is important to identify neuroimaging alterations that are specifically seen in patients who experience relapse in early-stage psychosis. We hypothesized that this specific patient group may be more homogenous in disease-associated signatures likely to be linked to relapse, compared with the overall patient group. Such sub-stratified patient group and neuroimaging signatures would be useful for the biological understanding of relapse. To address this goal, we conducted a cross-sectional study (85 patients with early-stage psychosis and 94 healthy controls) with the use of medical records in a retrospective manner. To define the specific sub-group with the past experience of relapse, we used hospitalization due to psychotic symptom exacerbation, according to many publications that used this factor as a proxy for relapse. By examining resting-state functional connectivity (FC) for the study subjects, we validated our hypothesis and defined 131 FCs possibly associated with relapse. Through these studies, 3 brain regions (the thalamus, precentral gyrus, and dorsal anterior cingulate cortex) were underscored.
BibTeX
@misc{mihaljevic2021neuroimaging,
  author = {Marina Mihaljević and Anisha Nagpal and Semra Etyemez and Zui Narita and Anna Ross and Rebecca Schaub and Nicola G. Cascella and Jennifer M. Coughlin and Gerald Nestadt and Frederik C. Nucifora and Thomas W. Sedlak and Koko Ishizuka and Vince D. Calhoun and Andréia V. Faria and Kun Yang and Akira Sawa},
  title = {{Neuroimaging alterations and relapse in early-stage psychosis}},
  year = {2021},
  doi = {10.1101/2021.12.10.472143},
  howpublished = {bioRxiv},
}
Figure