Unraveling the Neural Landscape of Mental Disorders using Double Functional Independent Primitives (dFIPs)

Najme Soleimani, Armin Iraji, Godfrey D. Pearlson, Adrian Preda, Vince D. CalhounbioRxiv 2024 · DOI

TL;DR

This paper introduces "Double Functional Independent Primitives" (dFIPs), a novel hierarchical method to analyze resting-state fMRI data and uncover distinct neural connectivity "fingerprints" for major psychiatric disorders like schizophrenia, autism, bipolar disorder, and depression. By decomposing functional network connectivity (FNC) into these primitives, the study reveals unique and shared patterns that could lead to more precise diagnoses and personalized treatments.

What Problem Does This Paper Address?

Mental illnesses are complex and often overlap, making diagnosis and treatment challenging. While neuroimaging techniques like resting-state fMRI (rs-fMRI) offer insights into brain function, precisely identifying the neural underpinnings of these disorders remains difficult. Traditional methods like seed-based functional connectivity (FC) can be limited, and even advanced techniques like Independent Component Analysis (ICA) can struggle with the complexity and overlap of brain signals. The field needs more sensitive and data-driven approaches to dissect the intricate functional network connectivity (FNC) patterns that characterize different mental health conditions.

Methods

The study proposes a two-step hierarchical approach:

  1. NeuroMark Pipeline: This step utilizes a spatially constrained ICA method (NeuroMark) to extract 53 intrinsic connectivity networks (ICNs) from rs-fMRI data for each individual. These ICNs represent distinct brain networks with synchronized activity.
  2. dFIP Extraction: The ICN time series are used to compute individual functional network connectivity (FNC) matrices (53x53). A second ICA is then applied to these vectorized FNC matrices from a large cohort (5805 individuals) to extract 15 "Double Functional Independent Primitives" (dFIPs). These dFIPs represent maximally independent patterns of FNC that covary across subjects, essentially acting as a "basis set" for functional connectivity.

The researchers then used these dFIPs to:

  • Calculate "reference" FNC patterns for each disorder group and their healthy controls.
  • Perform regression analyses to identify which dFIPs best explain these reference patterns and distinguish between patient groups and controls.
  • Visualize disorder-specific FNC profiles and validate their distinctiveness using cross-validation.

Datasets

The study utilized a large, multi-site dataset comprising 5805 individuals from seven clinical studies:

  • Bipolar-Schizophrenia Network on Intermediate Phenotypes (B-SNIP)
  • Center for Biomedical Research Excellence (COBRE)
  • Function Biomedical Informatics Research Network (FBIRN)
  • Maryland Psychiatric Research Center (MPRC)
  • Chinese schizophrenia dataset
  • Autism Brain Imaging Data Exchange (ABIDE)
  • Chinese MDD dataset

The dataset included patients diagnosed with schizophrenia (SCZ), autism spectrum disorder (ASD), bipolar disorder (BPD) with psychosis, and major depressive disorder (MDD), along with healthy controls.

Key Insights and Ideas

  • dFIPs as "Neural Fingerprints": The core idea is that combinations of dFIPs can serve as unique identifiers for different mental disorders, much like a fingerprint. This moves beyond identifying single network disruptions to understanding how multiple network interactions combine.
  • Hierarchical Decomposition: The two-step ICA approach allows for a more robust and nuanced decomposition of complex FNC data, first identifying intrinsic brain networks and then identifying the fundamental patterns of connectivity between these networks.
  • Shared and Specific Signatures: The study found that while some dFIP patterns are common across different disorders (suggesting shared underlying mechanisms), others are highly specific, allowing for differentiation between conditions.
  • Cerebellum's Role: The cerebellum appears to play a significant role, with heightened connectivity in SCZ and ASD, but reduced cerebellar-subcortical connectivity across multiple disorders.
  • Potential for Biomarkers: The distinct dFIP patterns identified hold promise as neurobiological markers for more precise diagnosis and the development of targeted interventions.

Key Results

  • Disorder-Specific Signatures: The study successfully identified distinct dFIP patterns associated with SCZ, ASD, BPD, and MDD. For example, SCZ showed heightened connectivity in sensorimotor, auditory, and visual networks, while ASD and BPD showed reduced cerebellar-subcortical connectivity.
  • Transdiagnostic Overlap: A notable correlation was found between the reference patterns of SCZ and ASD, suggesting shared neurobiological mechanisms or pathways.
  • Validation: A 10-fold cross-validation strategy confirmed that the derived disorder-specific FNC profiles were most similar to the training FNCs of the same disorder, indicating their specificity.
  • Significant dFIPs: Several dFIPs showed significant differences in component loadings between patient groups and controls, with specific dFIPs (e.g., #7, #14) highlighted as particularly discriminative for certain disorders. For instance, dFIP 14 was significant for SCZ and ASD but less so for BPD and MDD.

Limitations and Future Work

  • Cross-Sectional Data: The study's cross-sectional design limits the ability to infer causality. Longitudinal studies are needed to track changes over time and determine directionality.
  • Task-Based vs. Resting-State: The reliance on resting-state fMRI means insights into how these networks function during specific cognitive tasks are missing. Future work could integrate task-based fMRI.
  • Data Variability: Differences in scanner calibration and scanning protocols across sites can introduce variability.
  • Future Directions: The authors suggest integrating multimodal imaging data (PET, dMRI) for a more comprehensive view, studying risk in non-clinical samples, predicting transitions to psychosis in high-risk individuals, and exploring heterogeneity including medication effects and dimensional measures of psychosis.

Summary updated

Abstract

Mental illnesses extract a high personal and societal cost, and thus explorations of the links between mental illness and functional connectivity in the brain are critical. Investigating major mental illnesses, believed to arise from disruptions in sophisticated neural connections, allows us to comprehend how these neural network disruptions may be linked to altered cognition, emotional regulation, and social interactions. Although neuroimaging has opened new avenues to explore neural alterations linked to mental illnesses, the field still requires precise and sensitive methodologies to inspect these neural substrates of various psychological disorders. In this study, we employ a hierarchical methodology to derive double functionally independent primitives (dFIPs) from resting state functional magnetic resonance neuroimaging data (rs-fMRI). These dFIPs encapsulate canonical overlapping patterns of functional network connectivity (FNC) within the brain. Our investigation focuses on the examination of how combinations of these dFIPs relate to different mental disorder diagnoses. The central aim is to unravel the complex patterns of FNC that correspond to the diverse manifestations of mental illnesses. To achieve this objective, we used a large brain imaging dataset from multiple sites, comprising 5805 total individuals diagnosed with schizophrenia (SCZ), autism spectrum disorder (ASD), bipolar disorder (BPD), major depressive disorder (MDD), and controls. The key revelations of our study unveil distinct patterns associated with each mental disorder through the combination of dFIPs. Notably, certain individual dFIPs exhibit disorder-specific characteristics, while others demonstrate commonalities across disorders. This approach offers a novel, data-driven synthesis of intricate neuroimaging data, thereby illuminating the functional changes intertwined with various mental illnesses. Our results show distinct signatures associated with psychiatric disorders, revealing unique connectivity patterns such as heightened cerebellar connectivity in SCZ and sensory domain hyperconnectivity in ASD, both contrasted with reduced cerebellar-subcortical connectivity. Utilizing the dFIP concept, we pinpoint specific functional connections that differentiate healthy controls from individuals with mental illness, underscoring its utility in identifying neurobiological markers. In summary, our findings delineate how dFIPs serve as unique fingerprints for different mental disorders.
BibTeX
@misc{soleimani2024unraveling,
  author = {Najme Soleimani and Armin Iraji and Godfrey D. Pearlson and Adrian Preda and Vince D. Calhoun},
  title = {{Unraveling the Neural Landscape of Mental Disorders using Double Functional Independent Primitives (dFIPs)}},
  year = {2024},
  doi = {10.1101/2024.08.01.606076},
  howpublished = {bioRxiv},
}
Figure