Background: Mental illnesses extract personal and societal costs, leading to significant challenges in cognitive function, emotional regulation, and social behavior. These disorders are thought to result from disruptions in how different brain regions communicate with each other. Despite advances in neuroimaging, current methods are not always precise enough to fully understand the complexity of these disruptions. More advanced approaches are needed to better identify and characterize the specific brain network alterations linked to different psychiatric conditions.

Methods: We employed a hierarchical approach to derive Double Functionally Independent Primitives (dFIPs) from resting-state functional magnetic resonance imaging (rs-fMRI) data. dFIPs represent independent patterns of functional network connectivity (FNC) across the brain. Our study utilized a large multi-site dataset comprising 5805 individuals diagnosed with schizophrenia (SCZ), autism spectrum disorder (ASD), bipolar disorder (BPD), major depressive disorder (MDD), and healthy controls. We analyzed how combinations of dFIPs differentiate psychiatric diagnoses.

Results: Distinct dFIP patterns emerged for each disorder. Schizophrenia was characterized by heightened cerebellar connectivity and reduced cerebellar-subcortical connectivity. In ASD, sensory domain hyperconnectivity was prominent. Some dFIPs displayed disorder-specific connectivity patterns, while others exhibited commonalities across multiple conditions. These findings underscore the utility of dFIPs in revealing neural connectivity alterations unique to each disorder, serving as unique fingerprints for different mental disorders.

Conclusions: Our study demonstrates that dFIPs provide a novel, data-driven method for identifying disorder-specific functional connectivity patterns in psychiatric conditions. These distinct neural signatures offer potential biomarkers for mental illnesses, contributing to a deeper understanding of the neurobiological underpinnings of these disorders.

Keywords: Double Functionally Independent Primitive (dFIP); Functional Network Connectivity (FNC); Independent Component Analysis (ICA); Mental Disorders.

Copyright © 2025. Published by Elsevier Inc.

背景：精神疾病对个人和社会产生成本，导致认知功能、情绪调节和社交行为方面的重大挑战。这些障碍被认为是由于大脑不同区域之间的沟通中断所引起的。尽管神经影像学技术取得了进展，但目前的方法并不总是足够精确，无法完全理解这些中断的复杂性。需要更先进的方法来更好地识别和描述与不同精神疾病相关的特定脑网络改变。

方法：我们采用分层方法从静息态功能磁共振成像（rs-fMRI）数据中导出双功能性独立原语（dFIP）。dFIP代表大脑中独立的功能连接模式。我们的研究使用了一个大型多中心数据集，包括5805名被诊断为精神分裂症（SCZ）、自闭症谱系障碍（ASD）、双相情感障碍（BPD）、重性抑郁障碍（MDD）和健康对照的个体。我们分析了dFIP组合如何区分精神病诊断。

结果：每个疾病的dFIP模式各不相同。精神分裂症的特点是小脑连接增强和小脑-皮层下连接减弱。在自闭症中，感觉领域高连接性尤为突出。一些dFIP显示出特定于某种疾病的功能连接模式，而其他dFIP则在多种条件下表现出共同点。这些发现强调了dFIP揭示每个障碍独特神经连接改变的实用性，为不同精神障碍提供独特的指纹。

结论：我们的研究表明，dFIP提供了识别精神病状况中特定疾病功能连接模式的新颖、数据驱动的方法。这些独特的神经特征可能作为精神疾病的生物标志物，有助于更深入地理解这些障碍的神经生物学基础。