Changping Laboratory, Beijing, China.
Department of Medical Imaging, Zhengzhou University People's Hospital & Henan Provincial People's Hospital, Zhengzhou, China.
Neuropsychopharmacology. 2023 Mar;48(4):633-641. doi: 10.1038/s41386-022-01500-4. Epub 2022 Nov 19.
Substantial clinical heterogeneity and comorbidity inherent amongst mental disorders limit the identification of neuroimaging biomarkers that can reliably track clinical symptoms. Strategies that enable generation of meaningful and replicable neurobiological markers at the individual level will push the field of neuropsychiatry forward in developing efficacious personalized treatment. The current study included 142 adult patients with a primary diagnosis of schizophrenia (SCZ), bipolar (BP), or attention deficit/hyperactivity disorder (ADHD), and 67 patient ratings across four behavioral measures. Using functional connectivity derived from a personalized fMRI approach, we identified several candidate imaging markers related to dimensional phenotypes across disorders, assessed the internal and external generalizability of these markers, and compared the probability of replicating findings across datasets using individual and group-averaged defined functional regions. We identified subject-specific connections related to three different clinical domains (attention deficit, appetite-energy, psychosis-positive) in a discovery dataset. Importantly, these connectivity biomarkers were robust and were reproduced in an independent validation dataset. For markers related to neurovegetative symptoms (attention deficit, appetite-energy symptoms), the brain connections involved showed similar connectivity patterns across the different diagnoses. However, psychosis-positive symptoms were associated with connections of varying strength across disorders. Finally, we found that markers for symptom domains were replicable for individually-specified connections, but not for group template-derived connections. Our personalized strategies allowed us to identify meaningful and generalizable imaging markers for symptom domains in patients who exhibit high levels of heterogeneity. These biomarkers may shed new light on the connectivity underpinnings of psychiatric symptoms and lead to personalized interventions.
大量的临床异质性和精神障碍固有的合并症限制了能够可靠地跟踪临床症状的神经影像学生物标志物的识别。能够在个体水平上生成有意义和可重复的神经生物学标志物的策略将推动神经精神病学领域朝着开发有效的个性化治疗方法前进。本研究纳入了 142 名患有首发精神分裂症 (SCZ)、双相情感障碍 (BP) 或注意力缺陷/多动障碍 (ADHD) 的成年患者，以及 4 项行为测量中的 67 名患者评分。使用源自个性化 fMRI 方法的功能连接，我们确定了与跨障碍维度表型相关的几个候选影像学标志物，评估了这些标志物的内部和外部泛化能力，并使用个体和群体平均定义的功能区域比较了跨数据集复制发现的可能性。我们在发现数据集确定了与三个不同临床领域（注意力缺陷、食欲-能量、精神病阳性）相关的特定于个体的连接。重要的是，这些连接生物标志物是稳健的，并在独立验证数据集中得到了重现。对于与神经植物性症状（注意力缺陷、食欲-能量症状）相关的标志物，所涉及的大脑连接在不同诊断中显示出相似的连接模式。然而，精神病阳性症状与跨疾病的连接强度不同有关。最后，我们发现症状域的标志物对于个体指定的连接是可复制的，但对于基于组模板的连接则不可复制。我们的个性化策略使我们能够识别出具有高度异质性的患者中症状域的有意义且可推广的影像学标志物。这些生物标志物可能为精神症状的连接基础提供新的线索，并导致个性化干预。
