Spatiotemporal organization of brain areas’ spectral fingerprints informing a new generation of computational whole-brain models

认知需要结构和功能不同的大脑区域的协调互动。神经元振荡与许多认知过程有关，在神经精神疾病中受到干扰，最近通过非侵入性脑磁图（MEG）记录发现，每个区域都是独特的。然而，这些“光谱指纹”的详细时空组织，它们与大脑结构特性的潜在共变，以及它们的受试者特异性和个体差异尚未被表征。出于这个原因，我建议，首先，使用人类连接组计划提供的健康人群的大型现有MEG数据集来表征这些属性。总之，光谱，时间，空间和结构信息，以及它们在健康个体中的变化，将提供区域特异性神经处理的广泛表征，是建立光谱指纹作为生物标志物进行比较的关键步骤，在未来的工作中，与患病人群。此外，光谱指纹将为计算全脑模型提供信息和改进。具体而言，建立模型，在皮层区域的水平上进行空间分辨，具有不切实际的功率谱。因此，我建议，第二，开发新一代的全脑模型，更现实的功率谱，并推断出潜在的神经电路参数，通过约束这些模型与每个区域的频谱指纹。计算建模结合非侵入性可获得的生物标志物可以揭示患病个体中神经回路参数的干扰，从而可以暗示仍然知之甚少的神经精神疾病的神经回路机制。

Cognition requires the coordinated interaction of structurally and functionally diverse brain areas. Neuronal oscillations have been linked to a multitude of cognitive processes, are disturbed in neuropsychiatric diseases and have been recently revealed, through non-invasive magnetoencephalography (MEG) recordings, to be unique to each area. However, the detailed spatio-temporal organization of these “spectral fingerprints”, their potential co-variation with structural properties of the brain, as well as their subject-specificity and individual differences have not been characterized. For that reason, I propose, first, to characterize these properties using a large existing MEG data set of a healthy population provided by the Human Connectome Project. Together, spectral, temporal, spatial and structural information, as well as their variation in healthy individuals, will provide an extensive characterization of area-specific neural processing and is a crucial step for establishing spectral fingerprints as biomarkers to be compared, in future work, with a diseased population. Moreover, spectral fingerprints will inform and improve computational whole-brain models. Specifically, established models, spatially resolved at the level of cortical areas, possess unrealistic power spectra. I therefore propose, second, to develop a new generation of whole-brain models with more realistic power spectra and to infer the underlying neural circuitry parameters by constraining these models with each area’s spectral fingerprint. Computational modeling in combination with non-invasively obtainable biomarkers can reveal disturbances in neural circuit parameters in diseased individuals and can thereby hint at neural circuit mechanisms of neuropsychiatric diseases which remain poorly understood.