Plasticity of large-scale neural connectivity following working memory training

在老龄化社会中，保持和改善认知功能的技术变得越来越重要。工作记忆（WM）训练等方法被发现可以提高处理能力。在神经影像学研究中，这种改善与区域大脑信号振幅的改变有关，而大脑区域之间相互作用或连通性的变化尚不清楚。考虑到越来越多的证据表明人类大脑是一个动态的神经网络系统，有必要开发基于连接的神经可塑性标记。在本研究中，我们将在大尺度脑连接水平上研究训练诱导认知改善的神经元基质。鉴于神经可塑性的多种机制，我们将在一组中年健康参与者中应用神经成像技术，如功能磁共振成像、PET和DTI。使用这些模式，功能、代谢和结构连通性将被测量。实验对象将接受系统的WM训练，然后与积极的对照组进行比较。在训练前后，将在混合PET/MR系统上同时获取MRI和氟脱氧葡萄糖PET数据。在相同的两个时间点，以及3个月和12个月的随访期间，将进行广泛的认知测试。我们假设训练后默认模式和中央执行网络的连通性指数增加。然而，考虑到不同的时间动态和能量消耗，不同形式的训练诱导的神经可塑性预计与时间上的认知收益有不同的关系。具体来说，我们假设网络结构和代谢连通性的增加预测长期（训练后12个月）接近转移效应的准确性高于功能连通性。除了神经生物学的相关性外，该研究的结果将有助于开发针对年龄和疾病相关认知缺陷的认知干预措施。

In an ageing society techniques to preserve and improve cognitive function are gaining increasing importance. Approaches such as working memory (WM) training were found to improve processing capacities. In neuroimaging studies, this improvement was associated with alterations in amplitude of regional brain signal, while changes in interaction, or connectivity, between brain regions remain unclear. Considering increasing evidence that the human brain represents a dynamic system of neural networks, there is a need for development of connectivity-based markers of neural plasticity. In the proposed study we will investigate neuronal substrates of training-induced cognitive improvement at the level of large-scale brain connectivity. Given a variety of mechanisms underlying neural plasticity, we will apply a combination of neuroimaging techniques such as fMRI, PET, and DTI in a group of middle-aged healthy participants. Using these modalities, functional, metabolic, and structural connectivity will be measured. Experimental subjects will undergo a systematic WM training, followed by a comparison with an active control group. Simultaneous acquisition of MRI and fluorodeoxyglucose PET data on a hybrid PET/MR system will take place before and after training. An extensive cognitive test battery will be administered at the same two time points, as well as at 3 and 12 months follow-up. We hypothesize an increase in connectivity indices of the default mode and central executive networks following training. However, given the distinct temporal dynamics and energy costs, different forms of training-induced neuroplasticity are expected to be differently related to cognitive gains in time. Specifically, we hypothesize that increase in network structural and metabolic connectivity predicts long-term (12 months after training) near-transfer effects with higher accuracy than does functional connectivity. Beside neurobiological relevance, results of the proposed study will contribute to the development of cognitive interventions against age- and disease-related cognitive deficits.