Departments of Cognitive Sciences and Neurology, University of California, Irvine, Irvine, CA, 92697, USA.
Department of Communication Sciences & Disorders, Louisiana State University, 74 Hatcher Hall, Baton Rouge, LA, 70803, USA.
Brain Imaging Behav. 2018 Aug;12(4):1141-1149. doi: 10.1007/s11682-017-9771-2.
Resting state magnetic resonance imaging (rsfMRI) permits observation of intrinsic neural networks produced by task-independent correlations in low frequency brain activity. Various resting state networks have been described, with each thought to reflect common engagement in some shared function. There has been limited investigation of the plasticity in these network relationships after stroke or induced by therapy. Twelve individuals with language disorders after stroke (aphasia) were imaged at multiple time points before (baseline) and after an imitation-based aphasia therapy. Language assessment using a narrative production task was performed at the same time points. Group independent component analysis (ICA) was performed on the rsfMRI data to identify resting state networks. A sliding window approach was then applied to assess the dynamic nature of the correlations among these networks. Network correlations during each 30-second window were used to cluster the data into ten states for each window at each time point for each subject. Correlation was performed between changes in time spent in each state and therapeutic gains on the narrative task. The amount of time spent in a single one of the (ten overall) dynamic states was positively associated with behavioral improvement on the narrative task at the 6-week post-therapy maintenance interval, when compared with either baseline or assessment immediately following therapy. This particular state was characterized by minimal correlation among the task-independent resting state networks. Increased functional independence and segregation of resting state networks underlies improvement on a narrative production task following imitation-based aphasia treatment. This has important clinical implications for the targeting of noninvasive brain stimulation in post-stroke remediation.
静息态磁共振成像（rsfMRI）可以观察低频脑活动中任务独立相关性产生的内在神经网络。已经描述了各种静息态网络，每个网络都被认为反映了某种共同功能的共同参与。在中风后或通过治疗诱导后，这些网络关系的可塑性的研究还很有限。对 12 名中风后语言障碍（失语症）患者在治疗前（基线）和基于模仿的失语症治疗后进行多次成像。同时在相同时间点进行了使用叙事生成任务的语言评估。对 rsfMRI 数据进行了组独立成分分析（ICA），以识别静息态网络。然后应用滑动窗口方法来评估这些网络之间相关性的动态性质。在每个 30 秒窗口期间，使用网络相关性将数据聚类为每个窗口的每个时间点的十个状态。将每个状态中花费的时间变化与叙事任务的治疗收益进行相关性分析。与基线或治疗后立即评估相比，与叙事任务的治疗后 6 周维持间隔的行为改善相比，花费在（十个总体状态中的）单个动态状态中的时间量与叙事任务的行为改善呈正相关。该特定状态的特点是任务独立的静息态网络之间的相关性最小。基于模仿的失语症治疗后，叙事产生任务的改善与静息态网络的功能独立性和分离性增加有关。这对中风后康复中靶向非侵入性脑刺激具有重要的临床意义。
