Sensory-Biased Working Memory & Attention Networks in the Human Brain

人类通过多种感官方式体验世界--我们看，我们听，我们摸。每一种感官形式在表征我们周围世界的能力上都有其独特的优势和劣势。我们的大脑能够灵活地招募视觉或听觉大脑结构，当它们的优势与手头的任务相对应时。此外，我们在任何时刻的体验都取决于感官输入和大脑中的工作记忆和注意力机制。这些机制的能力非常有限，这反过来又限制了认知。该项目的总体目标是研究支持视觉，听觉和触觉中的注意力和工作记忆的人脑机制。研究小组将进行功能性MRI实验，研究健康成年人在执行要求苛刻的感觉工作记忆任务时的大脑活动。该研究小组的初步研究表明，有广泛的大脑网络，延伸到大脑皮层的额叶，专门用于每种感觉方式，以及支持和统一这三种感觉的共享网络。目前的研究计划将研究工作记忆表现和大脑网络组织的个体差异。它还将开发先进的计算模型，可以从大脑连接的独特模式或“指纹”预测个体大脑的功能组织。该项目将通过传播新的模型和计算工具来促进其他研究工作，并将招募和培训年轻科学家，包括在STEM中代表性不足的群体的成员，进行认知神经科学研究。该提案有4个主要智力目标：（1）确定触觉，视觉，听觉和独立于模态的注意力&工作记忆（WM）的精细尺度组织（2）揭示了每种通道下大脑皮层区域的工作记忆信息编码的特异性;（3）详细描述了注意&工作记忆回路的网络组织;（4）检验了关于内容特异性工作记忆机制和跨通道工作记忆编码的假设。个体受试者的功能磁共振成像分析允许精细尺度观察不同的功能区域。利用特定于受试者的皮层组织图，研究小组将重新研究一个备受争议的问题，即哪些大脑结构支持每种模式的刺激特定工作记忆。它将通过检查个体参与者在静息状态下的功能网络来研究感觉通道偏向区域的特异性。该研究小组将利用其研究结果，在人类连接组项目数据集中的1200名受试者中探索这些网络。该研究小组还将测试和验证一种机器学习方法，即连接体指纹识别，用于从个体大脑独特的功能连接体中预测特定于模态的工作记忆区域的位置。视觉在编码空间信息方面表现出色，但编码时间不太可靠;相反，听觉系统执行高保真的时间编码，但空间编码粗糙。研究小组已经观察到WM信息到皮层结构的跨通道重新编码，这些结构更喜欢“适当的通道”-听觉空间WM招募视觉偏向区域，视觉时间WM招募视觉偏向区域。为了探索特定于内容的工作记忆机制，研究人员将研究感觉形式工作记忆和空间/时间工作记忆之间的相互作用。研究小组假设视觉和空间WM存储是不同的，但听觉和时间WM存储是共享的。总的来说，这些研究将阐明支持感觉工作记忆的人脑网络和机制。这个奖项反映了NSF的法定使命，并被认为是值得通过使用基金会的智力价值和更广泛的影响审查标准进行评估的支持。

AbstractHumans experience the world via multiple sensory modalities -- we see, we hear, and we touch. Each sensory modality has unique strengths and weaknesses in its ability to represent the world around us. Our minds are able to flexibly recruit visual or auditory brain structures when their strengths correspond to the task at hand. In addition, our experience at any moment depends both on sensory input and on working memory and attention mechanisms in the brain. These mechanisms have very limited capacities, which in turn limit cognition. The overarching goal of this project is to examine the human brain mechanisms that support attention and working memory in vision, hearing, and touch. The research team will perform functional MRI experiments to study the brain activity of healthy adults while they perform demanding sensory working memory tasks. Preliminary studies by the research group suggest that there are extensive brain networks, extending into the frontal lobes of the cerebral cortex, that are specialized for each sensory modality, as well as a shared network that supports and unifies these three senses. The current research program will examine individual differences in working memory performance and brain network organization. It will also develop advanced computational models that can predict the functional organization of an individual's brain from their unique pattern or 'fingerprint' of brain connectivity. The project will facilitate other research efforts through the dissemination of new models and computational tools, and will recruit and train young scientists, including members of groups that are under-represented in STEM, in cognitive neuroscience research.This proposal has 4 primary intellectual goals: (1) identify the fine-scale organization of tactile, visual, auditory, and modality-independent attention & working memory (WM) regions within human cerebral cortex; (2) reveal the specificity of coding of WM information across cortical regions for each modality; (3) detail the network organization of attention & WM circuits; and (4) test hypotheses about content-specific WM mechanisms and cross-modality WM coding. Individual subject fMRI analyses permit fine-scale observation of distinct functional regions. Drawing on subject-specific maps of cortical organization, the research group will re-examine the highly debated question of which brain structures support stimulus-specific working memory for each modality. It will investigate the specificity of sensory modality biased regions by examining functional networks in the resting-state within individual participants. The research team will leverage its findings to probe these networks in 1200 subjects from the Human Connectome Project dataset. The research group will also test and validate a machine-learning approach, Connectome Fingerprinting, for predicting the location of modality-specific working memory regions in individual brains from their unique functional connectome. Vision excels in coding spatial information, but codes timing less reliably; conversely, the auditory system performs high-fidelity temporal coding but coarse spatial coding. The research team has observed cross-modal recoding of WM information into cortical structures that prefer the 'appropriate modality' - auditory spatial WM recruits visual-biased regions and visual temporal WM recruits auditory-biased regions. To probe content-specific WM mechanisms, the researchers will examine interactions between sensory modality WM and space/time WM. The research group hypothesizes that visual and spatial WM stores are distinct, but that auditory and timing WM stores are shared. Collectively, these studies will elucidate the human brain networks and mechanisms that support sensory working memory.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.