K. Hwang: Context representations in the human cognitive thalamus

Abstract
In a series of studies, we investigated the connectivity and representational properties of the human mediodorsal thalamus (MD) and its contributions to cognitive control. Our findings revealed several key insights. First, using fMRI data from subjects performing a diverse set of tasks, we found that task-evoked thalamic responses converge onto a low-dimensional structure. This architecture reflects a compact set of basis activity patterns in the thalamus, with anterior, medial, and posterior-medial regions exhibiting broad, domain-general activity profiles. These findings align with evidence identifying these thalamic regions as multi-domain network hubs. Second, this low-dimensional representation preferentially encodes context in the MD during hierarchical cognitive control and influences cortical activity to select context-relevant representations. To further understand how task-evoked thalamic responses contribute to hierarchical cognitive control, we developed a thalamocortical network model. This model demonstrated that MD representations, through functional connectivity with cortical networks, transform task-specific activity into cognitive representations. It outperformed models based on other brain structures, underscoring the critical role of thalamocortical interactions in supporting cognition. We validated the model through simulations and empirical data from patients with focal thalamic lesions. These lesions selectively disrupted task-specific cortical activity and impaired executive functions, directly linking thalamic network properties to cognitive deficits. Finally, our recent work revealed that MD context representations integrate with error signals, promoting context-dependent flexibility. Together, these studies demonstrate how the MD serves an active, representational role in cognition, facilitating flexible, goal-directed behavior.

Bio
Dr. Kai Hwang received his PhD from the University of Pittsburgh and completed his postdoctoral training at UC Berkeley. He is currently an Associate Professor in the Department of Psychological and Brain Sciences at the University of Iowa. His research focuses on the neural and cognitive dynamics of cognitive control, utilizing methodologies such as fMRI, EEG, TMS, lesion studies, eye tracking, and computational modeling. Dr. Hwang's work aims to understand the neural architecture and dynamic processes that enable brain networks to select, inhibit, transfer, and integrate information for goal-directed behaviors. His lab's recent studies particularly emphasize the thalamocortical system and its contribution to cognitive control.

Speaker's website

Lab's website

Zoom Link
Meeting-ID: 646 4455 4045
Passcode: f&%%A9wa