Neural population dynamics of human working memory

The activity of neurons in macaque prefrontal cortex (PFC) persists during working memory (WM) delays, providing a mechanism for memory.1-11 Although theory,11,12 including formal network models,13,14 assumes that WM codes are stable overtime, PFC neurons exhibit dynamics inconsistent with these assumptions.15-19 Recently, multivariate reanalyses revealed the coexistence of both stable and dynamic WM codes in macaque PFC.20-23 Human EEG studies also suggest that WM might contain dynamics.24,25 Nonetheless, how WM dynamics vary across the cortical hierarchy and which factors drive dynamics remain unknown. To elucidate WM dynamics in humans, we decoded WM content from fMRI responses across multiple cortical visual field maps.26-48We found coexisting stable and dynamic neural representations of WM during a memory-guided saccade task. Geometric analyses of neural subspaces revealed that early visual cortex exhibited stronger dynamics than high-level visual and frontoparietal cortex. Leveraging models of population receptive fields, we visualized and made the neural dynamics interpretable. We found that during WM delays, V1 population initially encoded a narrowly tuned bump of activation centered on the peripheral memory target. Remarkably, this bump then spread inward toward foveal locations, forming a vector along the trajectory of the forthcoming memory-guided saccade. In other words, the neural code transformed into an abstraction of the stimulus more proximal to memory-guided behavior. Therefore, theories of WM must consider both sensory features and their task-relevant abstractions because changes in the format of memoranda naturally drive neural dynamics.

Automated summary

The activity of neurons in the prefrontal cortex during working memory delays is dynamic, and the dynamics vary across different cortical visual field maps. Using fMRI responses, researchers found that the early visual cortex exhibited stronger dynamics compared to the high-level visual and frontoparietal cortex. By leveraging population receptive field models, they visualized and interpreted the neural dynamics. Specifically, during working memory delays, the V1 population encoded a narrow bump of activation centered on the peripheral memory target, which then spread inward along the trajectory of the forthcoming memory-guided saccade.