We seek to understand how we learn from our mistakes. Recognizing what we did wrong is an essential part of learning—imagine trying to learn a new song without realizing you played a wrong note. This principle extends far beyond music, shaping much of what we learn throughout life, from walking and talking to recognizing objects and faces. Our goal is to uncover how the brain detects errors and uses this information to guide learning.

We have recently made progress toward this goal by identifying a specific cell type in the cerebral cortex that signals errors. We initially discovered this error signal in the context of navigation within the posterior parietal cortex, a region implicated in decision-making. We subsequently found that the same cell type signals other forms of errors in additional cortical regions, including sensory and motor areas. Together, these findings suggest that common cellular mechanisms may underlie error signaling and learning across a wide range of cortical modalities. Moreover, the identification of a genetically targetable error-signaling cell type opens new avenues for uncovering these mechanisms.

Our goal moving forward is to expand beyond this single cell type to develop a more comprehensive understanding of how the cortex learns from errors. We systematically investigate the contributions of cortical cell types, the neuromodulatory signals that influence their activity, and the subcellular electrical and molecular processes that shape their function. To do so, we combine mouse genetics, cell-type-specific viral tools, spatial transcriptomics, two-photon imaging, and two-photon optogenetics during behavior. Finally, we have developed experimental paradigms that elicit errors across sensory, cognitive, and motor domains, allowing us to test whether error signaling and learning mechanisms generalize across the cortex. Through this comparative approach, we aim to uncover general principles of error-based learning and advance a unified framework for cortex-based intelligence.