Anterior cingulate cortex mixes retrospective cognitive signals and ongoing movement signatures during decision-making

The anterior cingulate cortex (ACC) is believed to play a key role in reward-based learning in uncertain and dynamic environments. Recent work suggests that in these environments, ACC neurons monitor decisions and their outcomes to inform behavioral strategy. Yet, it remains unclear whether ACC neurons similarly encode behavioral history in fully deterministic perceptual decision-making tasks and, if so, how behavioral history signals interact with other signals that are represented by ACC neurons. Here, we recorded the activity of ACC neurons in freely moving mice performing a visual evidence accumulation task. Many ACC neurons had mixed selectivity: they were strongly driven by non-linear combinations of previous choices and outcomes (trial history). We observed that trial history could be decoded well from the ACC population and that the neural representations of trial history remained stable over multiple seconds. Using linear encoding models, we demonstrate that both trial history and movements strongly drive neural activity in the ACC. We found no relationship between trial history encoding and behavioral biases and instead observed that the neural dynamics encoding trial history were low-dimensional, similar between sessions from the same subjects and conserved across different subjects. These findings suggest the ACC might implement a trial history monitoring process that is independent of behavioral biases and common to different subjects.