The geometry of domain-general performance monitoring in the human medial frontal cortex
Monitoring our own behavior without explicit feedback is a prominent human ability, enabling us to evaluate whether we made an error, are experiencing conflict, or are in an easy or difficult task environment. In real-life situations, we often need to rapidly learn to perform novel tasks with minimal instruction and time for practice, which requires domain-general performance monitoring that functions “out of the box.” At the same time, we need to resolve unforeseen errors, difficulty, and other performance disturbances with task-specific measures ad hoc. Doing so requires domain-specific performance monitoring processes so as to improve upon and specialize task performance. Are there neural representations that support both domain-specific and domain-general performance monitoring, and if so, how are these requirements satisfied simultaneously?
Flexibly adapting behavior to achieve a desired goal depends on the ability to monitor one’s own performance. A key open question is how performance monitoring can be both highly flexible to support multiple tasks and specialized to support specific tasks. We characterized performance monitoring representations by recording single neurons in the human medial frontal cortex (MFC). Subjects performed two tasks that involve three types of cognitive conflict. Neural population representations of conflict, error and control demand generalized across tasks and time while at the same time also encoding task specialization. This arose from a combination of single neurons whose responses were task-invariant and non-linearly mixed. Neurons encoding conflict ex-post served to iteratively update internal estimates of control demand as predicted by a Bayesian model. These findings reveal how the MFC representation of evaluative signals are both abstract and specific, suggesting a mechanism for computing and maintaining control demand estimates across trials and tasks.