Distinct Roles of Deep and Superficial Cortical Layers in Tone Prediction, Comparison, and Adaptation in Human Auditory Cortex
Auditory perception relies on prior sensations both to form expectations about upcoming sounds and to adapt to repeated inputs. Yet, how these processes interact with bottom-up frequency-specific stimulus drive, and how they are organised across cortical depth, remains unclear. Using ultra-high-field functional magnetic resonance imaging (7T fMRI), participants listened to stochastic tone sequences whose repetition structure and predictability varied over time. For each voxel independently, we used computational modelling to estimate responses for stimulus drive, repetition suppression, and expectations. We then partitioned the unique and shared variance of these components to estimate their relative importance in modulating BOLD activity across cortical layers in auditory cortices. Expectation-related signals explained most unique variance in deep (infagranular) layers, in line with accounts that place internal model representations and belief updates in deeper cortical populations. Variance jointly explained by stimulus drive and expectation was instead strongest in superficial layers, consistent with these compartments expressing computations that combine bottom-up input with top-down predictions, such as prediction error signalling. Finally, repetition suppression accounted for variance uniformly across depth, suggesting an adaptation-like mechanism that modulates responsiveness without a clear laminar bias. Together, the data suggest that deep layers contain content-specific predictive models, superficial layers register prediction-input alignment, and repetition suppression provides a depth-invariant, local gain control that complements predictive processing.