Explicit knowledge gates expectation suppression in the motor system: Evidence from a TMS oddball paradigm

Abstract

Models of predictive processing propose that the brain continuously generates predictions about incoming sensory input, updating an internal model of the environment through prediction errors when those predictions are violated. A foundational assumption of these models is that prediction errors are generated automatically, even in the absence of explicit knowledge of the predictions. Evidence from auditory oddball studies in unconscious patients appears to support this view, though findings are complicated by stimulus-specific adaptation confounds that make it difficult to isolate genuine predictive effects. To investigate whether expectation suppression or prediction-based attenuation extends to the motor system and whether it operates automatically, we developed a novel motor oddball paradigm using brain stimulation. Transcranial magnetic stimulation (TMS) delivered over the primary motor cortex elicits motor-evoked potentials (MEPs) in peripheral muscles, providing an index of corticospinal excitability. By varying stimulation intensity in an oddball-like manner using repeating and deviating sequences, we manipulated the predictability of TMS pulses and compared MEP amplitudes for expected versus unexpected intensity-matched stimulation. Incorporating experimental designs to control for adaptation and an instruction manipulation to test the role of explicit knowledge, expected TMS produced smaller MEPs than unexpected TMS. Critically, this attenuation was observed only in participants explicitly informed about the sequence structure. These findings extend expectation suppression effects to the motor system and support the domain-generality of prediction-based neural attenuation while challenging the assumption that predictive processing operates entirely automatically.