Theta-Band Temporal Interference Stimulation Targeting the Dorsal Anterior Cingulate Cortex Modulates Prediction Error Encoding in Food Addiction

BACKGROUNDFood addiction is characterized by persistently elevated expectations of food reward, suggesting impairments in prediction error processing. The dorsal anterior cingulate cortex (dACC) encodes prediction errors via theta-band activity, but causal evidence in humans remains limited. Transcranial temporal interference stimulation (tTIS) offers a non-invasive approach to modulate deep cortical regions including the dACC. METHODSThirty-four individuals with food addiction (15 males) were randomly assigned to active theta-band tTIS or sham stimulation. Participants completed a food incentive delay task in a satiated state while EEG was recorded pre- and post-stimulation. Feedback-related negativity (FRN) was analyzed at the mean ERP and single-trial levels using mixed-effects models linking FRN amplitudes to behavioral and model-based prediction errors; stimulus preceding negativity (SPN) was analyzed at the mean ERP level. RESULTSTrial-level mixed-effects models revealed robust prediction error encoding in FRN amplitudes during food reward feedback. Critically, tTIS altered FRN sensitivity to prediction errors, such that pre-to-post changes in FRN-prediction error coupling differed between the active and sham groups for both behavioral prediction errors ({beta} = 1.74, p = .049) and model-based prediction errors ({beta} = 1.77, p = .048). At the mean ERP level, FRN during food feedback became more negative after stimulation in the active group (t16 = 2.75, p = .014), whereas SPN was unchanged. CONCLUSIONSTheta-band tTIS targeting the dACC selectively enhances sensitivity to prediction errors during food reward processing, highlighting temporal interference stimulation as a tool to causally modulate feedback-evaluation mechanisms in humans.