Orbitofrontal cortex drives predictive filtering of sensory responses in auditory cortex

• Top-down projections from the orbitofrontal cortex carry predictive signals that grow with sound experience and suppress the auditory cortex via inhibitory circuits (Nature Neuroscience).¹
• Reveals a predictive mechanism for sensory habituation; informs predictive coding and closed-loop stimulation design.¹
• Study was conducted entirely in awake, head-fixed mice using chronic two-photon calcium imaging; 5,483 L2/3 neurons were recorded across 20 mice over 5 days of repeated tone exposure (70 dB SPL, 200 trials/day).¹
• Two timescales of habituation were identified—within-day and across-day—whose high-dimensional population trajectories were nearly orthogonal (78.2°), indicating distinct mechanisms; pupillometry confirmed across-day habituation is independent of arousal (pupil diameter).¹
• Pharmacological inactivation (muscimol) of the ventrolateral OFC (OFCvl) after 5 days of habituation reversed habituation in A1 pyramidal cells (change index 0.31 ± 0.04, P = 2.0 × 10⁻¹³), whereas inactivation in naïve animals had no effect (change index −0.002, P = 0.82), supporting a predictive rather than novelty-driven mechanism.¹
• OFCvl axon boutons in A1 increased their sound-evoked activity across days; by day 5, 17.7% showed significant potentiation (vs. 8.2% reduced, P = 0.0034); critically, bouton responses were sound-specific—they increased for the repeatedly exposed tone A but not for a novel tone B presented only on days 0 and 6 (tone A change index 0.55 ± 0.14, P = 0.0042; tone B 0.06 ± 0.17, P = 0.71).¹
• The identified circuit mechanism: OFCvl predictive signals selectively recruit SST (somatostatin) interneurons in A1, which suppress pyramidal, VIP, and PV cells; OFCvl inactivation strongly reduced SST responses (change index −0.50 ± 0.06, P = 1.1 × 10⁻¹¹) while increasing VIP and PV responses.¹
• Optogenetic activation of the OFCvl→A1 pathway was sufficient to suppress A1 activity (modulation index −0.23 ± 0.06, P = 0.012), establishing pathway sufficiency; silencing this pathway on day 6 selectively affected responses to habituated tone A but not novel tone B.¹
• NMDA receptor knockout experiments (GluN1 deletion) in A1 neurons and in SST cells both significantly reduced sensory habituation (Ctrl vs. KO-A1: P = 3.6 × 10⁻⁵; Ctrl vs. KO-SST: P = 1.8 × 10⁻⁵), while VIP-cell GluN1 deletion had no effect, implicating local NMDA-dependent synaptic plasticity in SST cells as essential for OFC-driven habituation.¹
• Predictive filtering was stronger during the sustained phase of the tone than at onset, consistent with a slow top-down feedback loop predicting stimulus continuation rather than merely its onset.¹
• This study was conducted at the University of North Carolina at Chapel Hill and provides the first causal identification of OFC as the frontal source that builds an internal predictive model of repeated sounds—placing OFC at the top of a sensory-prediction hierarchy and complementing intracranial EEG work in humans showing OFC leads predictive encoding in auditory deviance detection tasks.¹

Footnotes

1. https://www.nature.com/articles/s41593-026-02217-z ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹