Filtering the noise: a cerebellar-centered framework for understanding and treating mental illness -a paradigm shift in psychiatry

For the past half-century, psychiatric drug development has largely focused on tweaking neurotransmitter receptors and chemical pathways. Yet despite billions of dollars invested and major advances in neuroscience, truly innovative treatments for mental illness remain scarce. Disorders like depression, schizophrenia, and post-traumatic stress disorder (PTSD) continue to be managed with drugs discovered decades ago that often provide only partial relief, with remission rates of approximately 30-40% for treatment-resistant depression and 60-70% of schizophrenia patients experiencing persistent symptoms despite medication. This stagnation has prompted a paradigm shift - what if the key to treating mental illness is not just which receptor a drug targets, but how it changes the brain's processing of sensory information? In this treatise, I propose that many psychiatric conditions stem from breakdowns in the brain's sensory filtering mechanisms, the neural circuits that gate irrelevant stimuli before they consume valuable processing resources, and that effective therapies must restore these filtering functions. While computational psychiatry has long recognized that mental illness may reflect failures in predictive filtering, the specific neural substrate implementing this gating remains underspecified. Here the cerebellum emerges as a critical hub: neuroanatomically positioned to perform bottom-up sensory gating before cortical processing, housing more than half the brain's neurons in an architecture ideally suited for distilling signal from noise and showing state-dependent disruption of cerebellar-cortical connectivity during symptom provocation in PTSD. Intriguingly, psychedelic drugs may act as recalibration triggers for these neural filters, acutely disrupting entrenched filtering architectures and reopening windows of plasticity through which maladaptive sensory weightings can be reset. This cerebellar filtering framework offers a neuroanatomically specified extension of predictive processing theory, generates falsifiable predictions, and suggests novel therapeutic targets for conditions that have resisted a half-century of receptor-focused drug development.