SSRI treatment of obsessive-compulsive disorder as motion across a bistable fold: a calibrated circuit plasticity model of response, remission, and augmentation

SSRIs are first-line for obsessive-compulsive disorder (OCD), yet benefit is partial, weeks-delayed, and severity-gated, and augmentation is empirical. The stable-states view is established but not drug-specific. We supply the missing coupling: a calibrated object joining serotonin pharmacokinetics, a basal-ganglia-thalamocortical circuit, and corticostriatal plasticity, in which OCD is an elevated fixed point and SSRI treatment moves this bistable state across a saddle-node fold (a tipping point). With ~35 constants fixed, two shared parameters reproduce six SSRIs (RMS 0.83 Y-BOCS points, 84 arm-timepoints) and predict a held-out trial (0.58). A fold-free model fits the means equally well, so the mean cannot separate them; they diverge only on individual-level signatures it cannot produce - a discrete, durable off-drug remission subgroup and critical slowing - with qualitative support in OCD and adjacent disorders, not yet tested as bistability signatures. These, not the mean, decide who remits and relapses; the same geometry reframes memantine and antipsychotic augmentation.