Intraoperative motor evoked potentials include the D-wave as a surrogate for long-term motor outcome and muscle motor evoked potentials as a surrogate for early outcome. Their efficacy depends on excluding confounding factors and on warning criteria; insufficiently sensitive criteria could result in unpredicted deficits, whereas excessively sensitive ones could cause false alarms deterring surgical treatment and jading surgeons to alerts, eventually leading to deficits through failure to intervene. Although D-waves have few indications, they are nonsynaptic, linear, and stable-properties that support amplitude reduction criteria: >50% for intramedullary spinal cord tumor surgery and >30% to 40% for peri-Rolandic brain surgery. Muscle motor evoked potentials have many indications but are polysynaptic, nonlinear, and unstable-properties that challenge warning criteria and make them unusually capricious and sensitive. Disappearance is a remarkably frequent pathologic sign compared with other evoked potentials and is always a major criterion. Marked (>80%) amplitude reduction may be a minor or moderate spinal cord criterion, depending on the surgical circumstance. Modest (>50%) reduction may be a major criterion for brain, brainstem, and facial nerve monitoring, if justified by sufficient preceding stability. Acute ≥100-V threshold elevation may be a minor or moderate spinal cord criterion, depending on the surgical circumstance and on adherence to reported methodology. Morphology criteria lack support. Tailoring warning criteria to different monitoring situations based on anatomy, surgical goals, and published evidence seems advisable.