Comments of Bhambhwani and associates regarding our novel description of extraocular muscle (EOM) shape differences in superior oblique (SO) palsy highlight the clinical importance of compartmentalization. Our anatomic findings in human, monkey, cow, and rabbit of segregation of the SO into distinct neuromuscular compartments extend recent recognition of a fundamental anatomic feature of EOMs that has also been confirmed in the lateral (LR), medial (MR), and inferior rectus (IR) and inferior oblique muscles (Le A., Demer J. L., Poukens V. Compartmental innervation scheme for the mammalian superior oblique (SO) and inferior oblique (IO) muscles. Program No. 62.22. Neuroscience Meeting Planner. Washington, DC: Society for Neuroscience, 2014. Online.). Magnetic resonance imaging (MRI) demonstrates differential function in normal humans: for MR during convergence, for LR during ocular counter-rolling, and for LR, IR, and SO during vertical fusional vergence, where the paradoxical compartmental role of the SO is particularly remarkable. These findings challenge many longstanding assumptions about how EOMs function in health and disease.

Our study included only subjects with highly significant, unilateral reduction in SO size relative to a large group of normal control subjects. No bilateral or masked cases were included. Our paper noted that overelevation in adduction did not differ between SO atrophy groups. Duration of SO palsy had no significant effect on SO size, consistent with our finding that SO atrophy following experimental trochlear neurectomy reaches completion in no more than 5 weeks. We acknowledge that current MRI cannot distinguish medial from lateral compartment SO palsy in humans; an animal model is needed to clarify this issue.

We have accumulated MRI evidence that much of current teaching about SO palsy is incorrect. For example, the time honored “3-step” test establishing ipsilateral hypertropia that increases with contralateral gaze, and with ipsilateral head tilt, is only 70% sensitive and 50% specific for actual SO dysfunction. Indeed, the differences in torsion and vertical deviations that Bhambhwani and associates attribute to rectus and contralateral SO overaction occur even in the absence of SO palsy! We did not perform vertical forced duction tests in our study, for reasons including our concern that these tests, highly subjective for the examiner yet unpleasant for the alert patient, would be uninterpretable. We are unaware of evidence for the existence of “spread of comitance,” and it is evident neither in the monkey model of SO palsy nor in the few patients we have followed long-term without treatment.

We concur with Bhambhwani and associates that more must be learned about the nature of lesions producing selective compartmental SO palsy. However, this does not diminish the importance of objective studies of SO structure and function in the diagnosis of SO palsy itself. As we have demonstrated, there are no clinical motility findings, either singly or in “amalgamation,” that reliably correlate with SO size and function as demonstrated by MRI. Longstanding concepts of SO palsy, mainly based upon clinical anecdote and intuition with liberal application of inductive reasoning, require significant updating in light of modern, objective methods such as MRI.