In our article,1 we argued that changes in measurements on MRI scans over time in patients with schizophrenia should not be interpreted as evidence that neurodegeneration or neurotoxicity is a primary feature of the disease. The points that we raised to support this argument included the paucity of biologically plausible clinical-neuropathologic correlations, the remarkable inconsistency of the data, and the virtual absence of evidence of neurodegeneration in postmortem brain tissue, despite numerous studies. That Mathalon et al now state that concluding otherwise would be "misguided" suggests that our argument was convincing, as these authors in their earlier writings had advocated for a neurodegeneration interpretation (although their current use of the phrase "parenchyma lost" raises concerns that there is still some reluctance). In defending the MRI morphometry literature, they raise an important question: do longitudinal MRI morphometry studies contribute to our understanding of the brain disorder that accounts for the clinical condition? While our article addressed the problems with this literature in terms of a specific interpretation of the findings, we think that the question they now raise invites some additional comment.
Weinberger DR, McClure RK. Neurotoxicity, Neuroplasticity, and Magnetic Resonance Imaging Morphometry—Reply. Arch Gen Psychiatry. 2003;60(8):848-849. doi:10.1001/archpsyc.60.8.846