While white matter myelin primarily functions to accelerate conduction velocity and has been extensively studied in schizophrenia-spectrum disorders (SSD), less is known about the role of gray matter myelin in SSD. Cortical myelination occurs mostly on the proximal axons of parvalbumin positive (PV+) interneurons, where it assists in trophic support and experience-dependent plasticity. Given the role of PV+ interneuron dysfunction in SSD, it is critical to advance our understanding of cortical myelin pathology in this context. Here, we quantified myelin maps using the T1w/T2w ratio in a large group of antipsychotic medication-naïve, first-episode psychosis patients. We compared myelin content between patients (N = 91) and controls (N = 107) using a MANCOVA and calculated zero-order correlations with the discriminant function for each region, then used a machine learning approach to identify the most parsimonious constellation of cortical regions driving group differences using a stepwise algorithm. Group membership was significantly associated with T1w/T2w ratio (Wilks Lambda = 0.09, p < 0.01), where patients had higher myelin values compared to healthy controls. We identified a subset of 16 regions, primarily located in association cortices, that were sufficient to explain group differences. Here, we report an increase in the cortical T1w/T2w ratio in association cortices in first-episode psychosis. We suggest that faulty myelin compaction during this critical developmental period could contribute to PV+ interneuron pathology and cortical microcircuit disruptions resulting in the clinical phenotype. With additional empirical support from future studies, novel treatment strategies targeting cortical myelin could have potential to mitigate circuit dysfunction in the illness.