BACKGROUND: Computed tomography (CT) is the primary imaging investigation for many neurologic conditions with a proportion of patients incurring cumulative doses. Iterative reconstruction (IR) allows dose optimization, but head CT presents unique image quality complexities and may lead to strong reader preferences. OBJECTIVES: This study evaluates the relationships between image quality metrics, image texture, and applied radiation dose within the context of IR head CT protocol optimization in the simulated patient setting. A secondary objective was to determine the influence of optimized protocols on diagnostic confidence using a custom phantom. METHODS AND SETTING: A three-phase phantom study was performed to characterize reconstruction methods at the local reference standard and a range of exposures. CT numbers and pixel noise were quantified supplemented by noise uniformity, noise power spectrum, contrast-to-noise ratio (CNR), high- and low-contrast resolution. Reviewers scored optimized protocol images based on established reporting criteria. RESULTS: Increasing strengths of IR resulted in lower pixel noise, lower noise variance, and increased CNR. At the reference standard, the image noise was reduced by 1.5 standard deviation and CNR increased by 2.0. Image quality was maintained at ≤24% relative dose reduction. With the exception of image sharpness, there were no significant differences between grading for IR and filtered back projection reconstructions. CONCLUSIONS: IR has the potential to influence pixel noise, CNR, and noise variance (image texture); however, systematically optimized IR protocols can maintain the image quality of filtered back projection. This work has guided local application and acceptance of lower dose head CT protocols.