We report the effects of implementing the list mode image reconstruction method (LMR) in a new brain scanner based on positron emission tomography. The benefits of LMR include reduced load in imaging processes and conservation of fine spatial data-sampling, the latter of which may be lost by data binning in histogram mode image reconstruction (HMR). In our project, we aim to build a high-spatialresolution scanner that employs three-dimensional positionsensitive CdTe semiconductor detector units. We experimentally confirmed that the unit could measure the detected positions of annihilation photons with a position resolution of 1 mm3 full width at half maximum (FWHM). Therefore, the scanner can potentially minimize the parallax-induced decrease in the spatial resolution of off-center positions in the field of view (FOV), thereby providing high resolution throughout the FOV because it obtains accurate depth of interaction (DOI) information at a sampling pitch ~1 mm finer than that of conventional DOI technologies (~4 mm). We simulated the proposed scanner. using the GEANT4 Application for Tomographic Emission. Reconstructed images were obtained by LMR or HMR based on the maximum likelihoodexpectation maximization (ML-EM) algorithm. The system matrix for ML-EM reconstruction was calculated by Siddon’s ray-driven method to fully exploit accurate detection positions. The spatial resolution at the center of the FOV acquired by LMR is 0.6 mm FWHM, which is superior to that acquired by HMR (1 mm FWHM). We confirmed that the resolutions of both methods are maintained at off-center positions, and the resolutions acquired by LMR at these positions are superior to those acquired by HMR. Furthermore, we simulated Hoffmanbrain- phantom imaging to evaluate the image quality of the scanner and LMR in human brain studies.