Positron Emission Tomography (PET) has gained a tremendous momentum recently for clinical applications notably with the availability of (18)F-fluorodeoxyglucose for staging and evaluation of therapy efficacy in various types of cancers. Nonconventional positron emitting nuclides are now being investigated for the development of novel imaging and therapeutic strategies. However, these nuclides have less than ideal imaging properties. This article compares the performance for imaging of nonconventional nuclides such as (61)Cu, (68)Ga, (86)Y and (94m)Tc with the standard imaging nuclide (18)F for high-resolution small animal PET imaging. Quantitative imaging performance was evaluated in terms of spatial resolution and hot spheres recovery coefficients from image resolution and image quality phantoms representing the mouse. The data were reconstructed using algorithms of 2D filtered-back-projection, 2D ordered-subsets expectation maximization and maximum-a-posteriori. It is shown that the spatial resolution point spread function can be well explained by a double-gaussian function due to the generally long range of the positron. We show that, with the knowledge of the measured point spread functions, the accurate activity concentration in small lesions can be recovered when imaging with long-range positron emitters.