A method for determination of the timing stability of PET scanners

We report on the timing resolution and stability of the MicroPET R4 PET scanner. Its detectors have energy resolutions in the range of 25% and previously reported timing resolutions 3.2 ns. Our preliminary evaluation of this instrument showed that artefact-free normalization sinograms could only be obtained with a timing window of 10 ns or more in spite of a timing resolution of 3.2 ns. This instrument uses high-speed electronics albeit with 2-ns timing clock. We performed sham transmission scans with nothing in the field of view, and a range of timing windows from 2 to 14 ns and used a 14-ns timing blank scan to generate effective attenuation sinograms as a function of timing window. These showed trues count-rates which fit well to a ERF (/spl tau/) function. However, the effective attenuation value, which should be 1.0, changes from block to block and becomes very high in some blocks (>3.5 at 6 ns) suggesting the need for timing alignment. A method was devised to measure the timing stability to a precision better than the timing bin width of 2 ns.