Disjunct matrices for multiplexing PET detector signal readout

The readout channels of radionuclide imaging detectors are commonly multiplexed to reduce the data acquisition system complexity. Because the underlying detector signals have a sparse representation in time and space, sparse recovery methods may be applied. A new method multiplexing method is investigated here. We developed algorithms for creating practical-to-implement disjunct matrices and decoding methods that can accommodate light sharing in PET. Similar to cross-strip multiplexing, disjunct multiplexing is implemented with unweighted interconnections of pixels to the readout channel. Whereas cross-strips read out rows and columns, disjunct multiplexing readouts are interconnected to satisfy a disjunctness criterion. Methods are developed to create matrices that are easier to route. Disjunct multiplexing is evaluated by Monte Carlo simulation and shown to have superior SNR than cross-strip multiplexing at the same multiplexing ratio. Consequently, better energy resolution and positioning accuracy can be achieved. Further, disjunct multiplexing is shown to be highly accurate at positioning multiple interactions.