Stimulus reconstruction from the biphasic integrate-and-fire sampler

An algorithm to recover the encoded stimulus from a biphasic integrate-and-fire model has been proposed using the sampling framework for local averages. The algorithm converges for sparse sampling sets, common in extracellular neural recordings. Moreover, the spike region can be recovered with up to 50 dB accuracy even though the instantaneous and overall pulse rates are below the Nyquist bound.