Correspondence: Still no evidence for single photon detection by humans

The rod photoreceptors in the retina are known to be sensitive to single photons, but it has long been debated whether these single-photon signals propagate through the rest of the visual system and lead to perception. Recently, single-photon sources developed in the field of quantum optics have enabled direct tests of single-photon vision that were not possible with classical light sources. Using a heralded source based on spontaneous parametric downconversion to generate single photons which were sent to an observer at either an early or late time, Tinsley and Molodtsov et al. (2016) had observers judge when the photon was seen. Based on the above-chance accuracy in both a subset of high-confidence trials and in all post-selected trials, they claimed to show that humans can see single photons. However, we argue that this work suffers from three major issues: self-contradicting results, inappropriate statistical analyses, and a critical lack of statistical power. As a result, we cannot conclude that humans can see single photons based on the data of this study. We present a careful examination of the statistical analyses and the internal consistency of the data, which indicated that none of the key evidence holds.