Abstract
It has been shown that noise can improve signal detection in nonlinear systems (Wiesenfeld & Moss, Nature 1995, 373, 33-36). Recently we have shown that contrast detection sensitivity in human visual perception can be improved by superimposition of visual noise (Sasaki et al. Neurosci. Lett. 2006, 408, 94-97). Here we examined the effect of noise on signal detection using a temporal two-alternative forced-choice (2AFC) paradigm. Nineteen undergraduate students (10 male, 9 female, mean age of 19.8 years) with normal or corrected to normal vision participated in this study. Participants were asked to detect signal of a small light spot (white LED) with its intensity modulated by sine waves at frequency of 1Hz. Random flickering light modulated by white noise was superposed on the signal. The performance rate of correct response was 62.6 ± 3.7% at control without noise, increased to 77.3 ± 2.3% (t(18)=2.738, p<0.05) at noise intensity of -10 dB (referring noise threshold as 0 dB), then decreased to 47.3 ± 4.3% at 20 dB (t(18)=-2.870, p<0.05). These results confirmed our previous findings that addition of appropriate level of noise can improve signal detection in human visual perception. [J Physiol Sci. 2007;57 Suppl:S107]
