Binaral rivalry.

**(A)** All twelve subjects tested experienced switches between perceiving predominantly ‘rose’ and predominantly ‘marker’ (y-axis, similarity rating to ‘rose’ or ‘marker’ on a 100-unit visual analogue scale as shown in Figure 1B) over 20 intermittent samplings (x-axis) of PEA and n-butanol, one presented to each nostril. Dots above the middle line indicate an olfactory percept of predominantly ‘rose’. Dots below the middle line indicate an olfactory percept of predominantly ‘marker’.

**(B)** Illustration of the visual analogue scale used for olfactory similarity ratings.

**(C)** Histogram of the mean similarity ratings across the 20 samplings from the 12 subjects. How biased one was towards perceiving ‘rose’ or ‘marker’, as reflected by his/her mean similarity rating, follows a normal distribution with the mean at 53.9% similar to ‘marker’.

**(D)** Histogram of the similarity ratings (240 ratings from 12 subjects, each with 20 samplings). The distribution can be modeled with the sum of two normal distributions (dotted curve):

, where

*h*_{1}, μ

_{1}, σ

_{1} are the height, mean, and standard deviation, respectively, of the first normal distribution, and

*h*_{2}, μ

_{2}, σ

_{2} are the height, mean, and standard deviation, respectively, of the second normal distribution. Here μ

_{1} corresponds to 66% similar to ‘marker’, and μ

_{2} corresponds to 65% similar to ‘rose’.

## PubMed Commons