Scratching bouts are modeled as Bernoulli trials until successful itch-extinguishing in mice

Itching and subsequent scratching behavior have been observed in many species, including humans. The behavior was evolved to remove skin parasites. Yet, scratching is performed without reliable indicators of whether a parasite is present. We addressed this apparent paradox by studying scratching in mice. Video recordings of ~5000 scratching bouts were collected in free-moving C57BL6/J mice. The statistical properties of their temporal sequence were analyzed. Inter-bout time intervals preceding over 50% of 5000 bouts were <10 s. We hypothesized that episodes of repetitive scratching corresponded to the duration of discrete events of itch sensation and comprised bouts separated by inter-bout intervals of <10 s. The distribution of itch episodes comprising n (n = 1, 2, 3, ...) scratching bouts was well-approximated by the geometric distribution with success probability = 0.5 in healthy mice and lower probability in dry skin mice. This suggests that scratching bouts are modeled by probabilistic Bernoulli trials, and their repetitive sequence in each episode continues until the itch is successfully extinguished. Accordingly, we can presume the presence of parasites from the repeat length of scratching bouts determined by the probability of successful itch-extinguishing. This may provide a promising stochastic model to assess itchy phenotypes.