Mining Massive Archives of Mice Sounds with Symbolized Representations

Abstract

Many animals produce long sequences of vocalizations best described as "songs." In some animals, such as crickets and frogs, these songs are relatively simple and repetitive chirps or trills. However, animals as diverse as whales, bats, birds and even the humble mice considered here produce intricate and complex songs. These songs are worthy of study in their own right. For example, the study of bird songs has helped to cast light on various questions in the nature vs. nurture debate. However, there is a particular reason why the study of mice songs can benefit mankind. The house mouse (Mus musculus) has long been an important model organism in biology and medicine, and it is by far the most commonly used genetically altered laboratory mammal to address human diseases. While there has been significant recent efforts to analyze mice songs, advances in sensor technology have created a situation where our ability to collect data far outstrips our ability to analyze it. In this work we argue that the time is ripe for archives of mice songs to fall into the purview of data mining. We show a novel technique for mining mice vocalizations directly in the visual (spectrogram) space that practitioners currently use. Working in this space allows us to bring an arsenal of data mining tools to bear on this important domain, including similarity search, classification, motif discovery and contrast set mining.

Exact Primitives for Time Series Data Mining

“…The ordering is typically temporal; however other kinds of data such as color distributions [41], shapes [104] and spectrographs [117] also have a well defined ordering and can fruitfully be considered "time series" for the purpose of indexing and mining. It is possible there could be variable time spacing between successive points in the series.…”

Parameter-Free Audio Motif Discovery in Large Data Archives

“…Some representative examples include:  Acoustic wildlife monitoring has been shown to allow effective and non-invasive measurement of the health of ecosystems [36].  A powerful tool for investigating the role of genetics in human disorders modifies ("knocks out") various genes in mice and examines their vocalizations for changes that may be linked to those genes, and hence the analogue genes in humans [35] [40]. Figure 1 hints at the utility of this idea, which we revisit in Section V.D.…”

“…The problem with this approach is that the feature extraction step must be highly tuned to the domain. For example, Zakaria et al [40] demonstrate a technique to find motifs in vocalizations of a specific strain of lab mice called Fmr1-KO. However, it is not clear if this multi-stage algorithm (which requires significant human intervention) generalizes to other strains of mice, much less to other rodents.…”

“…Given a high quality symbolic representation of the data, the problem becomes much simpler; we can just use an off-the-shelf symbolic repeated pattern discovery tool. This approach has been used in music [3] and in mice vocalizations [40]. However, it is obvious that the symbol extraction algorithms used for pop songs are unlikely to generalize to classical music, much less mice or insects [40].…”

“…This approach has been used in music [3] and in mice vocalizations [40]. However, it is obvious that the symbol extraction algorithms used for pop songs are unlikely to generalize to classical music, much less mice or insects [40]. Likewise, it is not clear that the symbol extraction method discussed in [40] will generalize to other strains of mice, much less other mammals.…”