Skip to main content
Fig. 2 | BMC Biology

Fig. 2

From: TrackUSF, a novel tool for automated ultrasonic vocalization analysis, reveals modified calls in a rat model of autism

Fig. 2

Automated analysis of mouse mating calls using TrackUSF. A 3D t-SNE analysis of all USFs recorded from three C57BL/6J and three BalbC male-female pairs. Each USF is represented by a dot, color-coded for the strain. Black numbers represent the distinct clusters, manually defined by the drawn gray lines. Note the clear separation of cluster 1, which includes non-vocal signals defined as noise. B Examples of spectrograms showing USFs from all clusters, each marked as the number of the cluster it is linked to, superimposed by the TrackUSF software on their corresponding non-vocal (i) or vocal (ii–iv) signals. C PSD analysis of the distinct clusters shown in A. The total number of USFs in each cluster is detailed in the legend. Note the unique profile of cluster 1, which is mainly enriched with non-vocal signals (noise). D Above: co-localization of USFs (each of which is represented by a single colored dot) and manually defined USVs (black lines, the length of which is proportional to the USV duration) during a whole 10-min-long audio recording of mouse mating calls. Note the various sequences of USVs, separated by prolonged silent periods. Below — one USV sequence displayed in higher resolution, with the co-localized USFs (excluding cluster 1). Note the accurate detection of most USVs by the various types of USFs

Back to article page