Browsing by Subject "binomial logistic regression"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Alcohol-Naïve USVs Distinguish Male HAD-1 from LAD-1 Rat Strains
    (Elsevier, 2017) Mittal, Nitish; Thakore, Neha; Reno, James M.; Bell, Richard L.; Maddox, W. Todd; Schallert, Timothy; Duvauchelle, Christine L.; Psychiatry, School of Medicine
    Ultrasonic vocalizations (USVs) are mediated through specific dopaminergic and cholinergic neural pathways and serve as real-time measures of positive and negative emotional status in rodents. Although most USV studies focus primarily on USV counts, each USV possesses a number of characteristics shown to reflect activity in the associated neurotransmitter system. In the present study, we recorded spontaneously emitted USVs from alcohol-naïve high alcohol drinking (HAD-1) and low alcohol drinking (LAD-1) rats. Using our recently developed WAAVES algorithm we quantified four acoustic characteristics (mean frequency, duration, power and bandwidth) from each 22 – 28 kHz and 50 – 55 kHz frequency modulated (FM) USV. This rich USV representation allowed us to apply advanced statistical techniques to identify the USV acoustic characteristics that distinguished HAD-1 from LAD-1 rats. Linear mixed models (LMM) examined the predictability of each USV characteristic in isolation and linear discriminant analysis (LDA) and binomial logistic regression examined the predictability of linear combinations of the USV characteristics as a group. Results revealed significant differences in acoustic characteristics between HAD-1 and LAD-1 rats in both 22 – 28 kHz and 50 – 55 kHz FM USVs. In other words, these rats selectively bred for high- and low-alcohol consumption can be identified as HAD-1 or LAD-1 rats with high classification accuracy (approx. 92-100%) exclusively on the basis of their emitted 22-28 kHz and 50-55 kHz FM USV acoustic characteristics. In addition, acoustic characteristics of 22 – 28 kHz and 50 – 55 kHz FM USVs emitted by alcohol-naïve HAD-1 and LAD-1 rats significantly correlate with their future alcohol consumption. Our current findings provide novel evidence that USV acoustic characteristics can be used to discriminate between alcohol-naïve HAD-1 and LAD-1 rats, and may serve as biomarkers in rodents with a predisposition for, or against, excessive alcohol intake.