|
Activity Number:
|
361
- Contributed Poster Presentations: Section on Nonparametric Statistics
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2020 : 10:00 AM to 2:00 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
|
Abstract #312332
|
|
|
Title:
|
Quantifying Activity Patterns Using Non-Parametric Approaches in a Ferret Study
|
|
Author(s):
|
Thaddeus Haight* and Susan Schwerin and Sharon Juliano
|
|
Companies:
|
Center for Neuroscience and Regenerative Medicine and Uniformed Services University and Uniformed Services University
|
|
Keywords:
|
Loess;
actigraphy;
activity ;
smoothing function
|
|
Abstract:
|
Actigraphy represents a non-invasive method for collection of resting/activity patterns but quantification of the acquired data represents an important analytic challenge. We applied non-parametric methods to examine activity patterns in ferrets that experienced different experimental conditions. Locally estimated scatterplot smoothing (Loess) was applied to actigraphy counts for each individual animal in a 24-hour period. Loess generates a smoothing curve that is a compilation of locally fit regression lines across the data. A span parameter is used to control the adjacent data points that contribute to each locally fit regression and contributes to the smoothness of the resultant curve. Loess was applied using: 1) activity counts > 0, assuming only non-zero counts varied across animals, and 2) the same span for each animal. 95% confidence limits (CL) were calculated. Metrics were computed for each animal based on counts where the lower 95 CL exceeded a given threshold for activity, and were compared for the experimental groups. Loess offers a flexible approach for differentiating activity patterns from actigraphy data that can be implemented easily and has interpretative value.
|
Authors who are presenting talks have a * after their name.
