Using 1/ f Scaling to Study Variability and Dexterity in Simple Tool using Tasks

Abstract

1/f scaling quantifies the relationship between power spectral density and frequency of a signal by fitting a linear regression model to log-transformed data. Where the fitted slope is zero, the signal is assumed to be white noise arising from a random source, but where there is a negative slope, the signal is assumed to be pink noise arising from a source with metastability. The concept of metastability provides a very useful way of thinking about variability in performance in a dynamic systems framework. In this paper, we demonstrate the application of 1/f scaling to the study of simple tool-using tasks, with the intention of studying how activity exhibits consistency and variability across individuals and across task demands.