Hierarchical Bayesian Models for Analysis of Isochronous Sensorimotor Synchronization Data
-
1
Lund University, Lund University Cognitive Science, Sweden
Two of the most common measures analyzed in an isochronous sensorimotor synchronization task is constant error and timing variability. A straight forward way to estimate these measures is to calculate the sample mean and sample variance of the tone-to-tap asynchronies. However straight forward and common this approach is it does not necessary generate the most accurate estimates.
When the tempo of the pacing tones is slow it is possible to show that using the sample mean and sample variance will regularly underestimate both the constant error and the timing variance. This is because the distribution of the asynchronies becomes increasingly skewed and non-normal as participants start producing reactive responses rather than anticipating responses at long inter stimulus intervals ( >2000 ms).
Bayesian statistics is an increasingly popular statistical method in psychology that facilitates modeling the non-normal asynchrony distribution arising at slow tempi. By using hierarchical Bayesian modeling estimates can be further improved in the common situation where an experiment involves multiple participants.
A hierarchical Bayesian model was implemented using a censored normal distribution to model the distribution of timing asynchronies. Estimates made by the model were compared with the classical methods of estimating constant error and timing variability both using simulated data and experimental data from Repp & Doggett (2007). Using the Bayesian hierarchical model resulted in considerable less bias at slow tempo and outperformed classical methods with regards to accuracy. A freely available implementation of the model in the R programming language is available at https://github.com/rasmusab/bayes_timing.
Keywords:
Bayesian Models,
Rhythm perception,
variability,
constant error,
sensorimotor synchronization
Conference:
14th Rhythm Production and Perception Workshop Birmingham 11th - 13th September 2013, Birmingham, United Kingdom, 11 Sep - 13 Sep, 2013.
Presentation Type:
Oral Presentation
Topic:
Rhythm Production and Perception
Citation:
Bååth
R
(2013). Hierarchical Bayesian Models for Analysis of Isochronous Sensorimotor Synchronization Data
.
Conference Abstract:
14th Rhythm Production and Perception Workshop Birmingham 11th - 13th September 2013.
doi: 10.3389/conf.fnhum.2013.214.00023
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
18 Jul 2013;
Published Online:
24 Sep 2013.
*
Correspondence:
Mr. Rasmus Bååth, Lund University, Lund University Cognitive Science, Lund, Sweden, rasmus.baath@lucs.lu.se