Association of Musical Training with Auditory and Speech Neural Coding and Perception
A number of studies have suggested that intensive musical training is related to better perception and neural coding of sound. Published musician advantages range from enhanced speech perception in crowded environments to enhanced representation of a sound’s frequency in the brainstem. Researchers in CATSS are collaborating with a team of over 20 scientists across 6 universities to replicate and extend 8 key published studies on musician advantages across a large sample of over 300 participants. The study, "Association of musical training with auditory and speech neural coding and perception" is led by UMN principal investigator, Dr. Andrew Oxenham, and project coordinator, Dr. Kelly Whiteford. This project regularly uses the equipment in CATSS, which has been a key resource for collecting data at UMN.
Why bother replication research?
Most previous studies on musician advantages are limited in several important ways: (1) They have used relatively small sample sizes (e.g., n=20 or n=30), (2) only compared professional musicians to non-musicians but not people with moderate amounts of training, and (3) have primarily tested college-aged listeners. These types of samples are useful in that they are likely to find group differences if they exist, but they are limited in that they are not generalizable to people outside the type of population tested. The present study will be able to more clearly assess both whether the musician advantage replicates and whether the musician advantage generalizes beyond young, professional musicians. This set of experiments is part of a National Science Foundation grant to support large-scale replication efforts in neuroscience.
Efforts toward open science
One unique aspect of this study is that all materials, including experiment code, de-identified data, and analyses, will be made free and accessible to the public via the project website on the Open Science Framework (OSF), with periodic updates posted on a study blog Sound Brain Science. The methods and direct replication analyses were preregistered prior to beginning data collection, meaning they were written and timestamped on OSF. Such transparency is important because it allows the public to have a more detailed view of the scientific process, including what decisions were made prior to data collection. Providing open access to the study data once the study is complete will give other scientists an opportunity to reproduce the study results or use the data in exploratory follow-up analyses.
Implementing a multi-site setup
A major challenge for multi-site studies is to ensure all experiments are working in the same manner across sites. The experimental code was written by Dr. Kelly Whiteford and Dr. Juraj Mesik at UMN and subsequently shared with all sites. Dr. Whiteford then traveled to each site, including Boston University, Carnegie Mellon University, Purdue University, University of Rochester, and University of Western Ontario, to help with experimental setup. This semester, Dr. Whiteford will finish the second round of travel to each site to ensure study protocols are followed uniformly across sites during data collection.
Engaging UMN students
UMN has so far collected data from 45 participants out of a planned total of 65 people, across two to three 2-hour sessions. The experimental protocol is complex and requires a group of well-trained researchers to implement data collection efficiently. Undergraduate research assistants Angela Sim and Kara Stevens are funded by NSF and have been a huge help in keeping the UMN team on schedule to finish data collection by June 2020.
Dr. Whiteford wears a 32-electrode EEG cap in one of the CATSS sound booths. Undergraduate research assistants Angela Sim and Kara Stevens tested Dr. Whiteford on the full experimental protocol multiple times before study code was shared between sites.
For more information and updates, see Dr. Whiteford’s study blog: www.soundbrainscience.com/blog