A Demonstration of Machine Learning in Detecting Frequency Following Responses in American Neonates

Abstract

In this study, we sought to evaluate the efficiencies of multiple machine learning algorithms in detecting neonates’ Frequency Following Responses (FFRs). We recorded continuous brainwaves from 43 American neonates in response to a pre-recorded monosyllable/i/with a rising frequency contour. Recordings were classified into response and no response categories. Six response features were extracted from each recording and served as predictors in FFR identification. Twenty-three supervised machine learning algorithms with mean efficiency values of 86.0%, 94.4%, 97.2%, and 97.5% when 1, 10, 100, and 1000 random iterations were implemented, respectively. These high efficiency values obtained from the neonatal FFRs demonstrate that machine learning algorithms can help assess pitch processing in neonates and can be applied to auditory screening and intervention services for neonates at risk for disorders associated with decreased pitch processing.

Implementation of Machine Learning on Human Frequency-Following Responses: A Tutorial

“…Applicability : This model has been successfully implemented in FFR research. 20 62 66 68 69 Although the SVM model is originally intended for binary classification, it can be easily extended to multiclass classification. A soft margin can be used to widen the margin and thus enhance separation results.…”

“…A decision tree classifier 72 is made of nodes and branches where each node evaluates a specific feature of the dataset and makes a best split. 65 66 Because this procedure is executed recursively, a sequence of nodes and branches will be created, and thus forming a tree with the leaf nodes representing a class label. Decision trees are simple to implement and easy to interpret but may not be suitable for more complex tasks.…”

“…To date, a wide spectrum of ML techniques has been used on several FFR projects. 20 34 62 65 66 67 68 69 70 71 Table 1 provides a summary of those projects with an emphasis on their ML-related research questions, model designs, approaches, and findings.…”

“…64 Due to the large number of viable computational algorithms, it is impractical to include all of them in one article. As such, this article focuses on the algorithms and models that have been reported in the recent FFR literature 20,34,62,[65][66][67][68][69][70][71] or which, in the authors' opinion, may be useful in furthering FFR research. To begin with, three basic ML terminologies are defined as the following:…”

“…This practice dataset is based on Hart and Jeng. 66 These scripts and dataset are accessible from the second author's (Y-S.J.) GitHub repository: https://github.com/qxxxb/ffr_ml_tutorial.…”