Classification a sensorimotor task level of complexity for athletes based on physiological indicators using machine learning methods
Authors:
1.
Federal Research Center for Innovator and Emerging Biomedical and Pharmaceutical Technologies
(Rehabilitation and Sport Psychophysiology Lab, Leading Researcher)
Moscow Center of Advanced Sports Technologies (Sports Psychology Department, physiologist)
Moscow, Russian Federation
Moscow Center of Advanced Sports Technologies (Sports Psychology Department, physiologist)
Moscow, Russian Federation
Type:
Article
Pages:
from 17
to 20
Status:
Published
Received:
18.12.2023
Accepted:
22.01.2024
Published:
13.03.2024
Subject area:
1.5.5
5.3.3
1.2.1
5.8.4
3.1.33
159.91
159.93
004.93'14
355.233.22
796.012
77.00
20.00
02.00.00
06.00.00
09.00.00
32.00.00
44.00.00
28
73
75
5
COM014000
COM018000
BIO016000
PSY012000
5.3.3
1.2.1
5.8.4
3.1.33
159.91
159.93
004.93'14
355.233.22
796.012
77.00
20.00
02.00.00
06.00.00
09.00.00
32.00.00
44.00.00
28
73
75
5
COM014000
COM018000
BIO016000
PSY012000
Language:
Russian
Keywords:
task complexity, classifiers, athletes, autonomic indicators, machine learning
Abstract (English):
The study aimed to identify the most sensitive autonomic indicators reflecting the level of complexity of the sensorimotor task performed by athletes using various machine learning methods (classification algorithms). As tasks of two levels of difficulty, we used the audio-motor synchronization task: to tap in synchrony with a metronome rhythmic sound (a simple task) and to tap the same rhythm without auditory cues (rhythm memory task, a complex task). Heart rate, respiratory parameters, skin conduction, and EEG were recorded. The most accurate classification was demonstrated by the Classification and Regression Trees (C&RT) model – the error was 18.3%.
The study aimed to identify the most sensitive autonomic indicators reflecting the level of complexity of the sensorimotor task performed by athletes using various machine learning methods (classification algorithms). As tasks of two levels of difficulty, we used the audio-motor synchronization task: to tap in synchrony with a metronome rhythmic sound (a simple task) and to tap the same rhythm without auditory cues (rhythm memory task, a complex task). Heart rate, respiratory parameters, skin conduction, and EEG were recorded. The most accurate classification was demonstrated by the Classification and Regression Trees (C&RT) model – the error was 18.3%.
Keywords:
task complexity, classifiers, athletes, autonomic indicators, machine learning
task complexity, classifiers, athletes, autonomic indicators, machine learning
1. Charles R.L., Nixon J. Measuring mental workload using physiological measures: A systematic review. Applied Ergonomics, 2019, 74, pp. 221-232. DOI: https://doi.org/10.1016/j.apergo.2018.08.028
2. Zhang J., Yin Z., Cheng P., Nichele S. Emotion recognition using multi-modal data and machine learning techniques: A tutorial and review. Information Fusion, 2020, 59 (1), pp. 103-126. DOI: https://doi.org/10.1016/j.inffus.2020.01.011
3. Tarvainen M.P., Niskanen J.-P., Lipponen J.A., Ranta-Aho P.O., Karjalainen P.A. Kubios HRV-heart rate variability analysis software. Computer Methods and Programs in Biomedicine, 2014, 113 (1), pp. 210-220. DOI: https://doi.org/10.1016/j.cmp
