Related Experiment Video
Updated: Sep 29, 2025
06:34
Infant Auditory Processing and Event-related Brain Oscillations
Published on: July 1, 2015
16.6K
Oscillator decomposition of infant fNIRS data
Takeru Matsuda1, Fumitaka Homae2,3, Hama Watanabe4
1RIKEN Center for Brain Science, RIKEN, Wako, Japan.
Plos Computational Biology
|March 24, 2022
Summary
This study introduces the oscillator decomposition method (OSC-DECOMP) to analyze infant functional near-infrared spectroscopy (fNIRS) data, revealing distinct brain oscillation patterns and functional connectivity in sleeping infants.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Functional near-infrared spectroscopy (fNIRS) measures brain hemodynamic responses using oxygenated (oxy-Hb) and deoxygenated hemoglobin (deoxy-Hb) time series.
- Analyzing oscillatory changes in infant fNIRS signals is crucial for understanding brain development.
- Existing methods often require arbitrary band-pass filter selection, limiting data-driven analysis.
Purpose of the Study:
- To investigate oscillatory changes in infant fNIRS signals using a novel statistical method.
- To apply the oscillator decomposition method (OSC-DECOMP) for data-driven extraction of oscillation components.
- To analyze frequency-specific functional connectivity in sleeping 3-month-old infants.
Main Methods:
- Applied OSC-DECOMP, a Gaussian linear state space model-based method, to 18-channel fNIRS data from 21 sleeping infants.
- Extracted 5-7 oscillators per channel, identifying frequency peaks around 0.01-0.1 Hz, 1.6-2.4 Hz, and 3.6-4.4 Hz.
- Utilized Akaike Information Criterion (AIC) to compare state space models and confirm common oscillatory activity.
Main Results:
- Identified three distinct oscillator frequency peaks, with the lowest (0.01-0.1 Hz) linked to brain activity and hemodynamic changes.
- Attributed the second peak (1.6-2.4 Hz) to cardiac pulse waves and noise, and the third (3.6-4.4 Hz) as its harmonic.
- Demonstrated that OSC-DECOMP successfully separated oscillators with similar frequencies based on their unique projection patterns.
- Revealed frequency-specific functional connectivity, with brain oscillators showing connectivity while pulse/noise oscillators were spatially homogeneous.
Conclusions:
- OSC-DECOMP effectively decomposes biological time series data into meaningful oscillation components without arbitrary filtering.
- The method distinguishes between neural oscillations, physiological noise, and artifacts in infant fNIRS signals.
- OSC-DECOMP is a promising tool for advancing the analysis of infant brain activity and functional connectivity using fNIRS data.
