Journal article
Wavelet Coherence-Based Ambient Masking for Noise Removal in Functional Near-Infrared Spectroscopy Measurements
IEEE transactions on instrumentation and measurement
22 Sep 2026
Abstract
Functional near-infrared spectroscopy (fNIRS) has been widely used for monitoring brain activity in naturalistic settings, but the measurements are prone to non-stationary optical interference from external hardware such as eye trackers and virtual reality (VR) headsets. Common filtering approaches to tease out certain frequency bands are often inadequate when instrumental noise overlaps with the fNIRS frequency bands of interest. Hence, this study introduces Wavelet Coherence-Based Ambient Masking (WCA), a time-frequency denoising approach that uses the ambient light as a reference channel to suppress coherent interference in fNIRS measurements. An iterative thresholding strategy was implemented to adapt the masking level to the noise profile of each measurement channel. Data from two studies were processed to validate the efficacy of the WCA approach. In a controlled phantom study, a simulated hemodynamic response was generated with interference from an eye tracker. The WCA approach improved signal-to-noise ratio by up to 19.93 dB at 730 nm and by up to 15.44 dB at 850 nm light intensity measures relative to low-pass filtering only. Furthermore, WCA significantly outperformed conventional ambient-reference techniques, including ordinary least squares (OLS) regression and normalized least mean squares (NLMS) adaptive filtering, suggesting improved signal recovery under severe interference. In a human study using VR-based Basic Life Support training (n=19), this WCA denoising method reduced inter-subject standard deviation in oxygenated hemoglobin (HbO) by 62% during baseline and 41% during the cognitive task. This variance reduction increased the effect size and improved cognitive-state classification performance, i.e., the area under the curve (AUC) from Receiver Operating Characteristic (ROC) analysis improved from 0.69 to 0.74. These findings demonstrate that WCA is an effective method for removing non-stationary ambient interference and can help to improve the robustness of fNIRS measurements in immersive multimodal environments.
Metrics
1 Record Views
Details
- Title
- Wavelet Coherence-Based Ambient Masking for Noise Removal in Functional Near-Infrared Spectroscopy Measurements
- Creators
- Mert Deniz Polat - Drexel University, School of Biomedical Engineering and ScienceAshley Marie Bishop - Drexel University, School of Biomedical Engineering and ScienceKurtulus Izzetoglu - Drexel University, School of Education
- Publication Details
- IEEE transactions on instrumentation and measurement
- Publisher
- IEEE
- Resource Type
- Journal article
- Language
- English
- Academic Unit
- School of Biomedical Engineering and Science
- Other Identifier
- 991022234043104721