Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

Back

Biomedical Imaging and Instrumentation

Comparison Between Two Different fNIRs Analysis Models for Prefrontal and Vestibular Cortex Activation

(D-155) Comparison Between Two Different fNIRs Analysis Models for Prefrontal and Vestibular Cortex Activation

Saturday, October 14, 2023

9:30 AM – 10:30 AM PDT

Location: Exhibit Hall - Row D - Poster # 155

Presenting Author(s)

TW

Thomas A. Wenzel

Undergraduate Research Assistant
Boise State University
Boise, Idaho, United States

Co-Author(s)

DL

Daniel Lundy

Student Researcher
University of Maryland, Baltimore County
Columbia, Maryland, United States
BS

Brian Sylcott

Professor
East Carolina University, United States

Primary Investigator(s)

CL

Chia-Cheng Lin

Professor
East Carolina University, United States

Introduction::

Functional near-infrared spectroscopy (fNIRs) is a neuroimaging technique that utilizes low levels of near-infrared light (650-950 nm) to measure changes in hemodynamics of the brain. FNIRs data can be analyzed using different toolbox interfaces, including the NIRS Brain AnalyzIR toolbox (AnalyzIR). The AnalyzIR toolbox is a MatLab based open-source analysis package. Previous research has found that the robust pre-whitened autoregressive coherence statistical model within the AnalyzIR toolbox is effective at detecting and removing artificials, which may decrease the false positive rate [2]. However, when using this toolbox to analyze the vestibular cortex, the elongation of the response to sensory integration needs to be considered [1,3]. The purpose of this study is to investigate fNIRs analysis with and without vestibular cortex activation adjustment to confirm fNIRS data analysis for our sensory integration studies.

Materials and Methods:: Seven healthy, right-handed participants (34.9: ± 4.99, 1M/6F) were recruited to complete six optic flow tasks. Optic flow (OF) is the apparent motion in the visual field, previous research showed OF may induce perceived self-motion [4]. In this study, block design (A-B-A-B-A) was used with baseline (A) and stimulus (B). Each block was 30 seconds. OF was used as stimulus, which was represented by objects resembling stars and administered using a VR headset (HTC Vive with Tobii eye-tracking), at six different speeds (0, 5, 10, 15, 20, 25 m/s). Two sets of 64-channel fNIRs devices (NIRSort, NIRx, German) were used to monitor prefrontal (PFC) and vestibular cortexes (VEST) activation during each task. Brain activity data was collected, then processed using the NIRS Brain AnalyzIR toolbox [1]. All data was processed with pre-whitened autoregressive (AR-IRLS model) coherence statistical model for artificial correction. The AR-IRLS model also computed with and without vestibular cortex activation adjustment in the subject level analysis. The group level used the random-effect regression model. The Benjamin-Hochberg procedure will be used to adjust the p-value to decrease the false discovery rate (FDR). The significant level was set as FDRp< 0.05. The receiver operating characteristic (ROC) curve was used to compare false positive rates (FPR) with and without vestibular cortex adjustment.

Results, Conclusions, and Discussions:: The ROC curve showed that both models (non-adjusted and vestibular-adjusted) had the same control for type-1 errors (Fig. 1.a) and the actual FPRs (Fig. 1.b) of the data. These results demonstrated that adjusting the regression model for the vestibular cortex response time had no effect on the accuracy of data analysis. However, the vestibular adjustment did affect the computing of hemoglobin changes in PFC and VEST activation. The results showed significant differences between speeds (5 and 10, 10 and 15, 15 and 20 m/s FDRp < 0.04) in the right and left PFC for the non-adjusted model. The non-adjusted model also showed significant difference in the right VEST between speeds (0 and 25, 10 and 25 m/s FDRp < 0.01), and in the left VEST between speeds (20 and 25, 5 and 15, 15 and 20 m/s FDRp < 0.05), as well as speeds (0 and 15, 10 and 15 m/s FDRp < 0.01) for both sides of the VEST (Fig. 2.c). The vestibular-adjusted model showed significant differences in the PFC between speeds (0 and 5, 0 and 10, 0 and 15, 0 and 20, 0 and 25 m/s FDRp < 0.01) on both sides, while also showing difference between speeds (5 and 15, 5 and 25, 20 and 25 m/s FDRp < 0.01) on the right PFC. The VEST showed difference on both sides between speeds (0 and 15, 0 and 20, 5 and 15, 5 and 20, and 5 and 25 m/s (FDRp < 0.04)) when using the vestibular-adjusted model, also showing significance between speeds (10 and 15, 15 and 20 m/s FDRp < 0.04) and (0 and 25, 10 and 25 m/s FDRp < 0.01) on the right and left respectively (Fig 2.d). The non-adjusted model showed a decrease in PFC and VEST activation during all OF speeds (Fig. 2.c), while the vestibular-adjusted model showed an increase of brain activity as OF speed increased (Fig 2.d). The vestibular-adjusted model provides results that have more significant differences in the PFC and VEST, between speeds, which is an activation response that corresponds to our hypothesis regarding increasing stimuli.

Acknowledgements (Optional): : This material is based upon work supported by the National Science Foundation under Grant No. 1950507. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

References (Optional): :

[1] Santosa, H. et al, Algorithms (2018) 11(5), 73.
[2] Santosa H., J Biomedical Optics (2017) 22(5).
[3] Dieterich M. et al, Brain (2008) 131(10), 2538-2552
[4] Palmisano, S. et al, Front. Psychol. (2015) 6, 93.