LIONirs toolbox design for fNIRS data analysis.
Poster No:
2009
Submission Type:
Abstract Submission
Authors:
Julie Tremblay1,2, Eduardo Martínez-Montes3, Alejandra Hüsser1,2, Laura Caron-Desrochers1,2, Phetsamone Vannasing1,2, Anne Gallagher1,2
Institutions:
1Université de Montréal, Montreal, Canada, 2CHU Sainte-Justine, Montreal, Canada, 3Cuban Center for Neuroscience, Havana, Cuba
First Author:
Co-Author(s):
Introduction:
Functional near-infrared spectroscopy (fNIRS) is a non-invasive neuroimaging technique. Analogous to functional magnetic resonance imaging (fMRI), it measures changes of cerebral blood oxygenation related to neuronal processes in cortical regions. In order to facilitate fNIRS data analysis, we developed a MATLAB toolbox and we describe here examples of functionalities. The LIONirs toolbox uses the MATLAB Batch System of SPM toolbox and provides a complete visual interface for data inspection and topographical representation during each stage of the analysis. This makes it a user-friendly tool for those researchers that do not have programming skills. We show this property as well as its flexibility and other capabilities by, creating a processing pipeline and applying it to data gathered from a group of 14 healthy subjects when they perform a passive story-listening task.
Methods:
fNIRS data was acquired using an Imagent oximeter (ISS) through 156 channels over bilateral frontal, temporal and parietal areas (sampling rate 20 Hz, laser wavelength 630 and 830nm). Fourteen healthy adult subjects listened passively to a story presented during 18 segments of 20 seconds interleaved with 20 seconds of rest.
fNIRS processing pipeline: Semi-automatic artifact detection is done with an automatized procedure which detects movement and muscular artifacts using a roll average method [1] and correlations among channels. Then, data decomposition methods such as Parallel Factor (PARFAC) analysis [2], [3], where applied to identify and remove the remaining artifacts which contaminated the raw signal. All artifacts were visually reviewed in the GUI and manually reject or correct if needed. Delta optical density transformation, low pass filtering 0.1 Hz and modified Beer-Lambert law were applied to obtain estimation of oxyhemoglobin (HbO) concentrations. Physiological artifacts measured by a short distance channel [4] and the model HRF response were regressed out using a general linear model (GLM) to estimate the individual response. Afterwards, a one-sample t-test with an FDR correction for multiple comparisons [5] was applied on the HRF estimate. Magnitude square coherence was computed on 200 random segments of 60 seconds longs to obtain individual functional connectivity matrices [6]. The average connectivity matrix describes the relationships of hemodynamic fluctuation (0.03 to 0.08 Hz) between channels and regions of interest.
fNIRS processing pipeline: Semi-automatic artifact detection is done with an automatized procedure which detects movement and muscular artifacts using a roll average method [1] and correlations among channels. Then, data decomposition methods such as Parallel Factor (PARFAC) analysis [2], [3], where applied to identify and remove the remaining artifacts which contaminated the raw signal. All artifacts were visually reviewed in the GUI and manually reject or correct if needed. Delta optical density transformation, low pass filtering 0.1 Hz and modified Beer-Lambert law were applied to obtain estimation of oxyhemoglobin (HbO) concentrations. Physiological artifacts measured by a short distance channel [4] and the model HRF response were regressed out using a general linear model (GLM) to estimate the individual response. Afterwards, a one-sample t-test with an FDR correction for multiple comparisons [5] was applied on the HRF estimate. Magnitude square coherence was computed on 200 random segments of 60 seconds longs to obtain individual functional connectivity matrices [6]. The average connectivity matrix describes the relationships of hemodynamic fluctuation (0.03 to 0.08 Hz) between channels and regions of interest.
Results:
Results reveal a significant increase in HbO in left temporal median (T=2.68, p<0.05), left temporal posterior (T=3.92 p<0.05) and right temporal median (T=6.01, p<0.05) areas during the passive story listening task showing the auditory (temporal median bilateral) and receptive language (left temporal posterior close to Wernicke area) processes. Figure 1 illustrates the processing pipeline for the passive story listening paradigm until the group level analysis, as well as the typical visualization of intermediate results produced by the LIONirs Toolbox. Figure 2 shows the steps followed for functional connectivity analysis. The average connectivity matrix during the paradigm showed highest coherence network (COH > 0.3) among the fronto-temporal (F7,T3,T5,F8), the somatosensory (C3,F3,C4,F4) and pre-frontal (Fp1,Fp2) channels [7].
·Fig. 1 Illustration of the processing pipeline created in the LIONirs toolbox. A) Detailed hierarchical scheme of the different steps of analysis. B) Subject-level analysis. C) Group-level analysis.
·Fig.2 Functional connectivity in LIONirs. A) Detailed procedure. B) Regions of interest. C) The averaged coherence matrix. D) The connectogram illustrate links across regions of interest (COH>0.3).
Conclusions:
Standardization of data inspection and correction as well as data analysis are crucial steps for a valid interpretation of fNIRS data. The LIONirs toolbox uses a MATLAB Batch System to create data analysis templates easy to replicate to large datasets and provides a visual interface for data inspection and topographical representation during any stage of the analysis. We present here an example of application using a passive story listening paradigm; the task related HRF is predominant in the temporal lobe, similar result where observe in the fMRI literature [8]. This illustrates the techniques implemented, the flexibility and general usefulness of the proposed toolbox.
Modeling and Analysis Methods:
Exploratory Modeling and Artifact Removal
Motion Correction and Preprocessing 2
Neuroinformatics and Data Sharing:
Informatics Other
Novel Imaging Acquisition Methods:
NIRS 1
Keywords:
Design and Analysis
Development
Informatics
Workflows
1|2Indicates the priority used for review
My abstract is being submitted as a Software Demonstration.
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Was any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
Which processing packages did you use for your study?
Provide references using author date format
[2] Bro, R. and Kiers, H.A.L., (2003), 'A new efficient method for determining the number of components in PARAFAC models', Journal of Chemometrics, vol. 17, no 5, p. 274‑286.
[3] Hüsser, A., Caron-Desrochers, L., Tremblay, J., Vannasing, P., Martínez-Montes, E., Gallagher, A. (2019), 'Parallel Factor Analysis (PARAFAC) for multidimensional decomposition of fNIRS data - A validation study', bioRxiv, p. 806778
[4] Gagnon, L. , Cooper, R.J., Yücel, M.A., Perdue, K.L., Greve, D.N., Boas, D.A.,(2012), 'Short separation channel location impacts the performance of short channel regression in NIRS', Neuroimage, vol. 59, no 3, p. 2518‑2528.
[5] Benjamini, Y., Yekutieli, D., (2001), 'The control of the false discovery rate in multiple testing under dependency', Ann. Statist., vol. 29, no 4, p. 1165‑1188.
[6] Kida, T., Tanaka, E., Kakigi, R., (2016), 'Multi-Dimensional Dynamics of Human Electromagnetic Brain Activity', Front. Hum. Neurosci., vol. 9.
[7] Bellec, P., Rosa-Neto, P., Lyttelton, O.C., Benali, H., Evans, A.C., (2010), 'Multi-level bootstrap analysis of stable clusters in resting-state fMRI', NeuroImage, vol. 51, no 3, p. 1126‑1139.
[8] Vannest, J.J., Karunanayaka, P.R., Altaye, M., Schmithorst, V.J., Plante, E.M., Eaton, K.J., Rasmussen, J.M., Holland, S.K., (2009), 'Comparison of fMRI data from passive listening and active-response story processing tasks in children', J Magn Reson Imaging, vol. 29, no 4, p. 971‑976.