Independent component analysis of broadband near-infrared spectroscopy data acquired on adult human head

A deep convolutional neural network for estimating hemodynamic response function with reduction of motion artifacts in fNIRS

OBJECTIVE

Functional near-infrared spectroscopy (fNIRS) is a neuroimaging technique for monitoring hemoglobin concentration changes in a non-invasive manner. However, subject movements are often significant sources of artifacts. While several methods have been developed for suppressing this confounding noise, the conventional techniques have limitations on optimal selections of model parameters across participants or brain regions. To address this shortcoming, we aim to propose a method based on a deep convolutional neural network (CNN).

APPROACH

The U-net is employed as a CNN architecture. Specifically, large-scale training and testing data are generated by combining variants of hemodynamic response function (HRF) with experimental measurements of motion noises. The neural network is then trained to reconstruct hemodynamic response coupled to neuronal activity with a reduction of motion artifacts.

MAIN RESULTS

Using extensive analysis, we show that the proposed method estimates the task-related HRF more accurately than the existing methods of wavelet decomposition and autoregressive models. Specifically, the mean squared error and variance of HRF estimates, based on the CNN, are the smallest among all methods considered in this study. These results are more prominent when the semi-simulated data contains variants of shapes and amplitudes of HRF.

SIGNIFICANCE

The proposed CNN method allows for accurately estimating amplitude and shape of HRF with significant reduction of motion artifacts. This method may have a great potential for monitoring HRF changes in real-life settings that involve excessive motion artifacts.

Measurement of Adult Human Brain Responses to Breath-Holding by Multi-Distance Hyperspectral Near-Infrared Spectroscopy

A major limitation of near-infrared spectroscopy (NIRS) is its high sensitivity to the scalp and low sensitivity to the brain of adult humans. In the present work we used multi-distance hyperspectral NIRS (hNIRS) to investigate the optimal source-detector distances, wavelength ranges, and analysis techniques to separate cerebral responses to 30 s breath-holds (BHs) from the responses in the superficial tissue layer in healthy adult humans. We observed significant responses to BHs in the scalp hemodynamics. Cerebral responses to BHs were detected in the cytochrome C oxidase redox (rCCO) at 4 cm without using data from the short-distance channel. Using the data from the 1 cm channel in the two-layer regression algorithm showed that cerebral hemodynamic and rCCO responses also occurred at 3 cm. We found that the waveband 700–900 nm was optimal for the detection of cerebral responses to BHs in adults.

Suppressing physiological interferences and physical noises in functional diffuse optical tomography via tandem inversion filtering and LSTM classification

For performance enhancement of functional diffuse optical tomography (fDOT), we propose a tandem method that takes advantage of the inversion filtering and the long short term memory (LSTM) classification to simultaneously suppress the physiological interferences and physical noises in fDOT. In the former phase, the absorption perturbation maps over the scalp-skull (SS) and cerebral-cortex (CC) layers are firstly pre-reconstructed using a two-layer topography scheme. Then, the recovered SS-map is inversed into measurement space by the forward calculation to estimate the intensity changes associated with the physiological interferences. Finally, the raw measurements are adaptively filtered with reference to the estimated intensity changes for accomplishing the model-based full three-dimension (3D) reconstruction. In the later phase, for further removing the randomly distributed physical noises, mainly instrumental noise, a LSTM network based fusion strategy is applied, where a pixel-wise binary mask of "activated" or "inactive" state is generated by performing LSTM classification and then fused with the 3D reconstruction. The results of the simulative investigation and in-vivo experiment show the proposed tandem scheme achieves improved performance in fDOT using a cost-effective and physically explicable way. Thus, the proposed method can be promisingly applied in real-time neuroimaging to acquire cortical neural activation information with improved reliability, quantification and resolution.

Independent components analysis (ICA) at the "cocktail-party" in analytical chemistry

Independent components analysis (ICA) is a probabilistic method, whose goal is to extract underlying component signals, that are maximally independent and non-Gaussian, from mixed observed signals. Since the data acquired in many applications in analytical chemistry are mixtures of component signals, such a method is of great interest. In this article recent ICA applications for quantitative and qualitative analysis in analytical chemistry are reviewed. The following experimental techniques are covered: fluorescence, UV-VIS, NMR, vibrational spectroscopies as well as chromatographic profiles. Furthermore, we reviewed ICA as a preprocessing tool as well as existing hybrid ICA-based multivariate approaches. Finally, further research directions are proposed. Our review shows that ICA is starting to play an important role in analytical chemistry, and this will definitely increase in the future.

Cancellation of Superficial Blood Flow in Brain Function Measurements Using Near-Infrared Spectroscopy

In brain function measurement by near-infrared spectroscopy, reduction of signal fluctuation caused by the superficial blood flow is necessary. We proposed an equilateral-triangle-type sensor arrangement for cancelling superficial blood flow using short-distance detectors placed at the center of the triangle. We applied the independent components analysis (ICA) for this equilateral triangle configuration to improve the correction effect. We measured the change in blood volume caused by the motion artifacts according to the posture change and analyzed the correction effect of the ICA. It is shown that the ICA is more effective than simple subtraction and has small dependence on the measurement conditions and is applicable to this equilateral triangle configuration.

Hyperspectral near-infrared spectroscopy assessment of the brain during hypoperfusion

Two-thirds of out-of-hospital cardiac arrest patients, who survive to hospital admission, die in the hospital from neurological injuries related to cerebral hypoperfusion. Therefore, noninvasive real-time monitoring of the cerebral oxygen metabolism in cardiac arrest patients is extremely important. Hyperspectral near-infrared spectroscopy (hNIRS) is a noninvasive technique that measures concentrations of the key chromophores in the brain, such as oxygenated hemoglobin, deoxygenated hemoglobin, and cytochrome C oxidase (CCO), an intracellular marker of oxygen consumption. We tested hNIRS on 10 patients undergoing transcatheter aortic valve insertion, where rapid ventricular pacing (RVP) is required to temporarily induce sudden hypotension and hypoperfusion that mimic cardiac arrest. Using multidistance hNIRS, we found that tissue oxygen saturation changes in the cerebral tissue were lower than those in the scalp during RVP. CCO redox changes were detected in cerebral tissue but not in the scalp during RVP. We have demonstrated that hNIRS is feasible and can detect sudden changes in cerebral oxygenation and metabolism in patients during profound hypotension.

Adaptive algorithm utilizing acceptance rate for eliminating noisy epochs in block-design functional near-infrared spectroscopy data: application to study in attention deficit/hyperactivity disorder children

Functional near-infrared spectroscopy (fNIRS) signals are prone to problems caused by motion artifacts and physiological noises. These noises unfortunately reduce the fNIRS sensitivity in detecting the evoked brain activation while increasing the risk of statistical error. In fNIRS measurements, the repetitive resting-stimulus cycle (so-called block-design analysis) is commonly adapted to increase the sample number. However, these blocks are often affected by noises. Therefore, we developed an adaptive algorithm to identify, reject, and select the noise-free and/or least noisy blocks in accordance with the preset acceptance rate. The main features of this algorithm are personalized evaluation for individual data and controlled rejection to maintain the sample number. Three typical noise criteria (sudden amplitude change, shifted baseline, and minimum intertrial correlation) were adopted. Depending on the quality of the dataset used, the algorithm may require some or all noise criteria with distinct parameters. Aiming for real applications in a pediatric study, we applied this algorithm to fNIRS datasets obtained from attention deficit/hyperactivity disorder (ADHD) children as had been studied previously. These datasets were divided for training and validation purposes. A validation process was done to examine the feasibility of the algorithm regardless of the types of datasets, including those obtained under sample population (ADHD or typical developing children), intervention (nonmedication and drug/placebo administration), and measurement (task paradigm) conditions. The algorithm was optimized so as to enhance reproducibility of previous inferences. The optimum algorithm design involved all criteria ordered sequentially (0.047 mM mm of amplitude change, 0.029    mM    mm / s of baseline slope, and 0.6 × interquartile range of outlier threshold for each criterion, respectively) and presented complete reproducibility in both training and validation datasets. Compared to the visual-based rejection as done in the previous studies, the algorithm achieved 71.8% rejection accuracy. This suggests that the algorithm has robustness and potential to substitute for visual artifact-detection.

Multiwavelength time-resolved near-infrared spectroscopy of the adult head: assessment of intracerebral and extracerebral absorption changes

An optical technique based on diffuse reflectance measurement combined with indocyanine green (ICG) bolus tracking is extensively tested as a method for the clinical assessment of brain perfusion at the bedside. We report on multiwavelength time-resolved diffuse reflectance spectroscopy measurements carried out on the head of a healthy adult during the intravenous administration of a bolus of ICG. Intracerebral and extracerebral changes in absorption were estimated from an analysis of changes in statistical moments (total number of photons, mean time of flight and variance) of the distributions of times of flight (DTOF) of photons recorded simultaneously at 16 wavelengths from the range of 650-850 nm using sensitivity factors estimated by diffusion approximation based on a layered model of the studied medium. We validated the proposed method in a series of phantom experiments and in-vivo measurements. The results obtained show that changes in the concentration of the ICG can be assessed as a function of time of the experiment and depth in the tissue. Thus, the separation of changes in ICG concentration appearing in intra- and extracerebral tissues can be estimated from optical data acquired at a single source-detector pair of fibers/fiber bundles positioned on the surface of the head.

Event-related changes of the prefrontal cortex oxygen delivery and metabolism during driving measured by hyperspectral fNIRS

Recent technological advancements in optical spectroscopy allow for the construction of hyperspectral (broadband) portable tissue oximeters. In a series of our recent papers we have shown that hyperspectral NIRS (hNIRS) has similar or better capabilities in the absolute tissue oximetry as frequency-domain NIRS, and that hNIRS is also very efficient in measuring temporal changes in tissue hemoglobin concentration and oxygenation. In this paper, we extend the application of hNIRS to the measurement of event-related hemodynamic and metabolic functional cerebral responses during simulated driving. In order to check if hNIRS can detect event-related changes in the brain, we measured the concentration changes of oxygenated (HbO2) and deoxygenated (HHb) hemoglobin and of the oxidized state of cytochrome c oxidase, on the right and left prefrontal cortices (PFC) simultaneously during simulated driving on sixteen healthy right-handed participants (aged between 22-32). We used our in-house hNIRS system based on a portable spectrometer with cooled CCD detector and a driving simulator with a fully functional steering wheel and foot pedals. Each participant performed different driving tasks and participants were distracted during some driving conditions by asking general knowledge true/false questions. Our findings suggest that more complex driving tasks (non-distracted) deactivate PFC while distractions during driving significantly activate PFC, which is in agreement with previous fMRI results. Also, we found the changes in the redox state of the cytochrome C oxidase to be very consistent with those in the concentrations of HbO2 and HHb. Overall our findings suggest that in addition to the suitability of absolute tissue oximetry, hyperspectral NIRS may also offer advantages in functional brain imaging. In particular, it can be used to measure the metabolic functional brain activity during actual driving.

Effect of mechanical optical clearing on near-infrared spectroscopy

Near-infrared Spectroscopy (NIRS) is a broadly utilized technology with many emerging applications including clinical diagnostics, sports medicine, and functional neuroimaging, to name a few. For functional brain imaging NIR light is delivered at multiple wavelengths through the scalp and skull to the brain to enable spatial oximetry measurements. Dynamic changes in brain oxygenation are highly correlated with neural stimulation, activation, and function. Unfortunately, NIRS is currently limited by its low spatial resolution, shallow penetration depth, and, perhaps most importantly, signal corruption due to light interactions with superficial non-target tissues such as scalp and skull. In response to these issues, we have combined the non-invasive and rapidly reversible method of mechanical tissue optical clearing (MOC) with a commercially available NIRS system. MOC utilizes a compressive loading force on tissue, causing the lateral displacement of blood and water, while simultaneously thinning the tissue. A MOC-NIRS Breath Hold Test displayed a ∼3.5-fold decrease in the time-averaged standard deviation between channels, consequentially promoting greater channel agreement. A Skin Pinch Test was implemented to negate brain and muscle activity from affecting the recorded signal. These results displayed a 2.5-3.0 fold increase in raw signal amplitude. Existing NIRS instrumentation has been further integrated within a custom helmet device to provide a uniform force distribution across the NIRS sensor array. These results showed a gradual decrease in time-averaged standard deviation among channels with an increase in applied pressure. Through these experiments, and the development of the MOC-NIRS helmet device, MOC appears to provide enhancement of NIRS technology beyond its current limitations.

Independent components analysis to increase efficiency of discriminant analysis methods (FDA and LDA): Application to NMR fingerprinting of wine

Discriminant analysis (DA) methods, such as linear discriminant analysis (LDA) or factorial discriminant analysis (FDA), are well-known chemometric approaches for solving classification problems in chemistry. In most applications, principle components analysis (PCA) is used as the first step to generate orthogonal eigenvectors and the corresponding sample scores are utilized to generate discriminant features for the discrimination. Independent components analysis (ICA) based on the minimization of mutual information can be used as an alternative to PCA as a preprocessing tool for LDA and FDA classification. To illustrate the performance of this ICA/DA methodology, four representative nuclear magnetic resonance (NMR) data sets of wine samples were used. The classification was performed regarding grape variety, year of vintage and geographical origin. The average increase for ICA/DA in comparison with PCA/DA in the percentage of correct classification varied between 6±1% and 8±2%. The maximum increase in classification efficiency of 11±2% was observed for discrimination of the year of vintage (ICA/FDA) and geographical origin (ICA/LDA). The procedure to determine the number of extracted features (PCs, ICs) for the optimum DA models was discussed. The use of independent components (ICs) instead of principle components (PCs) resulted in improved classification performance of DA methods. The ICA/LDA method is preferable to ICA/FDA for recognition tasks based on NMR spectroscopic measurements.

Reduction of physiological effects in fNIRS waveforms for efficient brain-state decoding

This paper presents a methodology for online estimation of brain activities with reduction in the effects of physiological noises in functional near-infrared spectroscopy signals. The input-output characteristics of a hemodynamic response are modeled as an autoregressive moving average model together with exogenous physical signals (i.e., ARMAX). In contrast to the fixed design matrix in the conventional general linear model, the proposed model incorporates the temporal variations in the experimental paradigm as well as in the hemodynamics. The performance of the proposed method has been tested by using box-car type functions followed by individual tapping tasks. The results and their significance were verified using t-statistics indicating that ARMAX seems to be better able to track/reveal the hemodynamic response. Also, online brain-activation maps were generated for localizing brain activities. Experimental results are compared with those of the existing conventional GLM-based method.

Near-infrared spectroscopy (NIRS) as a new tool for neuroeconomic research

Over the last decade, the application of neuroscience to economic research has gained in importance and the number of neuroeconomic studies has grown extensively. The most common method for these investigations is fMRI. However, fMRI has limitations (particularly concerning situational factors) that should be countered with other methods. This review elaborates on the use of functional Near-Infrared Spectroscopy (fNIRS) as a new and promising tool for investigating economic decision making both in field experiments and outside the laboratory. We describe results of studies investigating the reliability of prototype NIRS studies, as well as detailing experiments using conventional and stationary fNIRS devices to analyze this potential. This review article shows that further research using mobile fNIRS for studies on economic decision making outside the laboratory could be a fruitful avenue helping to develop the potential of a new method for field experiments outside the laboratory.

State-space models of impulse hemodynamic responses over motor, somatosensory, and visual cortices

THE PAPER PRESENTS STATE SPACE MODELS OF THE HEMODYNAMIC RESPONSE (HR) OF FNIRS TO AN IMPULSE STIMULUS IN THREE BRAIN REGIONS: motor cortex (MC), somatosensory cortex (SC), and visual cortex (VC). Nineteen healthy subjects were examined. For each cortex, three impulse HRs experimentally obtained were averaged. The averaged signal was converted to a state space equation by using the subspace method. The activation peak and the undershoot peak of the oxy-hemoglobin (HbO) in MC are noticeably higher than those in SC and VC. The time-to-peaks of the HbO in three brain regions are almost the same (about 6.76 76 ± 0.2 s). The time to undershoot peak in VC is the largest among three. The HbO decreases in the early stage (~0.46 s) in MC and VC, but it is not so in SC. These findings were well described with the developed state space equations. Another advantage of the proposed method is its easy applicability in generating the expected HR to arbitrary stimuli in an online (or real-time) imaging. Experimental results are demonstrated.

Independent component analysis (ICA) algorithms for improved spectral deconvolution of overlapped signals in 1H NMR analysis: application to foods and related products

The major challenge facing NMR spectroscopic mixture analysis is the overlapping of signals and the arising impossibility to easily recover the structures for identification of the individual components and to integrate separated signals for quantification. In this paper, various independent component analysis (ICA) algorithms [mutual information least dependent component analysis (MILCA); stochastic non-negative ICA (SNICA); joint approximate diagonalization of eigenmatrices (JADE); and robust, accurate, direct ICA algorithm (RADICAL)] as well as deconvolution methods [simple-to-use-interactive self-modeling mixture analysis (SIMPLISMA) and multivariate curve resolution-alternating least squares (MCR-ALS)] are applied for simultaneous (1)H NMR spectroscopic determination of organic substances in complex mixtures. Among others, we studied constituents of the following matrices: honey, soft drinks, and liquids used in electronic cigarettes. Good quality spectral resolution of up to eight-component mixtures was achieved (correlation coefficients between resolved and experimental spectra were not less than 0.90). In general, the relative errors in the recovered concentrations were below 12%. SIMPLISMA and MILCA algorithms were found to be preferable for NMR spectra deconvolution and showed similar performance. The proposed method was used for analysis of authentic samples. The resolved ICA concentrations match well with the results of reference gas chromatography-mass spectrometry as well as the MCR-ALS algorithm used for comparison. ICA deconvolution considerably improves the application range of direct NMR spectroscopy for analysis of complex mixtures.

Noise reduction in functional near-infrared spectroscopy signals by independent component analysis

Functional near-infrared spectroscopy (fNIRS) is used to detect concentration changes of oxy-hemoglobin and deoxy-hemoglobin in the human brain. The main difficulty entailed in the analysis of fNIRS signals is the fact that the hemodynamic response to a specific neuronal activation is contaminated by physiological and instrument noises, motion artifacts, and other interferences. This paper proposes independent component analysis (ICA) as a means of identifying the original hemodynamic response in the presence of noises. The original hemodynamic response was reconstructed using the primary independent component (IC) and other, less-weighting-coefficient ICs. In order to generate experimental brain stimuli, arithmetic tasks were administered to eight volunteer subjects. The t-value of the reconstructed hemodynamic response was improved by using the ICs found in the measured data. The best t-value out of 16 low-pass-filtered signals was 37, and that of the reconstructed one was 51. Also, the average t-value of the eight subjects' reconstructed signals was 40, whereas that of all of their low-pass-filtered signals was only 20. Overall, the results showed the applicability of the ICA-based method to noise-contamination reduction in brain mapping.

Direct identification of breast cancer pathologies using blind separation of label-free localized reflectance measurements

Breast tumors are blindly identified using Principal (PCA) and Independent Component Analysis (ICA) of localized reflectance measurements. No assumption of a particular theoretical model for the reflectance needs to be made, while the resulting features are proven to have discriminative power of breast pathologies. Normal, benign and malignant breast tissue types in lumpectomy specimens were imaged ex vivo and a surgeon-guided calibration of the system is proposed to overcome the limitations of the blind analysis. A simple, fast and linear classifier has been proposed where no training information is required for the diagnosis. A set of 29 breast tissue specimens have been diagnosed with a sensitivity of 96% and specificity of 95% when discriminating benign from malignant pathologies. The proposed hybrid combination PCA-ICA enhanced diagnostic discrimination, providing tumor probability maps, and intermediate PCA parameters reflected tissue optical properties.

Combining independent component analysis and Granger causality to investigate brain network dynamics with fNIRS measurements

In this study a new strategy that combines Granger causality mapping (GCM) and independent component analysis (ICA) is proposed to reveal complex neural network dynamics underlying cognitive processes using functional near infrared spectroscopy (fNIRS) measurements. The GCM-ICA algorithm implements the following two procedures: (i) extraction of the region of interests (ROIs) of cortical activations by ICA, and (ii) estimation of the direct causal influences in local brain networks using Granger causality among voxels of ROIs. Our results show that the use of GCM in conjunction with ICA is able to effectively identify the directional brain network dynamics in time-frequency domain based on fNIRS recordings.