Akihiro Ishikawa

Temporal classification of multichannel near-infrared spectroscopy signals of motor imagery for developing a brain–computer interface

There has been an increase in research interest for brain-computer interface (BCI) technology as an alternate mode of communication and environmental control for the disabled, such as patients suffering from amyotrophic lateral sclerosis (ALS), brainstem stroke and spinal cord injury. Disabled patients with appropriate physical care and cognitive ability to communicate with their social environment continue to live with a reasonable quality of life over extended periods of time. Near-infrared spectroscopy is a non-invasive technique which utilizes light in the near-infrared range (700 to 1000 nm) to determine cerebral oxygenation, blood flow and metabolic status of localized regions of the brain. In this paper, we describe a study conducted to test the feasibility of using multichannel NIRS in the development of a BCI. We used a continuous wave 20-channel NIRS system over the motor cortex of 5 healthy volunteers to measure oxygenated and deoxygenated hemoglobin changes during left-hand and right-hand motor imagery. We present results of signal analysis indicating that there exist distinct patterns of hemodynamic responses which could be utilized in a pattern classifier towards developing a BCI. We applied two different pattern recognition algorithms separately, Support Vector Machines (SVM) and Hidden Markov Model (HMM), to classify the data offline. SVM classified left-hand imagery from right-hand imagery with an average accuracy of 73% for all volunteers, while HMM performed better with an average accuracy of 89%. Our results indicate potential application of NIRS in the development of BCIs. We also discuss here future extension of our system to develop a word speller application based on a cursor control paradigm incorporating online pattern classification of single-trial NIRS data.

Removal of the skin blood flow artifact in functional near-infrared spectroscopic imaging data through independent component analysis.

We investigate whether the functional near-infrared spectroscopic (fNIRS) signal includes a signal from the changing skin blood flow. During a locomotor task on a treadmill, changes in the hemodynamic response in the front-parietal area of healthy human subjects are simultaneously recorded using an fNIRS imaging system and a laser Doppler tissue blood flow meter. Independent component analysis (ICA) for fNIRS signals is performed. The skin blood flow changes during locomotor tasks on a treadmill. The activated spatial distribution of one of the components separated by ICA reveals an overall increase in fNIRS channels. To evaluate the uniformity of the activated spatial distribution, we define a new statistical value-the coefficient of spatial uniformity (CSU). The CSU value is a highly discriminating value (e.g., 2.82) compared with values of other components (e.g., 1.41, 1.10, 0.96, 0.61, and 0.58). In addition, the independent component signal corresponding to the activated spatial distribution is similar to changes in skin blood flow measured with the laser Doppler tissue blood flow meter. The coefficient of correlation indicates strong correlation. Localized activation areas around the premotor and medial somatosensory cortices are shown more clearly by eliminating the extracted component.

Neurofeedback Using Real-Time Near-Infrared Spectroscopy Enhances Motor Imagery Related Cortical Activation

Accumulating evidence indicates that motor imagery and motor execution share common neural networks. Accordingly, mental practices in the form of motor imagery have been implemented in rehabilitation regimes of stroke patients with favorable results. Because direct monitoring of motor imagery is difficult, feedback of cortical activities related to motor imagery (neurofeedback) could help to enhance efficacy of mental practice with motor imagery. To determine the feasibility and efficacy of a real-time neurofeedback system mediated by near-infrared spectroscopy (NIRS), two separate experiments were performed. Experiment 1 was used in five subjects to evaluate whether real-time cortical oxygenated hemoglobin signal feedback during a motor execution task correlated with reference hemoglobin signals computed off-line. Results demonstrated that the NIRS-mediated neurofeedback system reliably detected oxygenated hemoglobin signal changes in real-time. In Experiment 2, 21 subjects performed motor imagery of finger movements with feedback from relevant cortical signals and irrelevant sham signals. Real neurofeedback induced significantly greater activation of the contralateral premotor cortex and greater self-assessment scores for kinesthetic motor imagery compared with sham feedback. These findings suggested the feasibility and potential effectiveness of a NIRS-mediated real-time neurofeedback system on performance of kinesthetic motor imagery. However, these results warrant further clinical trials to determine whether this system could enhance the effects of mental practice in stroke patients.

Quantification of Cerebral Blood Flow and Oxygen Metabolism with 3-Dimensional PET and 15O: Validation by Comparison with 2-Dimensional PET

Quantitative PET with 15O provides absolute values for cerebral blood flow (CBF), cerebral blood volume (CBV), cerebral metabolic rate of oxygen (CMRO2), and oxygen extraction fraction (OEF), which are used for assessment of brain pathophysiology. Absolute quantification relies on physically accurate measurement, which, thus far, has been achieved by 2-dimensional PET (2D PET), the current gold standard for measurement of CBF and oxygen metabolism. We investigated whether quantitative 15O study with 3-dimensional PET (3D PET) shows the same degree of accuracy as 2D PET. Methods: 2D PET and 3D PET measurements were obtained on the same day on 8 healthy men (age, 21–24 y). 2D PET was performed using a PET scanner with bismuth germanate (BGO) detectors and a 150-mm axial field of view (FOV). For 3D PET, a 3D-only tomograph with gadolinium oxyorthosilicate (GSO) detectors and a 156-mm axial FOV was used. A hybrid scatter-correction method based on acquisition in the dual-energy window (hybrid dual-energy window [HDE] method) was applied in the 3D PET study. Each PET study included 3 sequential PET scans for C15O, 15O2, and \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{H}_{2}^{15}\mathrm{O}\) \end{document} (3-step method). The inhaled (or injected) dose for 3D PET was approximately one fourth of that for 2D PET. Results: In the 2D PET study, average gray matter values (mean ± SD) of CBF, CBV, CMRO2, and OEF were 53 ± 12 (mL/100 mL/min), 3.6 ± 0.3 (mL/100 mL), 3.5 ± 0.5 (mL/100 mL/min), and 0.35 ± 0.06, respectively. In the 3D PET study, scatter correction strongly affected the results. Without scatter correction, average values were 44 ± 6 (mL/100 mL/min), 5.2 ± 0.6 (mL/100 mL), 3.3 ± 0.4 (mL/100 mL/min), and 0.39 ± 0.05, respectively. With the exception of OEF, values differed between 2D PET and 3D PET. However, average gray matter values of scatter-corrected 3D PET were comparable to those of 2D PET: 55 ± 11 (mL/100 mL/min), 3.7 ± 0.5 (mL/100 mL), 3.8 ± 0.7 (mL/100 mL/min), and 0.36 ± 0.06, respectively. Even though the 2 PET scanners with different crystal materials, data acquisition systems, spatial resolution, and attenuation-correction methods were used, the agreement of the results between 2D PET and scatter-corrected 3D PET was excellent. Conclusion: Scatter coincidence is a problem in 3D PET for quantitative 15O study. The combination of both the present PET/CT device and the HDE scatter correction permits quantitative 3D PET with the same degree of accuracy as 2D PET and with a lower radiation dose. The present scanner is also applicable to conventional steady-state 15O gas inhalation if inhaled doses are adjusted appropriately.

Development of a new rehabilitation system based on a brain-computer interface using near-infrared spectroscopy.

We describe the set-up for an electrical muscle stimulation device based on near-infrared spectroscopy (NIRS), designed for use as a brain-computer interface (BCI). Employing multi-channel NIRS, we measured evoked cerebral blood oxygenation (CBO) responses during real motor tasks and motor-imagery tasks. When a supra-threshold increase in oxyhemoglobin concentration was detected, electrical stimulation (50 Hz) of the biceps brachii muscle was applied to the side contralateral to the hand grasping task or ipsilateral to the motor-imagery task. We observed relatively stable and reproducible CBO responses during real motor tasks with an average accuracy of 100%, and during motor imagery tasks with an average accuracy of 61.5%. Flexion movement of the arm was evoked in all volunteers in association with electrical muscle stimulation and no adverse effects were noted. These findings suggest that application of the electrical muscle stimulation system based on a NIRS-BCI is non-invasive and safe, and may be useful for the physical training of disabled patients.

Cerebral functional imaging using near-infrared spectroscopy during repeated performances of motor rehabilitation tasks tested on healthy subjects

To investigate the relationship between the frontal and sensorimotor cortices and motor learning, hemodynamic responses were recorded from the frontal and sensorimotor cortices using functional near infrared spectroscopy (NIRS) while healthy subjects performed motor learning tasks used in rehabilitation medicine. Whole-head NIRS recordings indicated that response latencies in the anterior dorsomedial prefrontal cortex (aDMPFC) were shorter than in other frontal and parietal areas. Furthermore, the increment rate of the hemodynamic responses in the aDMPFC across the eight repeated trials significantly correlated with those in the other areas, as well as with the improvement rate of task performance across the 8 repeated trials. In the second experiment, to dissociate scalp- and brain-derived hemodynamic responses, hemodynamic responses were recorded from the head over the aDMPFC using a multi-distance probe arrangement. Six probes (a single source probe and 5 detectors) were linearly placed 6 mm apart from each of the neighboring probes. Using independent component analyses of hemodynamic signals from the 5 source-detector pairs, we dissociated scalp- and brain-derived components of the hemodynamic responses. Hemodynamic responses corrected for scalp-derived responses over the aDMPFC significantly increased across the 8 trials and correlated with task performance. In the third experiment, subjects were required to perform the same task with and without transcranial direct current stimulation (tDCS) of the aDMPFC before the task. The tDCS significantly improved task performance. These results indicate that the aDMPFC is crucial for improved performance in repetitive motor learning.

Fast (100–175 ms) components elicited bilaterally by language production as measured by three-wavelength optical imaging

Optical imaging has been gradually utilized to investigate language functions in the brain. The majority of hemodynamic response (slow signal) measurements have been applied to receptive and productive language processing, while several event-related optical signal (EROS) measurements on neuronal response (fast signal) have focused on receptive language processing. Therefore, an investigation of language production based on fast signal measurement is yet to be realized. Using a continuous-wave near-infrared spectroscopic (CW-NIRS) method with three long wavelengths in close ranges (780, 805, and 830 nm), which are suitable for the detection of fast optical signals, the current work investigated whether absorbance-based EROS components during overt language production might be elicited bilaterally in each wavelength with a 25 ms sampling time. Healthy adult subjects read aloud Japanese noun phrases (NP) presented on a computer screen. Two conditions (short/long-vowel duration) included either initial [s]- or [k]-phoneme types in the first words of the NP. The cognitive subtraction method achieved by deducting short-duration from long-duration conditions showed that in both phoneme types, reliable fast optical components with a peak latency of about 100-175 ms post initial-consonant onset were bilaterally elicited by long vowels. This result suggests that the present CW-NIRS methodology can clearly detect such early optical signals with good temporal resolution and with a good signal-to-noise ratio (SNR) obtained from a small number of stimuli. The fact that optical absorbance values at all three wavelengths had the same positive deflections during the initial-syllable production demonstrates that the elicitation of fast optical components may directly represent neuronal activity.

Self normalization for continuous 3D whole body emission data in 3D PET

Continuous 3D scanning with a large-aperture PET scanner can provide high sensitivity over most of the axial FOV in whole-body studies. To minimum the artifact depended on the uniformity of the different lines-of-response (LORs) in sinograms, accurate normalizing algorithms will be needed. In this study, we propose self-normalization method in which full correction factors are derived from the emission data itself without using conventional normalize scan of cylinder phantom. In this method, transaxial block profile and crystal efficiency were calculated from the original emission data, and correction factors applied itself. We implemented proposed method to a 5 ring GSO PET scanner which has continuous 3D scan mode and evaluated. To accurate correction factor, components were calculated from dataset which summed in the direction of the bed movement. To investigate the performance, we compared the proposed method with conventional component based normalization using uniform cylinder phantom. And To evaluate clinical performance, we also /sup 18/FDG human studies were performed. In both phantom and human studies, the block profile and crystal efficiencies can be calculated correctly using proposed method. Evaluating percent standard deviation, self-normalization improved transaxial uniformity since it reduced ring artifacts. Especially, the accuracy of transaxial block profiles which influenced easily by some physical phenomena has unproved. Our self-normalize method was accurate enough for continuous 3D scanning with a large aperture PET scanner. The image quality using self-normalization was superior than conventional component normalization. This method contributes to the improvement of system operation , since regular acquisition of cylinder phantom for normalization is not necessary.

Cerebral Hemodynamic Responses During Dynamic Posturography: Analysis with a Multichannel Near-Infrared Spectroscopy System

To investigate cortical roles in standing balance, cortical hemodynamic activity was recorded from the right hemisphere using near-infrared spectroscopy (NIRS) while subjects underwent the sensory organization test (SOT) protocol that systematically disrupts sensory integration processes (i.e., somatosensory or visual inputs or both). Eleven healthy men underwent the SOT during NIRS recording. Group statistical analyses were performed based on changes in oxygenated hemoglobin concentration in 10 different cortical regions of interest and on a general linear analysis with NIRS statistical parametric mapping. The statistical analyses indicated significant activation in the right frontal operculum (f-Op), right parietal operculum (p-Op), and right superior temporal gyrus (STG), right posterior parietal cortex (PPC), right dorsal and ventral premotor cortex (PMC), and the supplementary motor area (SMA) under various conditions. The activation patterns in response to specific combinations of SOT conditions suggested that (1) f-Op, p-Op, and STG are essential for sensory integration when standing balance is perturbed; (2) the SMA is involved in the execution of volitional action and establishment of new motor programs to maintain postural balance; and (3) the PPC and PMC are involved in the updating and computation of spatial reference frames during instances of sensory conflict between vestibular and visual information.

Development of double density whole brain fNIRS with EEG system for brain machine interface

Brain-machine interfaces (BMI) are expected as new man-machine interfaces. Non-invasive BMI have the potential to improve the quality of life of many disabled individuals with safer operation. The non-invasive BMI using the functional functional near-infrared spectroscopy (fNIRS) with the electroencephalogram (EEG) has potential applicability beyond the restoration of lost movement and rehabilitation in paraplegics and would enable normal individuals to have direct brain control of external devices in their daily lives. To shift stage of the non-invasive BMI from laboratory to clinical, the key factor is to develop high-accuracy signal decoding technology and highly restrictive of the measurement area. In this article, we present the development of a high-accuracy brain activity measurement system by combining fNIRS and EEG. The new fNIRS had high performances with high spatial resolution using double density technique and a large number of measurement channels to cover a whole human brain.