MRI Data-Driven Functional Near-Infrared Spectroscopy Mapping Method for Conforming Functional Hemodynamic Neural Activity to Structural Cortical Information

摘要

Functional near-infrared spectroscopy (fNIRS) as a brain imaging technique is getting more and more interested in spite of its lower spatial resolution compared to functional magnetic resonance image (fMRI). Especially, fNIRS is gradually spread out in investigating the cortical activity of motion related tasks, because of its advantages of robustness to motion artifacts, patient acceptability, and the experimental simplicity. Since fNIRS uses internationally recognized 10-20 probe positing system, the accuracy of source localization is poor. In addition, fNIRS does not provide accurate structural information for different individual brains because its statistical parametric mapping (SPM) is a mapping method that projects the statistical change of oxygenated or deoxygenated hemoglobin in a standard brain template. This paper proposed fNIRS mapping procedure that can reflect the more accurate structural characteristics of individual brain. First, the reliability of the fNIRS localization method was verified by comparing 3D digitizer and neuro-navigator position errors. Second, the personalized fNIRS images was acquired by converting the structural information obtained from the individual MRI to the suitable form for the SPM. Two healthy subjects participated in the investigation of localization errors with carrying out two motor tasks comprising hand grasping and knee bending. From a total of 2 trials, the average of error was 4.16 mm. These results guaranteed the reliability of the fNIRS source localization method using 3D-digitizer because the error bound was within 30 mm of the measurement distance between the probes. In order to reconstruct the personalized fNIRS brain images, five stroke patients were enrolled in the body weighted support treadmill training (BWSTT) experiment. As a result of the comparison between the conventional NIRS-SPM image and the proposed personalized brain image, the proposed method clearly discovered the position of individual brain lesion and more clearly figured out the lateralization and intensity of individual brain activity. This suggests that the proposed personalized brain imaging technique is useful for observing not only the structural characteristics of brain (lateralization and localization of lesion or cortical activity) but also the functional properties (intensity of oxy-or deoxy-hemoglobin) of the brain.