Detection of the fast neuronal signal on the motor cortex using functional frequency domain near infrared spectroscopy

Using non-invasive near infrared spectroscopy fast changes in the range of ms in the optical properties of neurons during brain activity have been described. Since the signal is small, the system to detect it has to be highly noise optimized. We used a frequency-domain tissue oximeter, whose laser diodes were modulated at 110MHz and the amplitude (AC), mean intensity (DC) and phase (phi) of the modulated optical signal was measured at 96 Hz sample rate. In two volunteers, 36 and 37 years old, the probe consisting of 4 crossed source detector pairs was placed above the motor cortex (C3 position), contralateral to the hand performing the tapping exercise. The tapping frequency was set at 2.5 times the heart rate of the subject to avoid the influence of harmonics on the signal. An electronic device recorded the tapping movement. Control-data were obtained from a solid medium of approximately the same optical properties as the human head. To reduce physiological noise the arterial pulsatility was removed using an adaptive filter, the data was detrended by a high pass filter and a cross correlation function between the optical data and the tapping signal was calculated. The instrumental noise of the control data was very low (AC mean 0.0015% +/- SD 0.00092%, DC 0.00037% +/- 0.00023% and phi 0.00083degrees +/- 0.00042degrees). On the head the noise level was AC 0.0042% +/- 0.0031%, DC 0.0021% +/- 0.0012% and phi 0.0020degrees +/- 0.0017degrees. In 14 DC, 5 AC and 0 phi out of 30 locations a fast signal was detected, which was higher (p<0.001) than the noise level. This signal disappeared during non-tapping periods. With the signal to noise ratio that we have achieved single subject measurements become feasible.