A novel least-squares regression algorithm was developed to compensate for cardiac pulse artifacts in near-infrared optical imaging of brain activity.
Various factors, including variations in the concentration of hemoglobin, determine changes in the transparency of living tissue to near-infrared light. Hence, optical measures have been proposed as a noninvasive method for investigating regional changes in brain activity. However, the amount of near-infrared light traversing a region of the head is also influenced by the periodic changes in blood pressure that occur during the cardiac cycle (pulse). These large changes may obscure smaller, localized events associated with brain activity. We developed a least-squares regression algorithm for compensating for the artifact introduced by the pulse. This procedure takes into account beat-to-beat variability in heart rate and differences in the shape of the pulse among subjects and among recording conditions.
Gratton et al. (Mon,) studied this question.