2.1.

Subjects

31 healthy adults (21 men, 10 women; mean age $34\pm 8.9$ , range 23 to 56) gave written informed consent before the experiments. All the subjects showed right-handedness except one, and none reported a history of neurological disorders.

2.2.

NIRS Measurement

An NIR topography system with 24 measurement positions (ETG-100, Hitachi Medical Corporation, Japan) was used. The system irradiates light at 780- and $830\text{-}\mathrm{nm}$ wavelength through an optical fiber to the same measurement point simultaneously. The reflected light was detected every $100\mathrm{ms}$ using an avalanche photodiode (APD) located $30\mathrm{mm}$ from the incident position.

We regarded the midpoint of the source-detector distance as the measurement position because the sensitivity of NIRS to chromophore-concentration changes is highest there.^{46, 47, 48} Optical fibers were used for both the irradiating and detecting the lights. The average power of each light source was $1.5\mathrm{mW}$ , and each source was modulated at a distinctive frequency $\left(1.0\text{to}10\mathrm{kHz}\right)$ to enable separation using a lock-in amplifier after detection. Ten irradiated positions and eight detection positions were configured to measure the 24 positions (Fig. 1 ).

Fig. 1

Arrangement of measurement positions in probe patterns over left and right sensorimotor areas centered on locations C3 and C4, respectively.

The measurement area was determined for each subject based on the international 10–20 system.⁴⁹ We measured an area of $6\times 6\mathrm{cm}$ in the left and right parietal areas centered on ${\mathrm{C}}_3$ and ${\mathrm{C}}_4$ , respectively (Fig. 1). The $6\times 6\text{-}\mathrm{cm}$ square was defined as the measurement area for each hemisphere based on the arrangement of optical fibers (irradiation and detection positions). The probe patterns were positioned parallel to the line connecting Cz to T3 (left hemisphere) or T4 (right hemisphere). The centers of the bilateral measurement areas were considered to correspond to each primary sensorimotor area based on previous studies examining the relationship between the international 10–20 locations and cortical area.^{50, 51, 52} In particular, Okamoto et al. recently reported that the ${\mathrm{C}}_3$ and ${\mathrm{C}}_4$ locations correspond to the central fissure with a standard deviation of less than $10\mathrm{mm}$ among 17 Mongoloids,⁵⁰ the same race as our subjects.

2.3.

Task Paradigm

The sensorimotor cortices were activated using a finger-tapping task. The fingers of one hand were placed on the tip of the thumb in serial order (forefinger, second finger, third finger, little finger, third finger, second finger, forefinger). The subjects were asked to repeat the tapping sequence at $3\mathrm{Hz}$ timed to the rate of the term “finger tapping” blinking on a CRT monitor for $30\mathrm{s}$ (activation) followed by $30\mathrm{s}$ of rest (baseline). There were a total of five activation/baseline cycles per session. Two sessions of finger tapping (one left-hand tapping and one right-hand tapping) were conducted per subject in a counterbalanced order among subjects. During the measurements, the subject sat on a chair and was instructed to fix his or her gaze on the fixation point at the center of the screen and to concentrate on the task.

2.4.

Data Analysis

We defined a period of $55\mathrm{s}$ , consisting of a $5\text{-}\mathrm{s}$ prestimulation resting period, a $30\text{-}\mathrm{s}$ stimulation period in the finger-tapping state, and a $20\text{-}\mathrm{s}$ poststimulation resting period as one (task) block. The detected temporal data for the attenuation change at each wavelength were separated into five blocks. Each block was baseline corrected using the data for the pre- and poststimulation periods; the products of the effective optical path length and the concentration changes of the independent hemoglobin species ( $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\mathrm{deoxy}}^{\prime }$ ) were calculated by applying the modified Beer-Lambert Law¹¹ as follows:

The activation signals were statistically assessed. Assuming that the hemodynamic time courses induced by the task varied among the subjects, we defined a $25\text{-}\mathrm{s}$ activation period for each subject. We used a within-subject averaged time course over the five blocks to select the one $25\text{-}\mathrm{s}$ activation period with maximum absolute value of the mean change for each hemoglobin species. The $25\text{-}\mathrm{s}$ activation period was allowed to shift from the initial $5\mathrm{s}$ after task onset to starting $15\mathrm{s}$ after task onset; the earliest period was from $5\mathrm{s}$ after task onset to task completion, and the latest period was from $15\mathrm{s}$ after task onset to $10\mathrm{s}$ after task completion (Fig. 2 ).

Fig. 2

Schematic diagram of analysis parameters for a task block. Mean value of $5\mathrm{s}$ for prestimulation (R) and mean value of $25\mathrm{s}$ for activation (A) were used for statistical analysis. Note that A could shift from $5\mathrm{s}$ after task onset to $10\mathrm{s}$ after task completion, depending on the maximum absolute value for each case.

The mean changes in hemoglobin during the prestimulation period and those during the activation period were calculated for each block. We calculated the $t$ value (paired $t$ test) between the mean hemoglobin changes in the prestimulation periods of five blocks and those in the activation periods of the same five blocks, and identified measurement positions with significant $t$ values (two-tailed $t$ test, $p<0.1$ ) as activation points.^{14, 36} We determined the threshold for each NIRS signal ( $\Delta C_{\mathrm{oxy}}^{\prime }$ , $\Delta C_{\mathrm{deoxy}}^{\prime }$ , and $\Delta C_{\text{total}}^{\prime }$ ). This statistical analysis was designed to determine the consistency (reproducibility) of changes for the five activation periods. With this analysis, causeless changes are not detected as activation because the statistical value does not reach the threshold unless similar changes arise in every (or almost every) activation period. In addition, using activation periods $25\mathrm{s}$ long reduces the possibility of misidentifying an increase or decrease due to spontaneous oscillations during a cycle of around 10 as activation.⁵³ Moreover, a $t$ test using the variance of the mean value in the resting period and that in the activation period across task blocks reduces the effect of high-frequency system noise. Note that although the determination of the activation period based on the peak timing differs from the analytical method we used in our previous report,⁴⁵ these differences in the results were negligible.

3.1.

Occurrence Probability of Typical Activation Pattern

We assessed the occurrence probability of the typical activation pattern (positive $\Delta C_{\mathrm{oxy}}^{\prime }$ , negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ , and positive $\Delta C_{\text{total}}^{\prime }$ ) for each measurement-position (Fig. 3 ). Positive $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\text{total}}^{\prime }$ in the hemisphere contralateral to the tapping hand were observed with high probability. In addition, the measurement positions with high probabilities were around the center of the measurement area (left hemisphere: positions 6 and 9; right hemisphere: position 16). Contrary to that, negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ as a whole showed less probability, although the activation centers were almost the same as those for $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\text{total}}^{\prime }$ (left hemisphere: position 9; right hemisphere: position 16). The mean signal amplitudes for subjects who showed the typical activation pattern are given in Table 1 . The mean amplitudes for $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\text{total}}^{\prime }$ were 0.109 and $0.104\mathrm{mM}\mathbf{\bullet }\mathrm{mm}$ , respectively, while that for $\Delta C_{\mathrm{deoxy}}^{\prime }$ was about half $(-0.048\mathrm{mM}\bullet \mathrm{mm})$ . The amplitudes sometimes varied widely among subjects, from small [Fig. 4(a) ] to large [Fig. 4(b)].

Fig. 4

Representative time courses of $\Delta C_{\mathrm{oxy}}^{\prime }$ , $\Delta C_{\mathrm{deoxy}}^{\prime }$ , and $\Delta C_{\text{total}}^{\prime }$ . Typical pattern (positive $\Delta C_{\mathrm{oxy}}^{\prime }$ , negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ , and positive $\Delta C_{\text{total}}^{\prime }$ ) is shown in (a) and (b); changes in (b) are larger than those in (a). An all-positive pattern (positive $\Delta C_{\mathrm{oxy}}^{\prime }$ , $\Delta C_{\mathrm{deoxy}}^{\prime }$ , and $\Delta C_{\text{total}}^{\prime }$ ) is shown in (c) and (d). Pattern in (d) was a rare case— $\Delta C_{\mathrm{deoxy}}^{\prime }$ was positive similar to $\Delta C_{\mathrm{oxy}}^{\prime }$ . Two exceptional patterns are shown in (e) and (f). The pattern with only a negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ , (e), was observed in only two subjects. The pattern with a negative $\Delta C_{\mathrm{oxy}}^{\prime }$ and a positive $\Delta C_{\mathrm{deoxy}}^{\prime }$ , (f), was observed in only one subject.

Fig. 3

Color maps of occurrence probabilities of typical activation pattern (positive $\Delta C_{\mathrm{oxy}}^{\prime }$ , negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ , and positive $\Delta C_{\text{total}}^{\prime }$ ) for each measurement position.

Table 1

Mean signal amplitude of each hemoglobin species among subjects who showed typical activation pattern (positive ΔCoxy′ , negative ΔCdeoxy′ , positive ΔCtotal′ ). An activation position with the highest signal amplitude within the contralateral hemisphere to the tapping hand for each subject was used.

The numbers of subjects who showed a significant $\Delta C_{\mathrm{oxy}}^{\prime }$ , $\Delta C_{\mathrm{deoxy}}^{\prime }$ , or $\Delta C_{\text{total}}^{\prime }$ (either positive or negative) for the hemisphere contralateral to the tapping hand are shown in Table 2 . The most common change was a positive $\Delta C_{\mathrm{oxy}}^{\prime }$ (90%; left hemisphere $28∕31$ , right hemisphere $28∕31$ ). Moreover, in 91% of the subjects showing this change, at least one of the four measurement positions in the central area (left hemisphere: positions 4, 6, 7, and 9; right hemisphere: positions 16, 18, 19, and 21) was included. A negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ and a positive $\Delta C_{\text{total}}^{\prime }$ were less frequently observed ( $\Delta C_{\mathrm{deoxy}}^{\prime }$ : 76%, left hemisphere $24∕31$ , right hemisphere $23∕31$ ; $\Delta C_{\text{total}}^{\prime }$ : 73%, left hemisphere $23∕31$ , right hemisphere $22∕31$ ). In 91% of the subjects showing a negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ and in 89% of the ones showing a positive $\Delta C_{\text{total}}^{\prime }$ , at least one of the four measurement positions in the central area was included. Although $\Delta C_{\mathrm{deoxy}}^{\prime }$ was generally negative, 16% of the subjects showed a positive $\Delta C_{\mathrm{deoxy}}^{\prime }$ . A negative $\Delta C_{\mathrm{oxy}}^{\prime }$ was observed for only one subject for the hemisphere contralateral to the tapping hand.

Table 2

Number of subjects (out of 31) who showed a significant ΔCoxy′ , ΔCdeoxy′ , or ΔCtotal′ (two-tailed t test, p<0.1 ). Numbers in () show the number of subjects having a significant change at measurement positions in the central area (left hemisphere: positions 4, 6, 7, 9; right hemisphere: positions 16, 18, 19, 21).

3.2.

Variability of Activation Pattern

To determine the relationships among $\Delta C_{\mathrm{oxy}}^{\prime }$ , $\Delta C_{\mathrm{deoxy}}^{\prime }$ , and $\Delta C_{\text{total}}^{\prime }$ in detail, we classified the various patterns of combination. The patterns and appearance frequencies are shown in Table 3 . Note that a significance threshold was set for each NIRS signal, so a significant change in one signal ( $\Delta C_{\mathrm{oxy}}^{\prime }$ or $\Delta C_{\mathrm{deoxy}}^{\prime }$ ) will not result in a significant change in $\Delta C_{\text{total}}^{\prime }$ (sum of $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\mathrm{deoxy}}^{\prime }$ ) due to a subthreshold change (noise) in the other signal.

Table 3

Patterns of hemoglobin changes and their occurrence rate among 31 subjects.  *1 ↑ increase, ↓ decrease, -subthreshold change. The changes were assessed using a two-tailed t test (p<0.1) . When both changes appeared for one condition, the change for more measurement positions was used; when both changes appeared for the same number of measurement positions, the change with the maximum t value (absolute value) was used.  *2 A subject’s change characterized by positive ΔCdeoxy′ and negative ΔCoxy′ in the right hemisphere during both-hand tapping [Fig. 4(f)]. The subject showed only a negative ΔCoxy′ in the left hemisphere during right-hand taping (↓--), but a similar pattern of a negative ΔCoxy′ and a positive ΔCdeoxy′ was observed in the right (ipsilateral) hemisphere.  *3 Three cases including both hemispheres of one subject. ΔCdeoxy′ was consistently negative, though ΔCoxy′ tended to be noisy.  *4 The pattern with a positive ΔCdeoxy′ only in the right hemisphere (-↑-), while the subject showed all-positive pattern (↑ ↑ ↑) in the left (ipsilateral) hemisphere.

The most common pattern (positive $\Delta C_{\mathrm{oxy}}^{\prime }$ , negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ , positive $\Delta C_{\text{total}}^{\prime }$ ; ↑↓↑ in Table 3) was shown in 18 and 17 subjects (total 56%) for the left and right hemispheres, respectively. Similar to the most common pattern, the patterns in which $\Delta C_{\mathrm{oxy}}^{\prime }$ was positive and $\Delta C_{\mathrm{deoxy}}^{\prime }$ was negative without a positive $\Delta C_{\text{total}}^{\prime }$ (↑↓↓ and ↑↓-) were seen in four and five subjects (total 15%) for the left and right hemispheres, respectively. In addition, positive $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\text{total}}^{\prime }$ with subthreshold $\Delta C_{\mathrm{deoxy}}^{\prime }$ (↑-↑) and a positive $\Delta C_{\mathrm{oxy}}^{\prime }$ with subthreshold $\Delta C_{\mathrm{deoxy}}^{\prime }$ and $\Delta C_{\text{total}}^{\prime }$ (↑--) were observed in four and one subjects (total 6%) for the left and right hemispheres, respectively.

Another frequent pattern was the all-positive pattern (↑↑↑ in Table 3), which was observed in three and five subjects (total 13%) for the left and right hemispheres, respectively. There were mainly two types of changes in this pattern; a general type showed a strong positive $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\text{total}}^{\prime }$ with a positive $\Delta C_{\mathrm{deoxy}}^{\prime }$ [Fig. 4(c)], and a singular type that showed a strong positive $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\mathrm{deoxy}}^{\prime }$ , resulting in a stronger positive $\Delta C_{\text{total}}^{\prime }$ [Fig. 4(d)].

A significant characteristic of all these patterns was a positive $\Delta C_{\mathrm{oxy}}^{\prime }$ , which was observed in 90% of the cases.

The 10% that did not show a positive $\Delta C_{\mathrm{oxy}}^{\prime }$ were grouped into three patterns; the first is the pattern with a negative $\Delta C_{\mathrm{oxy}}^{\prime }$ and a positive $\Delta C_{\mathrm{deoxy}}^{\prime }$ in the right hemisphere during left tapping [Fig. 4(f); ↓↑- in Table 3]. In addition, the same subject showed only a negative $\Delta C_{\mathrm{oxy}}^{\prime }$ in the left hemisphere during right-hand tapping (↓--), but a similar pattern of a negative $\Delta C_{\mathrm{oxy}}^{\prime }$ and a positive $\Delta C_{\mathrm{deoxy}}^{\prime }$ was observed in the right (ipsilateral) hemisphere. We therefore classified the changes for this subject as the same pattern. Another is the pattern with a negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ only [Fig. 4(e); -↓-]. This pattern was observed in two subjects (both hemispheres in one; one hemisphere in the other). The last is the pattern with a positive $\Delta C_{\mathrm{deoxy}}^{\prime }$ only in the right hemisphere (-↑-) while the subject showed all-positive pattern (↑↑↑) in the left (ipsilateral) hemisphere [Fig. 4(d)]. Note that in every case of these three patterns there was a hemoglobin change in at least one of the hemispheres.

4.1.

Occurrence Probability of Typical Activation Pattern

First, we examined the occurrence probability of the typical activation pattern (positive $\Delta C_{\mathrm{oxy}}^{\prime }$ , negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ , and positive $\Delta C_{\text{total}}^{\prime }$ ) in the hemisphere contralateral to the tapping hand. These physiological changes were consistent with the accepted theory^{29, 30} and previous NIRS studies.^{11, 32} The occurrence probabilities in a previous study³² were nearly consistent with our results for both $\Delta C_{\mathrm{oxy}}^{\prime }$ (about 89% in the previous study and 90% in ours) and $\Delta C_{\mathrm{deoxy}}^{\prime }$ (about 84% in the previous study and 76% in ours). Moreover, the combination of a positive $\Delta C_{\mathrm{oxy}}^{\prime }$ and a negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ was observed in 73 and 71% of the cases for the previous and present study, respectively.

Our results showed that a negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ was observed less frequently in the activation center than positive $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\text{total}}^{\prime }$ (Fig. 3; left hemisphere: position 9; right hemisphere: position 16). This could be due to the smaller total number of $\Delta C_{\mathrm{deoxy}}^{\prime }$ activation positions compared to the other hemoglobin species for each subject, which is supported by our finding that the number of subjects with positive $\Delta C_{\mathrm{deoxy}}^{\prime }$ did not differ much from those with positive $\Delta C_{\text{total}}^{\prime }$ (Table 2).

One possible reason for this is the effect of systemic changes on the responses of $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\text{total}}^{\prime }$ . It has been suggested that a finger-tapping task can lead to systemic changes in blood pressure and heart rate that affect measurements of $\Delta C_{\mathrm{oxy}}^{\prime }$ in particular.^{54, 55} Both $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\text{total}}^{\prime }$ had a larger activation area, i.e., they were measured at more positions, while the activation area for $\Delta C_{\mathrm{deoxy}}^{\prime }$ was more focused. While monitoring the systemic effects and using an analytical method that subtracts the effects from the signals would be ideal,⁵⁴ a more practical approach may be to design a task paradigm that does not induce systemic variance between the rest period and task period.

Although it was difficult to distinguish the systemic effects from the cortical response in this study, the activation area for $\Delta C_{\mathrm{deoxy}}^{\prime }$ may actually have been smaller than those for $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\text{total}}^{\prime }$ for most subjects. A previous study of simultaneous recordings of fMRI and NIRS signals suggests that $\Delta C_{\mathrm{deoxy}}^{\prime }$ provides more specific information for focal cerebral responses than $\Delta C_{\mathrm{oxy}}^{\prime }$ .⁵⁶ Moreover, another study of simultaneous recordings showed that an activation map using $\Delta C_{\mathrm{oxy}}^{\prime }$ did not overlap maps using $\Delta C_{\mathrm{deoxy}}^{\prime }$ and BOLD signals.³⁵ Further study of the spatial and temporal diversities between the two signals ( $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\mathrm{deoxy}}^{\prime }$ ) is thus important. Higher spatial resolution, however, will be needed before NIR topography can be used to analyze the activation centers accurately.

We also observed that the probable positions of highest activation differed between the left (position 9) and right hemisphere (position 16). Possible reasons for this longitudinal asymmetry of the highest activation positions are a methodological problem and an anatomical characteristic of the cerebral cortex. The methodological problem is possible inaccurate placement of the measurement probes, since we placed them by hand though we did determine the positions according to the international 10–20 system. We need to develop a better method for accurate positioning. The anatomical characteristic is possible asymmetry due to the anatomical asymmetry of the cerebral cortex. According to a previous study,⁵⁷ the parietal and temporal cortices of the left hemisphere are larger than those of the right hemisphere in most subjects. As a result, even though the probes are placed symmetrically on the head surface, they may actually be shifted upward in relation to the cerebral cortex. A placement error of only a few millimeters can mislead the peak activation position to the next measurement position when the investigated area is placed at the center of the measurement area surrounded by four measurement positions.⁴⁶

Even with this possible low spatial resolution, our finding that measurement positions with high activation probabilities were around the center of the measurement area (left hemisphere: positions 4, 6, 7, and 9; right hemisphere: positions 16, 18, 19, and 21) suggests that the 10–20 international system is useful in determining the measurement area for NIR topography.

Examination of the activation probabilities in detail (Table 2) shows that a negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ , which can cause the typical BOLD signal pattern, was not as consistently observed (negative 76%, positive 16%) as a positive $\Delta C_{\mathrm{oxy}}^{\prime }$ (positive 90%, negative 3%). This peculiar feature of $\Delta C_{\mathrm{deoxy}}^{\prime }$ —the possibility of it being positive or negative—might be another reason for the less-focused activation in $\Delta C_{\mathrm{deoxy}}^{\prime }$ . Our results suggest that the $\Delta C_{\mathrm{deoxy}}^{\prime }$ signal can occasionally vary not only in amplitude but also in direction, which conflicts with the common idea of BOLD fMRI.

The variable behavior of $\Delta C_{\mathrm{deoxy}}^{\prime }$ may be due to the variability of the cross talk effect^{39, 40} among subjects; however, it is not possible to estimate the actual cross talk effect for an individual. Although we used a wavelength pair of 780 and $830\mathrm{nm}$ , which is not an optimal wavelength pair (e.g., 690 and $830\mathrm{nm}$ ) for reducing the cross talk effect, as suggested by some simulation studies,^{38, 58} we did previously determine that the primary shape of the activation signal (positive change or negative change) does not depend on the wavelength pair (either 692 and $830\mathrm{nm}$ or 782 and $830\mathrm{nm}$ ).⁴⁵ We therefore think the variability of $\Delta C_{\mathrm{deoxy}}^{\prime }$ was not due solely to the wavelengths used.

Another possibility is that the variability in $\Delta C_{\mathrm{deoxy}}^{\prime }$ reflects some physiological phenomena relating to brain activation, since unusual behavior in the $\Delta C_{\mathrm{deoxy}}^{\prime }$ or the BOLD signal has been previously reported.^{30, 33} The variability of hemodynamic responses is normally considered to be dependent on several factors such as cerebral blood flow (CBF), cerebral blood volume (CBV), and oxygen consumption rate $\left({\mathrm{CMRO}}_2\right)$ .³⁵ One previous study showed positive, negative, and silent BOLD signals with a negative, positive, and subthreshold $\Delta C_{\mathrm{deoxy}}^{\prime }$ , respectively, while $\Delta C_{\mathrm{oxy}}^{\prime }$ showed various behaviors by simultaneous measurement of fMRI and NIRS.³⁵ Another study demonstrated positive $\Delta C_{\mathrm{oxy}}^{\prime }$ and $\Delta C_{\mathrm{deoxy}}^{\prime }$ in the capillary area along with a negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ and positive $\Delta C_{\mathrm{oxy}}^{\prime }$ in the large vein area.³⁷

4.2.

Variability of Activation Pattern

Next, we examined the various patterns in more detail using classification analysis for the various combinations of hemoglobin changes (Table 3). We considered the patterns with a positive $\Delta C_{\mathrm{oxy}}^{\prime }$ and without a positive $\Delta C_{\mathrm{deoxy}}^{\prime }$ (↑ ↓ ↓, ↑ ↓-, and ↑-↑ in Table 3) to be the same as the most common pattern (↑ ↓ ↑) seen in this study. These patterns were observed in 25 and 23 subjects (77%) for the left and right hemispheres, respectively. As mentioned before, the variety of patterns for $\Delta C_{\mathrm{deoxy}}^{\prime }$ and $\Delta C_{\text{total}}^{\prime }$ could depend on CBF, CBV, and ${\mathrm{CMRO}}_2$ ^{29, 30, 33} and on the proportions of the capillary and large vein areas in the measurement region.³⁷ Simulation studies in which the proportions of these factors were changed might be useful in determining the physiological mechanism leading to each pattern. In addition, the cross talk effect due to anatomical characteristics or different SNRs in the $\Delta C_{\mathrm{deoxy}}^{\prime }$ signal may have resulted in the pattern with subthreshold $\Delta C_{\mathrm{deoxy}}^{\prime }$ (↑-↑).

The all-positive pattern (↑↑↑ in Table 3), which was occasionally observed (13%), was also reported in a number of previous NIRS studies.^{6, 22, 37} In one of the previous studies, this pattern occurred in capillary areas such as the inferior frontal gyrus (Broca’s area), while the common pattern with a negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ occurred in large vein areas such as the superior temporal area.³⁷ If the variability of the hemodynamics depends on the distribution of the vascular system, there could be a large variability in the distributive condition of vessels among subjects, even for the same cortical area.

Our observation on both the common pattern and the all-positive pattern in the same cortical area during the same activation paradigm suggests that the variation in hemoglobin changes depends on the subject’s anatomical characteristics and condition rather than the characteristics of the measurement area or the paradigm.

The patterns described earlier showing a positive $\Delta C_{\mathrm{oxy}}^{\prime }$ were observed in 90% of the subjects. This suggests high reliability of $\Delta C_{\mathrm{oxy}}^{\prime }$ as an activation signal, whereas some previous studies used only $\Delta C_{\text{total}}^{\prime }$ to identify cognitive-related activation.^{13, 16, 17}

The other 10% are interpreted as follows: an atypical pattern with a positive $\Delta C_{\mathrm{deoxy}}^{\prime }$ [Fig. 4(f); ↓↑ - and - ↑ - in Table 3] was observed in one subject. The combination pattern, which is similar to the deactivation in fMRI^{30, 33, 34, 35} and PET,^{59, 60} possibly appears when ${\mathrm{CMRO}}_2$ significantly increases during activation.³⁵ Another possibility for this pattern is that the signal pattern showed blood stealing or neural suppression due to neighboring activation, but the reason only deactivation could be observed remains unexplained. Another is the pattern with a negative $\Delta C_{\mathrm{deoxy}}^{\prime }$ only [Fig. 4(e); - ↓ - in Table 3], which was observed in two subjects (both hemispheres in one; one hemisphere in the other). This pattern might result from less sensitivity for $\Delta C_{\mathrm{oxy}}^{\prime }$ . While the $\Delta C_{\mathrm{oxy}}^{\prime }$ signal tended to increase more, the level did not reach the statistical threshold for these subjects. This is possibly due to greater noise in the $\Delta C_{\mathrm{oxy}}^{\prime }$ signal than in the $\Delta C_{\mathrm{deoxy}}^{\prime }$ one. Physiological noise, such as low frequency oscillations⁵³ or systemic changes,^{54, 55} can have more effect on $\Delta C_{\mathrm{oxy}}^{\prime }$ than on $\Delta C_{\mathrm{deoxy}}^{\prime }$ . Although it may be possible to subtract the systemic changes from NIRS signals by using simultaneously recorded data for the arterial saturation and heart rate,⁵⁴ simultaneous measurements of these physiological signals with NIRS is difficult.

Another pattern was a positive $\Delta C_{\mathrm{deoxy}}^{\prime }$ only in the right hemisphere (-↑-in Table 3) during left-hand tapping, while this subject showed an all-positive pattern (↑ ↑ ↑) in the ipsilateral (left) hemisphere as well as during contralateral right-hand tapping [Fig. 4(d)]. While activation in both hemispheres during unilateral finger movement has been reported,^{61, 62} predominant activation in the ipsilateral one rarely occurs in normal adults. Further examination with anatomical imaging will be necessary to explain this unusual lateralization.