Exploring the impact and influence of melanin on frequency-domain near-infrared spectroscopy measurements

Abstract

Significance: Near-infrared spectroscopy (NIRS) is a non-invasive optical method that measures changes in hemoglobin concentration and oxygenation. The measured light intensity is susceptible to reduced signal quality due to the presence of melanin.

Aim: We quantify the influence of melanin concentration on NIRS measurements taken with a frequency-domain near-infrared spectroscopy system using 690 and 830 nm.

Approach: Using a forehead NIRS probe, we measured 35 healthy participants and investigated the correlation between melanin concentration indices, which were determined using a colorimeter, and several key metrics from the NIRS signal. These metrics include signal-to-noise ratio (SNR), two measurements of oxygen saturation (arterial oxygen saturation, ${\mathrm{SpO}}_2$ , and tissue oxygen saturation, ${\mathrm{StO}}_2$ ), and optical properties represented by the absorption coefficient ( ${\mu }_a$ ) and the reduced scattering coefficient ( ${\mu }_s^{\text{'}}$ ).

Results: We found a significant negative correlation between the melanin index and the SNR estimated in oxy-hemoglobin signals ( $r_s=-0.489$ , $p=0.006$ ) and ${\mathrm{SpO}}_2$ levels ( $r_s=-0.413$ , $p=0.023$ ). However, no significant changes were observed in the optical properties and ${\mathrm{StO}}_2$ ( $r_s=-0.146$ , $p=0.44$ ).

Conclusions: We found that estimated SNR and ${\mathrm{SpO}}_2$ values show a significant decline and dependence on the melanin index, whereas ${\mathrm{StO}}_2$ and optical properties do not show any correlation with the melanin index.

Keywords: melanin index; near-infrared spectroscopy; optical properties; oxygen saturation; signal-to-noise ratio.

Fig. 1

Participant distribution according to the melanin level is shown in panel (a). The melanin index is measured with $DSM-III$. Panel (b) shows the optode placement in the forehead.

Fig. 2

In both cases, SNR estimation shows a negative correlation, where panel (a) shows the correlation from $\mathrm{\Delta }[\mathrm{HbO}]$ signal is significant ($r_s=-0.489$, $p=0.006$) and panel (b) shows that for the 690 nm DC intensity, the correlation is higher, and for 830 nm (830 nm: $r_s=-0.28$, $p=0.133$), DC intensity the correlation is not significant (690 nm: $r_s=-0.568$, $p=0.001$).

Fig. 3

(a) ${\mathrm{SpO}}_2$ shows a significant negative correlation with the melanin index ($r_s=-0.146$, $p=0.44$), but (b) ${\mathrm{StO}}_2$ shows no statistically significant correlation ($r_s=-0.413$, $p=0.023$).

Fig. 4

Both (a) ${\mu }_a$ (830: $r_s=-0.024$, $p=0.9$, 690: $r_s=0.092$, $p=0.627$) and (b) ${\mu }_s^{\prime }$ (830: $r_s=-0.032$, $p=0.867$, 690: $r_s=0.11$, $p=0.562$) do not show any significant correlation with the melanin index.

Fig. 5

Total photon count number does not change ($r_s=-0.008$, $p=0.967$), but the 830 nm photon count increases ($r_s=0.506$, $p=0.004$), and the 690 nm photon count decreases significantly ($r_s=-0.491$, $p=0.006$).