Figure 1 presents the fundamental principle of transillumination imaging ^{[1][2][3][4][5][6][7]}[21,22,23,24,25,26,49]. RWesearchers  illuminate the animal body from one side and capture the resulting image from the opposite side (Figure 1a). This technique provides a simple and safe tool to investigate the internal structure of an animal body noninvasively. However, for NIR light, an animal body is regarded as a scattering-dominant turbid medium containing an inhomogeneous, light-absorbing internal structure. The scattering coefficient (μ_s) of human body tissue is typically of approximately 10 mm⁻¹ in near-infrared (NIR) wavelengths ^{[8][9][10][11][12][13]}[15,16,17,18,19,20]. This fact implies that the non-scattered component of light decays rapidly within a few millimeters in body tissue. However, when the body is less than a few centimeters thick, the internal light-absorbing structure can be visualized through transillumination imaging ^[7][49]. For instance, in human and animal bodies, rwesearchers c can observe a subcutaneous blood vessel network because of the marked differences in the absorption coefficients (μ_a) of hemoglobin and the surrounding tissues ^{[8][9][10][11][12][13]}[15,16,17,18,19,20].

A transillumination image represents the spatial distribution of light that reaches an imaging device through collimation optics, such as a human eye or a camera. In the transmitted light emerging from an animal body, the image-composing light is not the non-scattered component but the component that results from repeated scattering in the forward direction. When rwesearchers  illuminate an animal body with a collimated light beam, researchers cwe can observe a portion of the light in the same direction as the incident beam even through a few centimeters of thickness. Animal body tissue is known to have an anisotropy factor (g) close to 1 ^{[8][9][10][11][12][13]}[15,16,17,18,19,20], which indicates that the scattering pattern of body tissue is highly oriented towards the forward direction. Consequently, some of the light can emerge through the body tissue in the same direction as the incident light after undergoing repeated forward scattering. RWesearchers refer to this component as near-axis scattered light (NASL) ^[14][50]. It shares a similar concept to the snake photon ^[15][16][17][51,52,53] or quasi-ballistic light ^{[18][19][20][21]}[54,55,56,57]. From experimentation, rwesearchers h have confirmed that the NASL intensity remains detectable even after passing through 1 cm thick animal tissue ^[14][50]. Because NASL emerges from the animal body in the same direction as the incident light beam, the internal structure can be visualized by employing collimated optics using a regular camera lens with a small aperture.

Figure 2 presents examples of transillumination images of a human and a mouse ^[7][49]. Details of the experiments are reported elsewhere ^[22][58]. It is noteworthy that this technique can provide real-time video images with simple instrumentation in safe settings, enabling repeated or continuous bedside monitoring. In the video image of the mouse abdomen, the peristaltic movement of its digestive tract during regular food consumption can be visualized, obviating the need for any unnatural contrast medium required for X-ray fluoroscopy.

In cases where the object body is too thick to allow for light transmission in the incident beam direction, indirect light illumination can be used. Figure 1b presents this illumination technique. Given the highly diffusive nature of general body tissue, characterized by a reduced scattering coefficient μ_s′= ~1 mm⁻¹ in NIR wavelengths, the illuminated light from the side propagates almost uniformly after several millimeters of propagation. Consequently, the illuminated light is dispersed uniformly within the body, serving as back-illumination for imaging subsurface absorbing structures. This illumination technique has found application in finger-vein personal authentication ^[23][24][59,60]. Additionally, its application in the daily management of arterio-venous fistulas in dialysis treatment has been reported ^[25][61].

One important merit of NIR transillumination imaging is its capability for the noninvasive visualization of physiological changes occurring within animal bodies from an external perspective. Along with the recent progress in optoelectronic devices ^[26][27][62,63], rwesearchers c can leverage the extensive utilization of spectroscopy in the biomedical realm. As explained above, a transillumination image represents the spatial distribution of NASL observed through collimation optics. The intensity is a consequence of attenuation through animal body tissue, which is a scattering-dominant turbid medium. If researchers we disregard rigor, then the received intensity can be expressed using the following equation.

In this equation, I_in, I_out, μ_s, μ_a, d_t, and S, respectively, denote the incident light intensity, output light intensity (i.e., the transillumination image), scattering coefficient, absorption coefficient, path length in body tissue, and increment factor that represents scattering from off-axis directions. The parameters (x, y) signify the spatial distribution in the plane perpendicular to the optical axis in the z-direction of the Cartesian coordinate system. If rwesearchers  ignore the terms μ_s and S, then this equation reduces to the conventional Beer–Lambert law. Because rwesearchers c cannot neglect these terms with animal body tissue, the Beer–Lambert law is typically deemed inapplicable.

However, by obtaining transillumination images under two different conditions, wherein only the absorption condition varies while the other conditions are kept constant, absorption changes in a linear manifestation of the Beer–Lambert law can be visualized as where designations 1 and 2, respectively, denote conditions before and after the change. Equation (2) elucidates that by calculating the logarithm of intensity ratios for two images, pre-change and post-change, the resulting image delineates the spatial distribution of absorption changes while mitigating the scattering effect. Because μ_a is proportional to the molar absorption coefficient ε and the absorber concentration C, the observed absorption changes typically reflect alterations in either ε or C. By capturing transillumination images I₁ and I₂ before and after physiological changes within animal bodies, transillumination imaging can capture the internal functional changes manifesting as variations in ε or C.

One such functional change within animal bodies is the alteration in hemoglobin oxygenation. Figure 3 portrays the absorption spectra of hemoglobin and typical arterial and venous blood, showcasing distinct wavelength-dependent variations in the NIR range. Using these variations, one can visualize the distributions of venous and arterial blood separately. Figure 4 presents the results of venous–arterial differentiation in transillumination images obtained from different human wrists ^[28][64].

In the isosbestic wavelength, oxyhemoglobin and deoxyhemoglobin exhibit equal absorption. Therefore, by using this specific wavelength in the NIR range (typically of approximately 800 nm for human and rodent blood), one can independently visualize alterations in blood volume irrespective of its oxygenation status. 

Figure 5 presents the outcomes of functional imaging in a rat brain ^[29][65]. Electro-stimulation was applied to one of the rat’s forelimbs or whisker roots, with the intention of visualizing the corresponding local changes in its brain. Before conducting optical imaging, rwesearchers  identified the cerebral cortex responsible for somatosensory perception. Figure 5a shows the electroencephalograms (EEGs) observed via metal electrodes in direct contact with the mouse head during right forelimb stimulation. Figure 5b presents the functional imaging results, obtained from the transillumination images taken before and after forelimb stimulation. The localized increase in light attenuation, brought about by an augmented blood volume, agreed well with the EEG-identified regions. Stimulation of the right forelimb yielded an activated region in the left cerebral hemisphere. Upon switching the stimulation from the right to the left forelimb, the activated area transitioned symmetrically across the longitudinal cerebral fissure that separates the right and left cerebrum hemispheres. As depicted in Figure 5c, stimulation of the right/left whisker root caused different activated patterns compared to the forelimb stimulation, but remained within the respective left and right hemispheres. These positions and the localized dispersion in the somatosensory area closely correlated with measurements obtained through microelectrodes inserted into the brain ^[30][31][66,67]. It is noteworthy that these observations were noninvasive, requiring no incision or skull opening. The sole intervention involved shaving the hair on the rat’s head to enhance the clarity of the transillumination images.