**3. Results**

#### *3.1. Spiral Volumetric Optoacoustic Tomography*

A whole-body (neck to hind paws) SVOT image, reconstructed using the frames selected with the proposed motion rejection approach, is displayed in Figure 2a. In these experiments, an average

of 32% of frames were rejected per transducer position. The effectiveness of the motion rejection approach is demonstrated by analyzing three specific regions (dashed squares in Figure 2a). Specifically, we compare the image combining all the frames, with the one obtained by averaging the selected (static) frames as well as the image obtained by averaging the rejected (motion) frames. Note that the red square partially captures the thoracic region. It can be observed that a small vessel, clearly visible in the image rendered from the selected frames (Figure 2b, red arrow), cannot be resolved in the image reconstructed using all the frames. Furthermore, the former image features a clear motion artifact in the form of a 'double vessel' (red arrow), thus concealing the small vessel. The green square captures the region around the liver. A vertical vessel appears regular and complete in the selected-frames image (red arrow), whereas the same vessel appears disrupted in the all-frames image. Also here, the rejected-frames image discloses an artifact responsible for distorting the all-frames image. Finally, the blue square captures part of the abdomen. Clearly, small vessels are better resolved in the selected-frames image (red arrows). Notably, different structures appear to be blurred (yellow arrows) in the rejected-frames images with respect to the selected-frames images. A comparison between amplitude profiles of structures labeled by dashed yellow lines in Figure 2b further emphasizes the effectiveness of the motion rejection algorithm (Figure 2c) with the signal amplitude typically improved by 10% to 30% in the selected-frames images. Likewise, fine details appear more prominent in the selected-frames images, which is evinced by additional, fine peaks in the amplitudes profiles.

**Figure 2.** Motion rejection results for spiral volumetric optoacoustic tomography (SVOT). (**A**) Sagittal maximal intensity projection (MIP) of a volumetric image of the mouse reconstructed with the selected-frames (scale bar—1 cm). (**B**) Zoom-in of three regions marked in red, green, blue, respectively, in (**A**). Each image is reconstructed with (left) all-frames, (center) the selected (static) frames (right) the rejected (motion) frames (scale bar—1 mm). Structural differences are marked (yellow and red arrows). (**C**) Amplitude profiles marked in b (yellow dashed lines) for images reconstructed from all the frames (dashed lines) versus selected frames (solid lines).

## *3.2. Cross-Sectional Optoacoustic Tomography*

Effectiveness of the algorithm in cross-sectional OAT was tested by comparing the selected- and rejected-frames images taken from two distinct regions of the animal (Figure 3a). Between 20% and 31% of the frames were rejected in the top and bottom cross-sections, respectively. The rejected-frames images reveal smearing artifacts caused by a breathing motion that are evident across the entire mouse cross-section. Fine structures (red arrows) within the abdominal space appear blurred in the rejected-frames image. Moreover, some superficial structures seem to be artificially 'doubled' (yellow arrows) in the rejected-frames images. Minor differences were observed in the all-frames images (data not shown) with respect to the selected-frames images. Likewise, amplitude profiles from selected structures (dashed yellow line in Figure 3a) are increased by ~10% in the selected-frames images (Figure 3b). The calculated QI clearly reveals distortions at the boundaries of major structures, located mostly superficially (Figure 3c).

**Figure 3.** Motion rejection results for cross-sectional imaging with the ring array system. (**A**) Reconstructed transverse slices of a mouse for two different locations rendered by considering the selected (left) versus rejected (right) frames (scale bar—1 cm). Distorted structures are marked (red and yellow arrows). (**B**) Amplitude profiles (of yellow dashed lines in (**A**)) for the images rendered with all (dashed line) versus selected (solid line) frames. (**C**) Distortion-based QI of the difference between the selected- and all-frames images (1 = high similarity; −1 = low similarity).
