Fig. 2: Validation of LAMI on lymph node samples.
From: Learned adaptive multiphoton illumination microscopy for large-scale immune response imaging
a The surface shapes of lymph nodes used for (top) training with standard candles and (bottom) testing. b Results with constant illumination power, illumination power predicted by the ray optics model that assumed a perfectly spherical shape, and illumination power predicted by LAMI in the test sample, which had been seeded with lymphocytes labeled with GFP (green), RFP (red), and eFluor670 (magenta). Constant illumination rapidly attenuates the signal with depth. The ray optics model generates contrast throughout the volume, but has visible non-uniformity and areas where the signal from cells is entirely missing. In contrast, LAMI gives good signal throughout the imaging volume up to the maximum excitation laser power. c A 3D view of the LAMI-imaged lymph node, with several XZ projections of representative areas with different surface curvature. Plots show Z-position vs mean intensity of top 5% of pixels to demonstrate good signal is maintained with depth using LAMI. d Popliteal lymph node imaged with LAMI along with XZ cross-section of predicted illumination.
