DeepCGH addresses the limitations of traditional iterative optimization methods by introducing a non-iterative approach based on a convolutional neural network with unsupervised learning. DeepCGH computes accurate holograms with fixed computational complexity, generating holograms orders of magnitude faster and with up to 41% greater accuracy compared to alternate CGH techniques. It has been demonstrated to substantially enhance two-photon absorption and improve performance in photostimulation tasks without requiring additional laser power. This innovative algorithm enables the efficient computation of accurate holograms in milliseconds, making it a valuable technique in various applications such as optogenetic photostimulation and holographic imaging. 

3D-SHOT (Three-dimensional scanless holographic optogenetics with temporal focusing) addresses the challenges of achieving precise three-dimensional targeting of custom neuron ensembles within the brain for optogenetic photostimulation. The technique utilizes computer-generated holography (CGH) and a spatial light modulator (SLM) to distribute a laser beam into multiple targets with custom 3D shapes, enabling simultaneous activation of large numbers of opsin molecules with high temporal precision. By employing CGH and temporal focusing, 3D-SHOT offers the potential for single-neuron spatial resolution and rapid initiation of action potentials with precise timing. This innovative approach enhances the capabilities of optogenetics for investigating neural circuits and their relationship to behavior, providing new opportunities for research in biology, neuroscience, and medicine.