The field of neuroimaging and image reconstruction is rapidly advancing, with a focus on developing more efficient and accurate methods for reconstructing visual stimuli from brain signals. Recent work has emphasized the importance of hierarchical and multimodal approaches, which can capture complex neural information and integrate multiple sources of data. These approaches have shown significant improvements in image reconstruction quality and semantic fidelity, and have the potential to enable new applications in fields such as medical diagnostics and neuroadaptive interfaces. Notable papers in this area include: Moving Beyond Diffusion, which proposes a coarse-to-fine fMRI-to-image reconstruction framework that achieves superior semantic fidelity and faster inference. GateFuseNet, which integrates QSM and T1w images for PD diagnosis using a gated fusion module, achieving 85.00% accuracy and 92.06% AUC. Brain-IT, which presents a brain-inspired approach that addresses the challenge of reconstructing images seen by people from their fMRI brain recordings through a Brain Interaction Transformer, allowing effective interactions between clusters of functionally-similar brain-voxels.