<p>Feeding is an innate behavior critical for survival but is also influenced by many non-nutritional factors such as emotion, social context and environmental conditions. Recently, tuberal nucleus somatostatin (^TNSST) neurons have been identified as a key feeding regulation node. To gain a deeper understanding of the ^TNSST neural networks, we quantitatively characterised the brain-wide input-output configuration of mice ^TNSST neurons using the VITALISTIC method (<Emphasis Type="Underline">Vi</Emphasis>ral <Emphasis Type="Underline">T</Emphasis>racing <Emphasis Type="Underline">A</Emphasis>ssisted by <Emphasis Type="Underline">Li</Emphasis>ght-<Emphasis Type="Underline">S</Emphasis>heet microscope and <Emphasis Type="Underline">Ti</Emphasis>ssue <Emphasis Type="Underline">C</Emphasis>learing) and single-cell projectomes by fluorescence micro-optical sectioning tomography (fMOST). We found that ^TNSST neurons receive direct inputs from and send outputs to a broad range of brain regions, including many cortical and subcortical areas. Differently from AgRP neurons, the extensively studied ‘hunger’ neurons, ^TNSST neurons receive more diverse inputs from extra-hypothalamic regions and neuromodulatory centers. Using the projection-specific input tracing, we further revealed fine-tuning of the input-output configuration of ^TNSST neurons that align with specific functional needs.</p>