 When you want to know what time it is, a quick glance at your watch is all you need. Even though there's only a slight difference in angle, cells in your brain respond differently to the orientation of a clock hand pointing at 3 versus 4. But not all cells care about orientation to the same extent. Why are some cells sensitive to minute changes while others are only roughly aware? Researchers have now found an answer, and in the process discovered a new type of computational power in neurons. Based on the information they receive from the retina, individual neurons in the brain's cerebral cortex respond to stimuli oriented at particular angles. Some prefer vertical edges, some prefer horizontal edges, and others prefer angles in between. But the extent to which neurons care about orientation or their degree of selectivity varies. Some are extremely choosy, distinguishing, say, between 45 and 50 degrees. Others are more lax, responding only to large differences, like between 45 and 90 degrees. The diversity of orientation preference and selectivity among neurons helps the brain build an accurate representation of natural scenes. Scientists know a lot about how orientation preference comes about, but why some cortical neurons are highly selective while others are less so has remained unclear. To uncover the mechanisms responsible for these differences in orientation selectivity, researchers in David Fitzpatrick's lab at the Mox Planck Institute for Neuroscience in Florida used in vivo 2-photon calcium imaging to visualize the responses of individual neurons in the ferret visual cortex. Cortical neurons receive information from earlier in the visual pathway through tree-like extensions called dendrites. The dendrites are covered with tiny bumps called spines, where information is transferred between neurons. Differences in orientation selectivity could simply reflect differences in the range of orientations the neuron receives information about through its spines. To test this possibility, the researchers recorded the calcium responses of many spines of many individual cortical neurons. Individual spines also preferred specific orientations, and the preferred orientation of all of a neuron's spines together reflected the neuron's overall preference. However, the selectivity of the inputs did not explain the selectivity of the neuron. Instead, selectivity was related to how a neuron's output compared with its input. Some neurons fired far fewer action potentials than expected based on the input they received, indicating that the neurons were filtering out some input. In the most extreme cases, this filtering allowed neurons to be even more selective than the sum of their individual inputs. For example, while both the neuron and the spines might prefer 45 degrees, the calcium response from all of its spines might drop only slightly between 45 and 55 degrees, whereas the neuron's firing rate would fall drastically. How was this filtering an additional selectivity created? The researchers found that spines with the same preferred orientation were clustered along the dendrites, and that the orientation preferences of the clustered spines were more likely to match the neuron's preferred orientation. Their responses were also amplified due to differences in calcium signaling within dendrites with clustered spines versus those without. As a result, the properties of the clustered inputs had a much stronger influence on the neuron's overall orientation preference and selectivity. The non-clustered, non-amplified inputs tuned to non-preferred orientations were filtered out. These results suggest dendrites aren't just passive cables faring information in the brain as previously thought. Instead, the branch has taken active role in processing information, shaping it, ultimately influencing how we see.