LC neuron types each capture specific visual features. However, Cowley.etal2024 present a 1-to-1 network model that hypothesizes that complex behaviors rely on combinations of LC neurons.
they show that almost all LC units in their network model respond to many behaviors - or visual inputs (multiple red squares per column) and many LC units contributed to each specific behavior (multiple squares per row). this highlights a possible distributed code for visual information.
