Computational and electrophysiological study demonstrates broad receptive fields enable efficient population coding in rodent barrel cortex, indicating dual parallel sensory processing schemes.
Neurons in primary sensory cortices are often grouped into functional and anatomical units with similar receptive fields (RFs), a principle known as columnar organization. According to this view, each column is an elementary computational unit processing a particular kind of sensory input. The barrels in layer 4 (L4) of the rodent somatosensory cortex exemplify this organization, as they are somatotopically aligned to individual facial whiskers and process input from single vibrissae. However, it has become increasingly clear that the same thalamic input that drives specific responses in L4, also evokes unspecific responses in the main output cells of the cortex, the layer 5 pyramidal tract (L5PT) neurons, which have broad RFs. We recently reported the mechanism underlying these broad RFs: layer 6 cortico-cortical (L6CC) cells, receive direct sensory input from thalamus and project horizontally to L5PTs across multiple columns, driving their responses. It still remains unclear what sensory information L5PTs could broadcast to their downstream targets given their unspecific responses. To address this question, we combine anatomically and functionally realistic multi-scale models of the barrel cortex with in vivo electrophysiological measurements. We find that the broad RFs of L5PTs enable a population code that broadcasts every sensory input simultaneously to any of its downstream subcortical areas. In fact, any small population of L5PTs, whether located in a single column or spanning several, can encode the identity of the sensory stimulus. The reliability and efficiency of this code depend on both the width and variability of the RFs: the broader and more diverse the RFs among L5PTs, the fewer neurons are needed to convey the same sensory information. Thus, sensory inputs appear to be represented in parallel by two complementary coding schemes: one based on the selective responses in L4, and another on the unselective, variable responses of L5PTs.
No takes yet. Share an insight, caveat, or question.
María Royo Cano (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: