Understanding Neuronal Diversity: Role of Input Dynamics and Selectivity

The diversity of neurons in the brain is traditionally studied through morphological, electrophysiological, and transcriptomic features. However, because neurons function as spatiotemporal filters of the presynaptic activity patterns, the input dynamics should contribute to neuronal diversity. We tested this hypothesis in barrel cortical neurons using a classical ‘step-and-hold’ input and a rich dynamical input mimicking input from a presynaptic network. We found that the stimulus type strongly determines classification. To understand this input dependence further, we systematically compared which attributes are the most informative about neuronal heterogeneity by performing classifications based on four different attribute sets that capture 1) action potential features, 2) passive biophysical features, 3) adaptation, and 4) linear input filter using the Spike Triggered Average (STA). We compared the variance explained by the shared structure across these four attribute sets. We observed that the linear input filter explains the highest amount of private variance and thus is the most informative about neuronal heterogeneity. These results demonstrate the importance of the interplay between the input dynamics and the postsynaptic neuron’s linear input filter for understanding functional neuronal diversity.