Neuronal activity is the result of the orchestrated actions of multitudes of synapses, acting and evolving not as individuals, but as populations. These are shaped by essential processes of learning and homeostasis. Meanwhile, theoretical work shows that distributions of synaptic strengths reflect the computations carried out by neural systems. Yet little is known about the properties of these distributions, across brain areas, or how they change over the lifespan of an individual. We therefore aggregated, across millions of synapses, a functionally relevant proxy for synaptic strength: postsynaptic protein content, measured through a fluorescent PSD95 variant. Profiling distribution shapes by statistical moments, we found that collective synaptic properties indeed suggest different architectures across the brain.
Our measurements (see Cizeron et al., 2020 for further details) cover complete parasagittal sections of the mouse brain, in animals one week to 18 months old. We identified a hierarchical organization of regions along the anterior–posterior axis, with three divergent clusters of synaptic population shapes. One cluster includes telencephalic regions, one is centered in the midbrain and hindbrain, one comprises mainly the thalamus and the cerebellum. The structure emerging from our approach aligns with discoveries of regional patterns of cellular gene expressions. Our results suggest that synaptic populations are dynamically regulated over the lifespan. Regions which at three months of age have thinner tails largely conserve their distribution shapes later in life, whereas heavy tails in other regions strikingly grow ever more so, indicating that different processes shape the distributions, and persist over time.