Blazed oblique plane microscopy reveals scale-invariant predictions of brain-wide activity

Due to the size and opacity of vertebrate brains, it has until now been impossible to
simultaneously image neuronal circuits at cellular resolution across the entire adult brain. Thus,
any recording is an implicit subsampling of the total activity, be it by local averaging (e.g.
mesoscopic calcium imaging, fMRI), local sampling (e.g.multi-photon microscopy), or spatially
distributed sparse sampling (e.g. electrode arrays). Understanding the precise relationship
between these different sampling strategies has been a major challenge in neuroscience.
Importantly, it is unknown how to trade off these levels of description in terms of their predictive
power for global brain activity. To gain insights into this question we use our recently developed
blazed oblique plane microscopy (OPM) to perform the first brain-wide recording of neuronal
activity at single-cell resolution in an adult vertebrate, Danionella, at a volume rate of 1 Hz. We find
scale-free decay of cellular correlations and, remarkably, spatial self-similarity of predictive power.
In other words, a set of randomly sampled neurons and the same number of coarse-grained
macrovoxels predict cellular activity throughout the brain-wide population equally well.