This inventor has a partial theory, about the sonar of dolphins, that guides his designing of software for imaging sonar:
The chin, and perhaps other parts of the face, of a bottlenose dolphin contain "echotriggers," sensors that are exquisitely and especially sensitive to
echoes of the consistently-shaped "clicks" that dolphins of that species emit. The echotriggers,
responding to features of the clicks, send brief signals directly to neurons, call them "torons," in the
brain, by-passing the cochlea entirely. An hypothesized "sonic retina" in the brain has at least one
thin layer L of torons, like a layer in the retina of an eye. For each toron T
in the sonic retina, there is imagined to be, in front of the dolphin, a thin conical region D of rays, directions
from the head, and an interval [R1,R2] of ranges such that T is fired by a click that arrives from
some echoing point in a direction in D and at a distance from the dolphin that is in [R1,R2].
There is, for toron T, a set S of echotriggers such that the firing of T results from the nearly
enough simultaneous firings of a sufficient subset of S. When T fires, this somehow indicates to
some part of the brain that an object exists roughly in the direction of the axis of the cone D.
Range-resolution can be augmented by cues and neuronal layers not discussed herein. The
hypothesized torons of the biosonar might be analogous to ganglion neurons in the
retina of an eye.
Numbers of neurons in an eye give an indication of possible numbers of torons involved
in sonic imaging. The more than 90 million photoreceptors of a human retina communicate with
more than a million retinal ganglion neurons and these communicate into the brain. The tip of a
human finger can contain more than 1000 touch-sensors; so, it seems plausible that there are
1000 or more echotriggers in the chin of a dolphin. This "1000" is combinatorially consistent with there existing a million torons.