Postscript - Brain Theory from a Circuits and Systems Perspective - page 177

the cycling via changing loop delay, and then sampling the waveform with a single
pulse, they take on some of the properties of quantum qubits. They become
simulated qubits.
Simulated qubits are not capable of teleportation, since they are basically
classical (non-quantum) devices. But like qubits, they hold true and false simulta-
neously with respective probabilities in probability space, where coordinates may
be transformed by modifying the frequency and phase of the multivibrator wave-
form. When multivibrator functions are included as defined in a chapter of this
topic, simulated qubits possess interesting possibilities for function identification
and function satisfiability. Without simulated qubits, these operations are difficult
or impossible.
Recursive neurons operating as simulated qubits are better than ordinary logic
for certain purposes, for example, for a cue editor which randomly removes cues to
achieve a match during a drawn out memory search, as described in this topic.
The most important function of simulated qubits is perhaps not probabilistic
logic, but toggling, which can be accomplished with fairly simple non-sampled
circuits as given in this topic. Recursive neurons can be made to toggle between true
and false with a single pulse acting as a trigger. This type of element was shown to
be quite helpful in achieving multi write and multi read capability in human
associative memory. Furthermore, arrays of controlled toggles operating in parallel
may accomplish rapid mental arithmetic, all subliminal, for the purposes of a quick
computation of the priority of a return. This helps assure that only important recalls
are permitted into conscious short-term memory.
The power of reversible parallel arithmetic using controlled neural toggles
cannot be stressed enough. It has been proposed that controlled neural toggles are
the secret weapon used by savants who perform amazing mental calculations, such
as multiplying large numbers, or identifying large prime numbers.
Toggle arithmetic is accomplished when a set of source toggles control a disjoint
set of destination toggles. If all of the source toggles have true outputs, then all of
the destination toggles are made to change their states, or flip true to false, or false
to true. Arithmetic is accomplished by a meaningful sequence of such operations.
This approach is logically reversible, and may be physically reversible as well,
which implies energy efficiency.
There are situations under which toggles are unconditionally reset to zero, and
this of course is not reversible. Resetting effectively occurs, for example, when cues
in a cue editor are replaced, or when accumulated priorities in a recall editor are
overwritten at the beginning of a priority calculation.
Another interesting circuit element is the short-term memory neuron. To work
within a system, this neuron must emit regular pulses that can be recognized by
connected neurons. The explanation underlying its operation is that an extended
dendritic pulse occurs when its membrane is charged, but is not permitted to
discharge normally. Charge is held on the inside surface of the membrane relative
to the outside. A short-term memory membrane discharges at a slow controlled rate,
long enough to accomplish an extended burst of regular pulses, depending on
internal ionic conditions and on conductance to the external surface.