Page 126 - Biomimetics : Biologically Inspired Technologies

Bar-Cohen : Biomimetics: Biologically Inspired Technologies DK3163_c003 Final Proof page 112 21.9.2005 11:41pm

112 Biomimetics: Biologically Inspired Technologies

only active (or highly excited) neurons can launch such a process and while the transponder neurons
are excited, they are not active or highly excited (i.e., active, or highly excited, neurons — a rare
state that can only exist following a confabulation information processing operation — are the only
ones that can unconditionally excite other neurons) However, as with transponder neurons, if
a neuron receives a high-enough number of simultaneous inputs from active neurons — even
through unstrengthened synapses, and in the absence of any operation command input — it will
become excited. Finally, excited neurons can excite other neurons if those other neurons reside in
a lexicon which is simultaneously also receiving operation command signal input (this is what
happens when knowledge is used and when short-term memory learning takes place, as will be
discussed below).
The wiring of the symbol and transponder neuron axons is (largely) completed in childhood and
then remains (at least for our purposes here) essentially fixed for life. Again, the gross statistics of
this wiring are genetically determined; but the local details are random.
A relatively small number (say, 1 to 25% — a genetically controlled percentage that deliberately
varies across cortex) of the target region neurons representing symbol l will just happen to each
receive many synaptic inputs from a subset of the transponder neurons (Figure 3.A.5 illustrates the
axonal connections from c transponder neurons for only one of these few l neurons). These
particular l neurons complete the knowledge link. If all of the neurons representing symbol l are
already active at the moment these synaptic inputs arrive, then (in the event that they have not been
previously permanently strengthened) the transponder neuron synapses that land on this subset of
them will be temporarily strengthened (this is called short-term memory). During the next sleep
period, if this causal pairing of symbols c and l is again deliberately rehearsed, these temporarily
strengthened synapses may be more lastingly strengthened (this is medium-term memory). If this
link is subsequently rehearsed more over the next few days, these synapses may be permanently
strengthened (this is long-term memory). It is important to note that the synapses from the c neurons
to the c transponder neurons are generally not strengthened. This is because the transponder
neurons are not meaningfully active at the time when these inputs arrive. Only deliberate usage
of a link with immediately prior co-occurrence of both source symbol and target symbol
activity causes learning. This was, roughly, the learning hypothesis that Donald Hebb advanced 56
years ago (Hebb, 1949).
Note again that the transponder neurons that represent a symbol c will always be the same;
independent of which target lexicon(s) are to be linked to. Thus, c transponder neurons must send
a sufficiently large number of axons to all of the lexicons containing symbols to which symbol c
might need to connect. The theory posits that genetic control of the distribution of axons (nomin-
ally) ensures that all of the potentially necessary knowledge links can be formed. Obviously, this
postulated design could be analyzed, since the rough anatomy and statistics of cortical axon
fascicles are known. Such an analysis might well be able to support, or raise doubts, that this
hypothesis is capable of explaining cortical knowledge.
Cognitive functions where confabulations always yield zero or one winners, because at most one
symbol has anything close to enough knowledge links from the assumed facts, do not need precisely
weighted knowledge links. In cortical modules which only require such confabulations, knowledge
links terminating within that module are hypothesized by the theory to be essentially binary in
strength: either completely unstrengthened (i.e., as yet unused) or strong (strengthened to near
maximum). Such modules together probably encompass a majority of cortex.
However, other cognitive functions (e.g., language) do require each knowledge link to have a
strength that is directly related by some fixed function to p(cjl). The theory’s hypothesis as to how
these weightings arise is now sketched.
Although the mechanisms of synaptic modification are not yet well understood (particularly
those connected with medium-term and long-term memory), research has established that ‘‘Heb-
bian’’ synaptic strengthening does occur (Cowan et al., 2001). This presumably can yield a
transponder neuron to target symbol neuron synapse strength directly related to the joint probability