At the output end, we find 
not only various sets of body 
muscles that we can 
voluntarily control, in the 
trunk, limbs, eyes, and tongue, 
but also sets that subserve the 
less voluntary or involuntary 
housekeeping functions, such 
as making our stomachs churn, 
our water pass or bowels move, 
and our sphincters (between 
these events) hold orifices 
closed.

    We also need to qualify our 
model with respect to direction 
of information flow. The 
prevailing direction in our 
diagram is obviously from left to 
right, from input to output, but 
in almost every case in which 
information is transferred from 
one stage to the next, 
reciprocal connections feed 
information back from the 
second stage to the first. (We 
can sometimes guess what such 
feedback might be useful for, 
but in almost no case do we 
have incisive understanding.) 
Finally, even within a given 
stage we often find a rich 
network of connections 
between neighboring cells of 
the same order. Thus to say that 
a structure contains a specific 
number of stages is almost 
always an oversimplification.

    When I began working in 
neurology in the early 1950s, 
this basic plan of the nervous 
system was well understood. But 
in those days no one had any 
clear idea how to interpret this 
bucket-brigade-like handing on 
of information from one stage to 
the next. Today we know far 
more about the ways in which 
the information is transformed 
in some parts of the brain; in 
other parts we still know almost 
nothing. The remaining 
chapters of this text are devoted 
to the visual system, the one we 
understand best today. I will 
next try to give a preview of a 
few of the things we know about 
that system.