221
line, and not to cross the threshold to a living being, because the current ethical maturity
of humans is not sufficient for this. “Conscious machines” are also in principle uncontrol
lable and risky. Fortunately, however, we are relatively far away from this in bioinformatic
modelling because of a number of breakthroughs that are still necessary. Nevertheless, it
is advisable to take great care already during the design phase (for example, of increas
ingly powerful Internet tools or increasingly autonomous weapon systems) to ensure that
the design prevents the worst-case scenario (the greatest accident that can be assumed),
namely the autonomous machine with consciousness or superior intelligence that begins
to control or kill humans, from the outset.
Conclusion
• Our brain is given the ability to process information very well due to its modular
design. Our genome encodes different proteins that lead to different activating
and inhibiting nerve cell connections (synapses) in numerous different nerve
cells, depending on the cell type. Nerve cell associations thus have new proper
ties (emergence). In particular, our brain is particularly good at recognizing pat
terns. Human nerve cell associations are trained in the same way as neuronal
networks in computers (see previous chapter). Training and practice strengthen
or erase synapses. Practice thus optimizes learning success over time. There are
so many nerve connections in the associative regions of our cerebrum that it
becomes advantageous to process information in an integrated rather than local
ized manner. Interference patterns similar to a hologram are created.
• We describe with our own current simulations that environmental stimuli, but
also one’s own position as well as one’s own actions can be encoded in a holo
gram for all neurons participating in the pattern equally and simultaneously. Such
new emergent effects in our particularly complex brain presumably underlie our
consciousness (“fulguration” according to Konrad Lorenz). However, bioinfor
matics already makes important contributions to neurobiology by decoding and
describing coded molecular signals at all levels. First of all, this concerns genetic
factors of neuronal maturation and diseases, which can be elucidated with the
help of the OMIM database, genome and transcriptome analyses. At the level of
the nerve cell, protein structures, in particular receptors and their activation, can
be described in detail using protein structure analyses, molecular dynamics and
databases (e.g. DrumPID, PDB database), as well as underlying cellular net
works, protein-protein interactions and signalling cascades involved.
• Brain blueprints, so-called connectomes, are already available for C. elegans and
are being intensively developed for other model organisms and humans. A connec
tome contains computer-readable information on how each nerve cell is linked to
another and which receptors and ion channels play a role in this process. Suitable
programming languages allow the direct simulation of information processing in
the brain, especially for C. elegans. Numerous special software are available for
clinical evaluations (EEG, computer tomograms) (“medical informatics”), but also
for neurobiological experiments (e.g. a neuronal activity detection tool).
15.4 Possible Objectives