http://www.pnas.org/content/early/2013/11/27/1316909110
The data in the study concerned are derived from a technique called
diffusion tensor imaging (DTI), which estimates how water diffuses along
white matter tracts in the brain. The big advantage of DTI is that it
can be used in living humans. Its big disadvantage is that even at best
it gives only a very low resolution and thus partial picture of possible
connections in the brain, and DTI is also rather poorly validated
against 'gold standard' anatomical tracing methods that are routinely
used in animal experiments. Thus, of the billions of connections made
via the white matter of the human brain, DTI detects a tiny fraction of
a percent and the method only indicates hypothetical anatomical
connections and not any function. As a reality check one should take the
example of the nematode worm, C elegans, which is the only animal for
which we have a complete map of every connection in its nervous system.
It has 302 neurons, compared to 85 billion in the human brain, yet no
neuroscientist (or journalist) can tell you what is in the mind of this
worm. Our knowledge of the connections in the human brain is
poverty-stricken by comparison, yet this has not stopped some
neuroscientists from linking their hypothetical structures of human
neural networks, derived from techniques like DTI, directly to complex
psychological process, as in the paper under discussion. Worse still,
the media reinterpret and amplify the scientists conjectures to build
their own confection of irresponsible speculation that bears little
relation to the original data, as the present case so ably illustrates.
- K. Martin
